Research and analysis

Research on the measurement of loneliness

Published 30 October 2025

Glossary and acronyms

Glossary

Term	Definition
Construct validity	The extent to which a tool accurately measures the concept it is intended to assess.
Dimensionality	The number and nature of distinct concepts or factors a scale measures (e.g., whether a loneliness scale captures emotional and social loneliness as separate dimensions).
Factor analysis	A statistical method used to identify underlying dimensions within a set of items or questions.
Internal consistency	A measure of whether items within a scale assess the same underlying concept.
Loneliness	A subjective, unwelcome feeling of lack or loss of companionship, which happens when there is a mismatch between the quantity and quality of the social relationships that a person has and those they desire. Specific types of loneliness that are mentioned in this report include: Chronic loneliness: Loneliness that persists over an extended period of time and may have significant impacts on a person’s health and well-being. Emotional loneliness: A type of loneliness stemming from the absence of a close emotional attachment, such as a partner or best friend. Social loneliness: A type of loneliness resulting from a lack of a broader social network or group belonging.
Measurement invariance	Indicates whether a measure works similarly across different groups (e.g., age, gender, culture).
Psychometric properties	Characteristics that indicate how well a measure performs across multiple dimensions.
Responsiveness (or sensitivity to change)	The ability of a measure to detect meaningful changes over time.
Social isolation	The objective lack of social contacts and relationships. People who experience social isolation are not necessarily lonely.
Test-retest reliability	The degree to which a measure produces consistent results when administered at different points in time.

Acronyms

Acronym	Full name
ELQ	Existential Loneliness Questionnaire
SELSA	Social and Emotional Loneliness Scale for Adults
TILS	Three-Item Loneliness Scale, also known as UCLA-3 (3-item University of California, Los Angeles Scale)

Executive summary

The overarching aim of this research was to explore how changes in loneliness can be effectively measured over time, particularly in the context of evaluating interventions. Specifically, the research sought to identify which existing measures are most valid for this purpose, what further work would support the development of loneliness measurement where gaps are identified, and what lessons can be drawn from the measurement of related constructs.

A Rapid Evidence Assessment (REA) offered a clear picture of which loneliness measures are currently used to assess change over time. However, the included studies provided limited insight into the relative strengths, limitations, and suitability of these tools to capture changes in an individual’s loneliness over time. To address this gap, interviews and focus groups with academics, researchers and representatives of sector organisations involved in loneliness interventions were conducted. Together, these findings reveal insight into the challenges of measuring loneliness over time in intervention contexts and considerations for any further development of the measurement of loneliness. 

The research aimed to answer the following questions:

1. What measures currently exist to capture changes in loneliness over time?

The Rapid Evidence Assessment (REA) identified 50 studies in which changes in individual loneliness levels were measured to evaluate the impact of a programme, intervention, or activity on targeted participants. Three loneliness measures clearly emerged as the most commonly used for this purpose: (1) University of California, Los Angeles (UCLA) Loneliness Scale; (2) De Jong Gierveld Loneliness Scale; and (3) Three-Item Loneliness Scale (TILS).

Three additional measures were identified, but appeared far less frequently, each being used only in one study in the REA sample: (1) Social and Emotional Loneliness Scale for Adults (SELSA); (2) Loneliness and Social Dissatisfaction Questionnaire (or Children’s Loneliness Scale); and (3) Campaign to End Loneliness Measurement Tool.

2. What measures are the most valid for this purpose?

The academic literature does not provide evidence for comparing loneliness measures in terms of their validity for capturing change over time. While many studies discuss the psychometric properties of loneliness scales – such as internal consistency, construct validity, and reliability – these qualities do not directly address a measure’s ability to detect meaningful changes in an individual’s loneliness levels across repeated assessments.

That said, one way to approach this question is by considering which measures are most frequently used in studies that aim to evaluate change over time. In this respect, the three most widely used scales identified in the REA – the UCLA Loneliness Scale, the De Jong Gierveld Loneliness Scale, and the Three-Item Loneliness Scale (TILS) – stand out clearly. Their widespread use was also confirmed by stakeholders engaged through interviews and focus groups.

Another way of interpreting this question is to ask which measures are best suited for evaluating the effectiveness of interventions aimed at reducing loneliness. In this context, many additional considerations must be taken into account, such as feasibility in intervention setting, sensitivity of the wording of questions, and many others. These practical dimensions were a central focus of the conversations with stakeholders and are explored further in the recommendations section that follows.

3. What can be learnt from measures of similar, sensitive topics?

The review of measures of sensitive topics similar to loneliness highlighted a number of key dimensions to consider when designing, selecting, or implementing a measure. Stakeholders shared general guidelines for identifying a robust measure of sensitive topics, including the importance of strong psychometric properties, the practicality of shorter measures, and careful consideration of the measure’s content. Moreover, stakeholders stressed the multi-faceted and complex variables underpinning sensitive emotional experiences, including loneliness. They recommended that policymakers define which variables they are most interested in measuring relative to their policy objectives before selecting a measure, particularly when considering shorter measures, as these capture fewer variables. Stakeholders also noted four unique dimensions to consider when measuring sensitive topics: frequency, duration, intensity, and impact. While frequency-based measures were identified as most popular for measuring sensitive emotional states, other dimensions were considered equally important and could be accounted for by making minor adaptations to the measure.

Co-development and close collaboration between researchers, practitioners, and the target population were emphasised throughout both the development and administration of a measure. This ensures that measures are accurate, appropriate, and meaningful to the target population – particularly when considering specific groups, such as young people, older adults, or minoritised populations. This should extend to the measure’s implementation, for example through co-designing data collection approaches or embedding local researchers within evaluation teams.

Overall, stakeholders agreed that there was no perfect, one-size-fits-all measure for assessing sensitive topics similar to loneliness that works for all applications, contexts, and populations. For instance, although lengthier measures may be able to capture multiple dimensions of a sensitive emotional experience, these were often impractical and disliked by practitioners in intervention settings due to their complexity. While the limitations of existing measures were acknowledged, stakeholders suggested that these measures remained, on the whole, well-suited to capturing change in emotional experiences over time.

4. What factors should be considered when implementing loneliness measures across different intervention types and settings?

Insights from the research point to a set of practical considerations for implementing loneliness measures effectively across diverse intervention types and populations:

• Align the measure with intervention goals: The selected loneliness measure should reflect the specific aims of the intervention. This can be supported by clearly articulating how the intervention is expected to influence loneliness, and which dimensions of the experience (e.g., perceived social support, sense of belonging, frequency of interaction) are most likely to change. Where possible, using tools that allow some flexibility – such as selecting relevant subscales or items from a broader measure – can help ensure the evaluation captures the most meaningful outcomes.
• Balance depth with feasibility: Longer, more comprehensive scales are generally better suited to capturing nuanced experiences and detecting meaningful change. Where context and resources allow, their use is encouraged. However, it’s important to consider what is realistically feasible. In some settings or with certain groups – such as older adults or people with cognitive impairments – longer tools may reduce engagement, affect data quality, or be impractical. One approach is to test comprehensive measures early to assess whether they are manageable. If too burdensome, shorter versions or alternative tools may be more appropriate.
• Ensure the measure is understood by participants: Before full implementation, it can be helpful to test the measure with a small number of participants to check whether the questions are understood as intended and feel relevant to their experience. This helps ensure the language and concepts used in the scale are clear, relatable, and appropriate for the population the intervention is targeting.
• Train those administering the measure: Many loneliness interventions are delivered by volunteers or non-specialist staff. In these cases, appropriate training is essential – not only to ensure consistent use of the tool, but also to help staff feel confident and comfortable asking potentially sensitive questions. Poorly administered measures risk undermining data quality and participant engagement.
• Adapt to the needs of the population: Certain groups may require tailored approaches to ensure accessibility and comfort. For example, older adults or people with cognitive impairments may benefit from simplified language or visual formats. In some cases, participants with limited literacy, complex life circumstances, or a lack of trust in formal processes may respond better to conversational approaches rather than standardised written tools.
• Choose follow-up methods that suit the intervention format: In interventions where participation is brief, irregular, or highly variable (such as drop-in sessions or programmes for mobile populations), traditional before-and-after measurement may not be practical. In such cases, simplified follow-up tools or immediate feedback mechanisms may be more appropriate than formal longitudinal measurement.
• Be alert to emotional risk: Loneliness is a deeply personal and sometimes painful experience. For participants in vulnerable situations (such as those experiencing grief, health challenges, or caregiving stress), completing a loneliness measure can trigger emotional responses. Practitioners should administer these tools with sensitivity and ensure appropriate support is available if needed.

4. Is there a need to develop a new metric or approach to measure changes in an individual’s loneliness over time?

The literature review and stakeholder engagement revealed limited support for developing an entirely new loneliness measure. While some argued that a purpose-built tool could better address specific shortcomings of current measures, this view was not widely shared. This does not imply that existing measures are without flaws, but rather that creating a new tool is unlikely to resolve the core challenges associated with loneliness measurement, particularly regarding the limitations of a one-size-fits-all solution (i.e., a single measure that performs equally well across all population groups and intervention settings) and the difficulty of applying these tools in the context of activities to reduce loneliness.

Stakeholders repeatedly emphasised that loneliness is a deeply subjective experience, shaped by cultural background, demographic characteristics, and individual circumstances. The issue is not that current measures are incapable of capturing change across diverse groups, but that no single metric can reasonably perform well across all populations and contexts.

Moreover, some of the more robust existing measures – such as the full version of the UCLA loneliness scale – are favoured for their comprehensiveness. Longer tools are seen as better equipped to reflect the complexity of loneliness and to detect meaningful change over time. At the same time, these longer tools are less feasible for use in real-world intervention contexts, where time constraints and participant burden are key concerns. As such, a new measure that is at once comprehensive and brief, universally valid and easy to implement, is unlikely to be realistically achievable.

Recommended pathways for developing the measurement of loneliness to better capture change over time

While this report does not identify a need to develop an entirely new loneliness measure to optimally capture changes in loneliness over time, it does present five potential options for any future development of loneliness measures for this purpose:

1. Continue with the current ONS-recommended approach

2. Adopt a more comprehensive existing scale

3. Modify and adapt an existing tool to improve its suitability for intervention contexts

4. Introduce a flexible toolkit that supports context-specific selection of loneliness measures

5. Develop a completely new measure from scratch.

While each option has benefits and limitations, it is important to highlight that there is no single best approach a priori. The most appropriate choice ultimately depends on balancing strategic priorities, such as whether greater weight should be given to robustness and sensitivity in capturing change, or to feasibility and ease of implementation within intervention settings, as well as practical considerations, including the urgency of establishing a standardised approach and the resources available to support development and implementation.

That said, based on the research findings, it is recommended to focus on either adapting and refining an existing, validated scale or developing a flexible ‘toolkit’ approach. Both alternatives would require additional research, development and testing to implement, but offer the greatest potential to balance robustness (i.e., the reliability and sensitivity of the measure in detecting meaningful changes over time), relevance (i.e., ensuring the measure captures experiences that matter to participants and aligns with the aims of loneliness interventions), and feasibility (i.e., being practical and accessible to implement across a wide range of intervention settings and populations).

1. Introduction

Alma Economics was commissioned by the Department for Culture, Media & Sport (DCMS) to explore the current landscape of loneliness measurement. This project specifically aimed to identify the tools and approaches available for measuring changes in an individual’s experience of loneliness over time, particularly in the context of evaluating programmes and interventions. It also sought to assess whether any existing measures are well-suited for this purpose, and – where gaps exist – to consider the work that could support the development of loneliness measurement.   

Background

What is loneliness?

This report defines loneliness as a subjective, unwelcome feeling of lack or loss of companionship, which happens when there is a mismatch between the quantity and quality of the social relationships that a person has and those they desire. Loneliness takes various forms, including:  

• Chronic loneliness: Loneliness that persists over an extended period of time and may have significant impacts on a person’s health and well-being
• Emotional loneliness: A type of loneliness stemming from the absence of a close emotional attachment, such as a partner or best friend
• Social loneliness: A type of loneliness resulting from a lack of a broader social network or group belonging

Why measure loneliness?

Research shows that frequent loneliness is linked to early death and poses health risks comparable to smoking or obesity (Holt-Lunstad et al. 2015). It increases the likelihood of unhealthy behaviours, heart disease, stroke, depression, cognitive decline, and Alzheimer’s (Cacioppo and Hawkley 2009; Valtorta et al. 2016; Shankar et al. 2011). Loneliness also places pressure on public services, with lonely or isolated individuals more likely to visit a GP or an A&E department, experience unplanned hospital admissions, and move into residential care (Ellaway, Wood, and Macintyre 1999; Geller et al. 1999; Molloy et al. 2010; D. W. Russell et al. 1997).

Measuring loneliness is key for further understanding its impact, identifying its causes, and finding effective ways to address it. Measurement helps to demonstrate the prevalence of loneliness across population groups. By examining its determinants and contributing factors, measurement provides valuable insight into the root causes of loneliness. Beyond this, it helps assess its effects on individuals, including mental and physical health consequences, as well as broader societal impacts. Most importantly for this research, measuring loneliness is crucial for evaluating the effectiveness of interventions and programmes, offering evidence on what works in reducing or preventing loneliness.

How is loneliness measured?

Measuring loneliness in research or evaluation settings typically involves using standardised questionnaires. These tools use one or more items asking respondents about their feelings of loneliness, or of connection and belonging. Measures vary in their wording: some are direct (e.g., ‘How often do you feel lonely?’), while others assess related experiences without mentioning loneliness explicitly, helping to reduce stigma and the consequent response bias (e.g., respondents underreporting loneliness to avoid perceived judgment). Response formats also differ, with answers provided through Likert-type scales (e.g., ‘Never’ to ‘Often’), binary options (‘Yes’ or ‘No’), or other rating systems. Responses are then scored and aggregated into an overall loneliness score. Although many measures have been developed over time, only a few have seen widespread use in research and policy evaluations.

Loneliness measurement in the UK and key challenges

In 2018, the Office for National Statistics (ONS) published a comprehensive package for measuring loneliness within a population (Office for National Statistics 2018a). This framework includes three indirect questions assessing feelings of isolation, lack of companionship, and exclusion (i.e., the Three Items Loneliness Scale), along with a single direct question asking how often someone feels lonely. A slightly adapted version of these questions was also developed for children and young people. The measure was developed through a structured process that included a review of existing scales, expert input, and a cognitive and survey testing programme, with the goal of identifying a concise and reliable set of questions for inclusion in large-scale population surveys.

The government uses the ONS-recommended approach as the standard for measuring loneliness at a population level. To enable comparisons of the effectiveness of different approaches to tackling loneliness, charities and service providers collecting loneliness data as part of their programmes are also encouraged to adopt this framework. However, while the ONS-recommended approach is statistically rigorous for measuring loneliness among the population, it may not be well-suited for assessing changes in an individual’s loneliness over time, particularly in the context of evaluating interventions.

Research commissioned by DCMS (MacIntyre and Hewings 2023) gathered feedback from stakeholders implementing and evaluating loneliness interventions, who raised concerns about the suitability of the recommended measures for this purpose. These concerns included the measure’s limited sensitivity to change, its potential unsuitability for certain individuals, and its susceptibility to external factors beyond the intervention itself. In particular, one key concern was that the ONS-recommended measure may not effectively capture the kind of change an intervention seeks to bring about and may lack sensitivity to specific intervention effects. Many programmes, such as befriending services, target particular aspects of loneliness (e.g., companionship) rather than loneliness as a whole, meaning improvements might not be reflected in overall measurement scores.

The negative wording of the questions was also flagged as a potential issue, as some practitioners worried that negatively framed questions could influence participants’ perceptions or engagement with the survey. Further concerns were raised about the suitability of loneliness measures for certain groups. While adapted versions of the ONS questions exist for children and young people, other groups – such as individuals with learning difficulties or people from specific ethnic backgrounds – may still interpret scale questions differently or struggle with their meaning, suggesting a need for more inclusive or tailored approaches in some contexts. Lastly, stakeholders noted that when used within small-scale interventions, measurement scores could be affected by external factors such as life events or participants’ current mood, making it difficult to isolate the actual impact of the intervention.

Objectives of the research

In commissioning this project to Alma Economics, DCMS aims to understand the tools available for measuring change in an individual’s loneliness levels over time. This research seeks to consider the extent to which these measures are suitable for tracking changes in loneliness over time, and – where gaps are identified – consider what further work would support the development of loneliness measurement. More specifically, this research aimed to address the following questions:

1. What metrics or approaches currently exist both within the UK and internationally to measure change in loneliness over time?

2. Which of these existing metrics or approaches are most valid for measuring an individual’s loneliness over time?

3. Is there a need to develop a new metric or approach to measure changes in an individual’s loneliness?

4. What can be learned from how similar, sensitive topics, such as well-being or self-esteem, are measured?

5. What recommendations or considerations (if any) should be taken into account when implementing loneliness measures across different types of interventions and settings?

Overview of methodology

To address the objectives of this research, project activities were structured into two phases:

• Phase 1 – Evidence review and assessment of measures: This phase involved conducting a Rapid Evidence Assessment (REA) to identify existing loneliness measures that have been used to track changes in an individual’s experience of loneliness over time, particularly in the context of evaluating the effectiveness of interventions. Identified measures were then assessed using a structured framework to examine their key characteristics, design features, and practical applications. The aim was to determine whether any of these tools could be suitable for DCMS to recommend as a standard measure for evaluating loneliness interventions. In addition, a high-level review was undertaken of measures used for similar sensitive topics – that is, subjective experiences that may carry social stigma, such as well-being and self-esteem – to identify potential lessons and best practices that could inform the refinement or development of loneliness measurement tools.
• Phase 2 – Stakeholder engagement: In the second phase, expert insights were gathered through three focus groups and ten one-to-one interviews with UK-based experts. Stakeholders included academics and researchers with expertise in loneliness (n=8), evaluation experts from sector organisations delivering loneliness interventions (n=6), and academics and researchers with expertise in similar sensitive topics (n=4). This phase aimed to explore key challenges and considerations in measuring changes in loneliness over time, gather views on the strengths and limitations of existing measures, and capture stakeholder recommendations for selecting, adapting, or developing a standardised measure to evaluate loneliness interventions in the UK.

A detailed description of the methodology is provided in Appendix 2: Methodology.

Limitations of the research

This research was designed to provide timely and practical insights and therefore used a Rapid Evidence Assessment (REA) rather than a full systematic review. While appropriate for the project scope and timeframe, this approach means some relevant studies may not have been captured. Additionally, the stakeholders engaged in the research were all UK-based academics and evaluation experts. While this provided rich and valuable insight into key themes around loneliness measurement, the findings may not reflect the full range of perspectives, including those of other stakeholders such as intervention participants and frontline practitioners, as well as those of non-UK-based experts. Limitations were also encountered in recruiting experts in the measurement of similar sensitive topics, which may have narrowed the range of transferable insights captured.

2. Measures of loneliness

This chapter presents the main findings from the evidence review and stakeholder engagement on existing tools for measuring loneliness. It begins with an overview of the measures identified through the REA, followed by a summary of stakeholder views on the most widely used scales. Together, these insights inform the strengths, limitations, and practical considerations associated with different measurement approaches in the context of evaluating loneliness interventions.

Outcomes of the REA

Through the systematic search strategy outlined in Appendix II, the protocol returned a total of 2,109 records from academic literature databases. Following a first screening based on titles and abstracts, a longlist of 141 potentially relevant studies was compiled. These were then reviewed in full against the predefined inclusion criteria, resulting in a final shortlist of 50 studies in which changes in individual loneliness levels were measured to evaluate the impact of a programme, intervention, or activity on targeted participants. Each of these 50 studies was reviewed in detail, and relevant information was captured using a structured data extraction framework.

From this final sample, three loneliness measures emerged as the most widely used for this purpose:

• University of California, Los Angeles (UCLA) Loneliness Scale – used in 21 studies
• De Jong Gierveld Loneliness Scale – used in 12 studies
• Three-Item Loneliness Scale (TILS)^{[footnote 1]} – used in 11 studies

In addition to these widely adopted measures, three other loneliness scales were identified, although they appeared far less frequently, each being used in only one study:

• Social and Emotional Loneliness Scale for Adults (SELSA)
• Loneliness and Social Dissatisfaction Questionnaire (or Children’s Loneliness Scale)
• Campaign to End Loneliness Measurement Tool

A full list of the items included in each of these scales can be found in Appendix 1: Loneliness measure items.

Finally, single-item loneliness measures were encountered in 8 studies. These consist of a standalone question directly asking about loneliness (e.g., ‘How often do you feel lonely?’), as opposed to the multi-item scales listed above, which assess loneliness through several related questions or statements. In five of the eight cases, the single-item measure was used in combination with a validated multi-item scale, rather than as the sole tool for assessing loneliness.

Brief overview of loneliness measures identified

The following section provides a brief introduction to each loneliness measure identified in the rapid evidence review. A detailed discussion of each measure – covering their conceptual foundations, item structure, response formats, scoring methods, and use across populations and interventions – is provided in Appendix 3: In-depth discussion of loneliness measures.

UCLA Loneliness Scale: The UCLA Loneliness Scale is a widely used self-report tool designed to assess subjective feelings of loneliness and social isolation. Originally developed in 1978 and revised over time, it comprises 20 items with a mix of positive and negative statements. Several shorter versions exist, including 8- and 16-item versions, as well as the TILS, which is discussed as a separate measure. It is a unidimensional scale, and its high reliability and adaptability have made it a popular choice across various populations and interventions.

De Jong Gierveld Loneliness Scale: The De Jong Gierveld Loneliness Scale measures both social and emotional loneliness using two separate subscales. The original 11-item version, and a shortened 6-item version, are composed of indirect statements and mix positive and negative wording. The scale was originally developed for older adults, and is currently mostly used with this population group. Researchers can calculate separate subscale scores or a combined loneliness score. The scale has strong psychometric properties and is especially popular in ageing research.

Three-Item Loneliness Scale (TILS): The Three-Item Loneliness Scale is a brief, unidimensional tool developed for large-scale surveys. Derived from the UCLA Loneliness Scale, it includes three negatively worded items assessing how often individuals feel isolated, left out, or lacking companionship. Using a simple three-point response scale, it is easy to administer and score. Although concise, it correlates well with longer loneliness measures and has been widely used across different age groups and intervention settings. The TILS is also part of the ONS-recommended loneliness measure, used alongside a single direct question asking, ‘How often do you feel lonely?’

Social and Emotional Loneliness Scale for Adults (SELSA): The SELSA was designed to measure loneliness as a multidimensional experience, distinguishing between social, romantic, and family loneliness. It is grounded in Weiss’s distinction between emotional and social loneliness, further refining emotional loneliness into romantic and familial dimensions. The original version contains 37 items, while a shorter 15-item version is also available. It has been primarily used with adults and university students, and its detailed structure makes it suitable for research exploring different relational contexts.

Loneliness and Social Dissatisfaction Questionnaire (Children’s Loneliness Scale): Developed for school-age children, the Loneliness and Social Dissatisfaction Questionnaire assesses subjective feelings of loneliness and dissatisfaction with peer relationships. It includes 24 items, with 16 directly related to loneliness and eight ‘filler’ items. The tool mixes direct and indirect, as well as positively and negatively worded statements. Designed to be child-friendly, it captures general feelings of loneliness and social integration, and is commonly used in developmental and educational psychology research.

Campaign to End Loneliness Measurement Tool: This three-item tool was developed by the UK’s Campaign to End Loneliness to measure loneliness among older adults. It focuses on satisfaction with relationships, availability of support, and companionship, using positively worded, indirect questions to avoid stigma. Though simple, the tool is designed to be effective in community settings and public health evaluations, offering a practical, non-intrusive way to assess loneliness in older populations.

Stakeholders’ views on main loneliness measures

During the interviews and focus groups in Phase 2 of the project, stakeholders were invited to share their views on the three most widely used loneliness measures: the UCLA Loneliness Scale, the De Jong Gierveld Loneliness Scale, and the Three-Item Loneliness Scale. Discussions focused on the strengths and limitations of these tools, particularly in relation to evaluating the effectiveness of loneliness interventions. Below is a table summarising key insights.

Measure	Strengths	Limitations
UCLA Loneliness Scale	Widely validated and highly reliable Captures multiple dimensions of loneliness Granular and nuanced – useful for research settings	Length makes it time-consuming and burdensome Language can be abstract or repetitive Less suitable for community settings or older adults with cognitive impairments
De Jong Gierveld Loneliness Scale	Differentiates between emotional and social loneliness Items are relatable, straightforward, and grounded in everyday experiences Feels more personal and empathetic, improving engagement	Emotional/social split may not always be meaningful statistically Full version (11 items) may still be too long; shorter version (6 items) loses nuance Less widely used than UCLA, reducing comparability Response options (‘yes,’ ‘no,’ ‘more or less’) may limit sensitivity
Three-Item Loneliness Scale (TILS)	Very brief and easy to administer Commonly used in national surveys – supports benchmarking Useful when time or resources are limited	Too brief to capture the full range of loneliness experiences Limited response options and score range reduce sensitivity All items are negatively worded – may introduce bias or discomfort Less effective at detecting low or moderate levels of loneliness

Other measures worth considering

Stakeholders were also asked whether there were any additional measures they believed should be considered. While most stakeholders couldn’t identify any strong candidates, some acknowledged the existence of other scales but noted that these are either less established or not well-suited for broad application. Three scales were mentioned by stakeholders:

• Social and Emotional Loneliness Scale for Adults (SELSA): Concerns were raised regarding the SELSA’s wording – particularly its ‘romantic loneliness’ subscale, which assumes all individuals seek romantic relationships, a premise that may not be appropriate for everyone.
• Campaign to End Loneliness Measurement Tool: It was developed to address known gaps in existing measures. However, it was noted that its use has largely been confined to non–peer-reviewed research (e.g., evaluation reports led by sector organisations).
• Existential Loneliness Questionnaire (ELQ): Developed by Mayers, Khoo, and Svartberg (2002), the ELQ is a relatively niche tool designed to assess existential loneliness. This construct refers to a deep sense of isolation tied to the human condition itself. The ELQ consists of 22 items, exploring themes like personal meaning, alienation, and existential anxiety, typically rated on a Likert scale (e.g., from ‘strongly disagree’ to ‘strongly agree’). It is particularly aimed at use in contexts like palliative or end-of-life care, where existential concerns often become prominent. The ELQ captures unique, underexplored dimensions of loneliness and can deepen understanding beyond social and emotional constructs. However, it has not been widely validated across diverse populations, and it is less suited for evaluating most loneliness interventions, being primarily useful in the specific contexts for which it was designed.

3. Perspectives on measuring loneliness and capturing changes over time

This chapter presents the key insights from discussions with stakeholders about the practical and conceptual challenges of measuring loneliness over time. These conversations highlighted limitations in the current measurement landscape, explored potential trade-offs in selecting or designing appropriate tools, provided insights into the measurement of similar sensitive topics, and offered reflections on how a standardised approach might be shaped going forward.

Challenges in measuring loneliness and the impact of interventions

The subjective experience of loneliness, shaped by cultural and demographic factors

One of the primary challenges in measuring loneliness is its inherently subjective nature. Loneliness is experienced and reported in widely diverse ways, shaped by cultural and demographic differences, making it difficult to capture accurately using standardised tools. Variations in how different populations understand and express loneliness can lead to potential misrepresentations in assessments.

Stakeholders noted, for example, that research suggests men may be less likely to openly acknowledge loneliness due to stigma, yet they might resonate more with specific aspects of social loneliness when assessed using certain scales. Similarly, age influences how individuals express loneliness. Younger people may be more comfortable using the term ‘loneliness’, whereas older generations, particularly those who have endured historical hardships, may view loneliness differently or avoid labelling themselves as lonely. More transitory factors, such as living arrangements, can also shape how loneliness is experienced. For instance, children in care or individuals in multi-generational households may encounter distinct forms of loneliness that differ from those experienced by the wider population.

Cultural norms around social interactions further affect perceptions of loneliness. In some societies, communal gatherings and public socialising are central to daily life, whereas others foster more individualistic lifestyles. Stakeholders emphasised that failing to account for these cultural variations may lead to misinterpretation of both baseline levels of loneliness and the impact of interventions. For example, what constitutes ‘adequate’ social contact may vary across cultures, meaning that standardised measures developed in one setting may not translate well to another. This highlights the importance of culturally sensitive tools that reflect differing norms, expectations, and values around social connection.

The influence of life circumstances on loneliness levels

Another key challenge in measuring changes in loneliness, particularly when evaluating the impact of an intervention, is its inherent variability and susceptibility to a wide range of external influences. Loneliness is not a static condition; rather, it fluctuates in response to multiple factors operating at different levels, including life events, daily experiences, and broader social and environmental conditions. Loneliness levels can shift depending on the time of day, recent interactions, or temporary emotional states. These short-term fluctuations can create noise in the data, making it difficult to detect whether observed changes are a result of the intervention or simply a reflection of ordinary emotional fluctuations.

Stakeholders highlighted that major life events, such as changes in health status, employment, or living arrangements, can have a particularly strong impact on loneliness. These external factors may either mask or mimic the effects of an intervention. If such changes are not accounted for, there is a risk of drawing misleading conclusions, either overstating or understating the intervention’s effectiveness.

As a result, simple pre- and post-intervention measurements may not provide a complete or reliable picture of real change. More sophisticated approaches may be needed to differentiate between incidental variation and meaningful, sustained improvements in loneliness levels – for instance, through repeated measurement, use of control groups, or collecting contextual data about participants’ wider circumstances.

Distinguishing real change from measurement limitations

Stakeholders pointed to examples from intervention evaluations where, despite observed improvements in outcomes such as mental health and well-being, loneliness scores remained unchanged. This raised the question: were loneliness levels genuinely unaffected, or were the employed measures not sensitive enough to detect subtle changes? In one example, a large-scale study involving link workers in England found positive effects on well-being but no measurable shift in loneliness. It remained unclear whether this reflected limitations in the intervention, issues with the wording of loneliness questions, or the possibility that loneliness is inherently more persistent over time and less immediately affected by interventions.

Sensitivity and discomfort in asking about loneliness

Beyond conceptual and methodological complexities, several practical challenges arise when loneliness tools are used in real-world settings. A widely reported concern was the emotional sensitivity of loneliness-related questions. Even when phrased indirectly or using positive language, questions in loneliness scales were often described as intrusive or uncomfortable. Volunteers delivering interventions sometimes expressed reluctance to ask them, worrying they might upset participants or feel inappropriate in the context of otherwise light or social activities. In one case, a large-scale intervention ultimately abandoned loneliness measurement altogether due to volunteer discomfort.^{[footnote 2]}

Participants were also described as hesitant to engage with loneliness surveys, sometimes refusing to complete assessments due to the personal nature of the questions or broader stigma. Timing also plays a role: when baseline assessments are conducted before any relationship or trust is established, participants may be less willing to disclose their true feelings. As a result, pre-intervention evaluations may not accurately capture lived experiences, undermining the reliability of later comparisons.

Stakeholders also noted that when interventions are not explicitly framed around loneliness – which is most often the case – the shift from a social, informal activity to an intimate self-assessment can feel abrupt and out of place. This can discourage participation and reduce the quality of data collected.

Additional practical challenges of measuring loneliness in real-world settings

In addition to concerns about the sensitivity of loneliness questions, stakeholders highlighted a range of practical difficulties related to administering loneliness measures in intervention settings:

• Challenges with repeated measures for specific groups. The use of repeated measures, such as before-and-after surveys, can be challenging to implement for certain groups. In highly mobile populations like asylum seekers, participants are often relocated frequently, making follow-up assessments unfeasible. In such cases, stakeholders suggested that evaluations might need to focus on immediate or short-term benefits rather than long-term change.
• Inconsistency in measurement delivery. Many interventions rely on volunteers or frontline workers rather than trained researchers. In some cases, questions were adapted or reworded to make them feel more conversational or suitable for the setting. While often well-meaning, these changes can compromise data comparability and reduce the overall quality of evaluation findings.
• Risk of response bias. Participants may sometimes give answers they believe reflect positively on the programme, rather than sharing their true experiences. This desire to support the intervention or please the providers can lead to inflated results, making it more difficult to assess the real impact.
• Challenges in telephone-based interventions. The mode of delivery can also affect how loneliness is measured. In telephone-based programmes – particularly common with older adults – administering standardised questions was described as more difficult. Without visual cues, it can be harder to judge participants’ emotional reactions, and structured survey questions may feel intrusive or awkward in the context of informal conversations.

Limitations of the current landscape of loneliness measures

In discussions about the challenges of measuring loneliness and tracking changes over time, concerns were raised about several limitations in the current measurement landscape. These were not framed as critiques of individual tools, but as broader reflections on the existing approaches.

Overemphasis on frequency, neglecting duration and intensity

Frequency-based questions (i.e., those asking how often someone feels a certain way) are easy to administer, but they assume that loneliness occurs in discrete, countable episodes. This overlooks key dimensions of the experience, such as the duration and intensity of loneliness – that is, how long feelings persist and how strongly they are felt (Yang 2019). For example, someone experiencing loneliness intensely and for a long time but infrequently might score the same or even lower than someone who feels briefly but often lonely.

Limited sensitivity to change

Many loneliness measures were originally designed for population-level monitoring rather than for detecting change over time. As a result, they often lack the sensitivity to register small but meaningful shifts in an individual’s experience. This limitation was frequently linked to response formats, especially those used in frequency-based scales. Broad categories may fail to capture more subtle progress. For instance, a shift from ‘often’ to ‘sometimes’ lonely can represent significant personal improvement, but is registered as just a single step on the scale. Responses also tend to cluster around middle categories like ‘sometimes’, suggesting that more granular or clearly defined options might better capture change.

Accessibility and usability across populations

Existing loneliness measures may require a degree of emotional intelligence, literacy, or self-awareness that not all individuals possess. People with lower educational backgrounds, cognitive impairments, or limited self-reflection capacity may struggle to engage meaningfully with abstract or introspective questions. This challenge becomes especially pronounced when working with hard-to-reach or vulnerable groups. Some groups, such as young offenders, refugees, or those facing extreme social exclusion, are often reluctant or unable to engage with standardised surveys. Alternative formats, such as visual tools or conversational methods, may be more appropriate, but are rarely included in current practice.

Accounting for cultural and contextual factors

All loneliness scales were originally developed in specific contexts (e.g., university students in California, older adults in the Netherlands) and have since been widely adopted, not always with full consideration of their cross-cultural validity. Applying these tools universally assumes a shared understanding of loneliness, which may not exist.

Some stakeholders questioned whether existing measures remain relevant in light of changing forms of social connectedness. Most widely used scales were developed decades ago, before digital communication became central to many people’s social lives. For example, someone might feel highly connected through online communities or group chats, yet this may not be reflected in existing loneliness measures.

More broadly, selecting a suitable loneliness measure for specific population groups remains a major challenge. Individuals from different cultural backgrounds, age groups, or life circumstances may conceptualise and express loneliness in distinct ways, making it unlikely that a single tool will be universally effective. In practice, identifying the right measure often requires extensive theoretical work, comparative testing, and reviewing the literature – a resource-intensive process that is not feasible for every demographic. As a result, generalising findings across groups or assuming equivalence in measurement validity can be highly challenging.

Recommendations and trade-offs in selecting or designing loneliness measures

Having explored the key challenges in measuring loneliness and tracking changes over time, as well as limitations of the current measurement landscape, stakeholders were invited to reflect on potential ways to address these issues. Specifically, stakeholders were asked to consider what features make a loneliness measure suitable for use in intervention contexts, and how existing challenges might be mitigated through design choices for loneliness measures or implementation approaches.

Number of items: comprehensiveness (and sensitivity to change) vs. feasibility

Longer scales were seen to provide a more nuanced picture of loneliness. By probing different aspects of the experience, they can capture a wider range of individual situations and are more likely to detect meaningful change over time. In contrast, shorter measures – though easier to administer – risk oversimplifying the experience and missing subtle shifts that might be significant to individuals.

However, the practical realities of service delivery and data collection present clear constraints. Interventions are often delivered in community settings by staff or volunteers. In these contexts, longer scales can feel burdensome to both participants and facilitators. Asking too many questions can deter participation or result in incomplete data. For some programmes, a brief set of well-phrased questions was seen as sufficient to indicate whether an intervention had helped build social connection.

The choice of measure length may also depend on what else is being assessed. Many evaluations cover loneliness and additional measures, such as those for mental health, well-being, or quality of life. In these cases, a short loneliness scale might help reduce respondent burden. However, if loneliness is the main focus of the evaluation, a longer and more comprehensive scale may be more appropriate.

Wording of the items

Direct vs. indirect approaches

The use of the terms ‘lonely’ or ‘loneliness’ in survey items can lead respondents to downplay or deny their experiences, due to the social stigma surrounding loneliness. To address this, many measures use indirect wording. This approach was often favoured as it may encourage more honest responses and avoid discomfort. Even mentioning loneliness when introducing a scale was said to risk affecting disclosure.

However, not all agreed on excluding direct language. Including a question like ‘How often do you feel lonely?’ was seen by some as offering valuable insight. While direct questions may be uncomfortable for some, others may find the term helpful in recognising and articulating their experience. One proposed solution was to combine both direct and indirect items, similar to the ONS approach.

Positive vs. negative wording

Positively framed questions were seen as less intrusive, more encouraging, and more suitable for working with emotionally vulnerable groups. These types of questions were seen as more manageable and less emotionally distressing, especially for vulnerable groups.

At the same time, negatively worded items were thought to more accurately capture loneliness. Since loneliness is inherently an aversive state, positively phrased items may instead reflect related but distinct concepts, such as social connectedness. While more comfortable to answer, they may not offer the clearest picture of a person’s experience of loneliness.

Using clear and understandable language

Clarity and simplicity of language were considered essential. Questions should avoid abstract or complex phrasing, which can lead to confusion or disengagement. Concrete, everyday language—such as ‘There is always someone I can talk to’—was seen as more effective in helping respondents reflect on their relationships. In contrast, more abstract phrases like ‘I feel lonely’ or ambiguous terms like ‘companionship’ may be interpreted in varied ways.

Establishing a timeframe

A frequent recommendation was the importance of defining a clear timeframe when asking about loneliness. Many current measures ask participants how often they feel a certain way, but do not specify the period to which these responses should refer. This lack of clarity introduces considerable variation in how individuals interpret and respond to questions, reducing the comparability and reliability of results, especially when measuring change over time in intervention contexts.

Without a set timeframe, responses can be influenced by recent or particularly memorable events, rather than reflecting the period relevant to the evaluation. Including a clearly defined reference period – such as ‘in the last week’ or ‘over the past month’ – was seen as a key step toward improving both the interpretability and the responsiveness to change of loneliness measures.

There was also recognition that trade-offs exist in deciding the length of the time frame. Shorter reference periods (e.g., ‘yesterday’) may offer a more immediate and realistic snapshot but can introduce high variability depending on daily events. In contrast, longer timeframes (e.g., ‘the past year’) can smooth out fluctuations but may be harder for respondents to recall accurately and risk overgeneralisation.

Response format and number of response options

A commonly cited limitation of existing scales, particularly those with only three or four response options, is that they offer too little granularity. These broad categories (e.g., ‘hardly ever’, ‘some of the time’, ‘often’) may feel too vague to respondents and reduce the sensitivity of the measure to detect meaningful change.

Expanding the number of response options – such as moving from three to five or six points – was seen as a simple way to allow for more nuanced answers and finer distinctions in loneliness experiences, without altering the structure of existing scales. However, it was also acknowledged that offering too many choices – such as a 0-10 scale – could introduce noise, as participants may struggle to meaningfully differentiate between closely spaced values (e.g., a 6 versus a 7).

Overall, while there was no consensus on the ideal number of response options, stakeholders broadly agreed that moving beyond binary or three-point scales can make questions more comprehensible for respondents and enhance the measure’s ability to detect change.

Other suggestions and insights

A number of additional suggestions were offered that, while raised less frequently, provide further perspectives worth considering in the development and implementation of loneliness measures.

• The three components of a loneliness measure: One perspective emphasised that a loneliness measure comprises three equally important elements: (a) the introduction to the scale (i.e., how participants are guided to interpret and respond to the questions); (b) the items themselves (i.e., the questions being asked); and (c) the response format. It was noted that while researchers often focus on the items and response categories, the framing introduction is just as important to ensure the measure is interpreted consistently across settings.
• Capturing frequency, duration, and intensity: A practical example was shared of adapting an existing loneliness scale originally focused only on frequency, to also capture intensity and duration. This approach involved asking the same core question four times: first in general terms with a ‘Yes / No / Sometimes’ format (e.g., ‘In general, do you feel left out?’), followed by three additional items asking, respectively: ‘How often does this happen?’, ‘How intense is the feeling?’, and ‘How long does it last?’ – each answered using a specific 5-point response scale.
• Using frequent assessments to track fluctuations: Loneliness can change rapidly, sometimes from day to day. Some suggested that using more frequent, short-term assessments could better reflect these dynamics and reveal patterns that single pre- and post-intervention measurements may miss.
• Tracking change over longer periods: To better assess long-lasting impacts, follow-up assessments over at least a 3–4 month period after an intervention were recommended. Shorter-term assessments may capture volatile changes, such as an immediate boost from social contact or heightened awareness of loneliness, that do not reflect sustained progress. Longer follow-up periods can help determine whether reported changes are durable and meaningful.

Considerations for the development of a standardised loneliness measure

Stakeholders shared their views on the key considerations that could inform the selection of a standard loneliness measure, reflecting on whether adopting a single, universal tool is feasible or advisable, what characteristics such a measure should possess, and whether developing a new measure might be necessary for this purpose.

Key features of a standardised measure

A number of essential characteristics were identified by stakeholders as important for a standard loneliness measure to capture changes in loneliness over time and be effective in real-world intervention settings. While specific preferences varied, several recurring themes emerged, partially overlapping with elements discussed earlier in this chapter.

• Psychometric robustness and sensitivity to change: Any standardised measure intended for evaluating interventions should demonstrate strong psychometric properties – such as reliability and validity – across a range of populations and contexts. This also includes responsiveness, or the ability to detect changes in a person’s experience of loneliness over time.
• Simplicity and clarity in wording: There was a consistent emphasis on the importance of clear and accessible language. A standard measure should be easily understood across population groups, regardless of age, education level, or cultural background. Questions should reflect real-life situations and avoid abstract or ambiguous terms.
• Minimising discomfort and stigma: It was widely agreed that loneliness measures should minimise discomfort for both respondents and those administering them. This includes avoiding language that may carry stigma or cause distress, especially for individuals in vulnerable circumstances. Some suggested using only positively phrased items (e.g., focusing on connection rather than loneliness) to encourage engagement.
• Appropriate length: Practicality was a recurring theme in the conversations. A short and concise measure was considered preferable to avoid overburdening respondents and to ensure feasibility in frontline settings. At the same time, there was recognition of the need to balance brevity with comprehensiveness so that meaningful changes over time can still be detected.
• Moving beyond frequency: The dominance of frequency-based questions was questioned. While these are common, they may not fully capture the complexity of loneliness. A more multidimensional approach – one that also incorporates intensity and duration – was suggested as more reflective of lived experience. #### Questioning the feasibility of a single standard measure Several stakeholders questioned the feasibility and appropriateness of selecting a single standardised measure to evaluate all loneliness interventions in the UK. Loneliness is not a uniform experience, and interventions often target different dimensions of it, which may not be equally captured by one tool. A measure designed to suit the majority could risk overlooking specific experiences or outcomes that matter in specific interventions.

Others raised concerns about the diversity of the population and the variation in how loneliness is experienced across cultural, demographic, and life stage contexts. From this perspective, a one-size-fits-all approach may ultimately serve no one well. Instead, loneliness assessments should be adaptable, allowing for flexibility in selecting or tailoring tools that align with the specific goals of an intervention and the characteristics of the population it serves.

A flexible ‘toolkit’ approach

Although there were concerns about the feasibility of one universal measure, there was also recognition of the value of consistency. One proposed solution was a more flexible ‘toolkit’ model, offering a standardised but adaptable framework for measuring loneliness in intervention settings.

Under this approach, researchers and practitioners would select from a pool of validated items and response formats, guided by the intervention’s goals and the characteristics of the target population. For example, the UCLA 20-item scale was identified as a foundational resource, due to its strong psychometric properties, wide applicability across population groups, and comprehensive coverage of the loneliness experience. From this base, users could identify a smaller number of items most relevant to their context – for example, selecting the four highest-scoring items at baseline for follow-up assessments, or choosing those that best align with the expected impact of the intervention. The toolkit could also offer multiple response formats (e.g., for frequency, duration, and intensity of loneliness), accompanied by practical guidance to support implementation. Pilot testing was also highlighted as a key feature of this model: before deploying a measure in a real-world setting, researchers should trial it with the relevant population to ensure the wording, structure, and focus are appropriate.

However, not all stakeholders supported this model. Concerns were raised that a toolkit model could undermine comparability across studies – already a challenge in loneliness research. From this perspective, a more flexible model, while potentially more inclusive and sensitive, could further fragment the evidence base and make it harder to assess which types of interventions are most effective.

Developing a ‘common currency’

Another proposed direction was developing a ‘common currency’ approach. Recognising that no single measure is likely to be appropriate for all populations and intervention types, this idea shifts the emphasis away from prescribing one universal tool, toward enabling comparability between the results produced by different validated measures.

Under this model, researchers would continue to use the loneliness scale most appropriate to their context, but with the support of a mapping system or conversion framework that translates scores and changes across different measures. Some reflected on previous efforts to develop such equivalences between loneliness scales, such as mapping score changes from the UCLA scale to the De Jong Gierveld scale. Although this earlier attempt was ultimately abandoned due to its complexity, the concept was still seen as holding potential, particularly as a way to preserve flexibility while improving comparability across studies and evaluations.

Risks and challenges of developing a new measure

In the final part of the focus groups and interviews, the discussion explored whether there might be value in developing a new standardised measure to evaluate loneliness interventions in the UK. Stakeholders were asked to reflect on the potential opportunities, challenges, and trade-offs between creating a new tool and continuing to use (or adapt) existing measures. The majority advised caution and emphasised a range of risks and limitations associated with developing an entirely new measure.

A central concern was the risk of losing comparability with existing literature. It was highlighted that loneliness research has a long history and that introducing a new measure could sever continuity with the existing body of work. Maintaining comparability – both with earlier studies and across groups – remains a key strength of widely used scales such as the UCLA loneliness scale, which enables researchers and practitioners to position their findings within broader national and international literature.

Stakeholders highlighted the practical difficulties of achieving widespread adoption of a new tool. Lessons from other sectors, such as health economics, suggest that even improved tools may struggle to gain traction when existing ones are already embedded in monitoring systems. Familiar measures tend to be preferred because they are widely accepted, require less training, and are already in routine use.

Concerns were also raised about the resource implications of developing a new measure from scratch. Designing, validating, and testing a new loneliness scale would be time-consuming and costly, requiring investment that might be better directed towards understanding and improving how existing tools can be used or adapted.

Crucially, stakeholders questioned whether a new measure would even solve the core challenges that practitioners face, noting that no single tool is likely to work well across all populations and interventions, given the diverse ways loneliness is understood and experienced across different age groups, cultural backgrounds, and personal contexts. A standardised measure – even one designed specifically for this purpose – is unlikely to be both comprehensive enough to capture this diversity and brief enough to be usable in intervention settings.

Rather than developing a completely new tool, several stakeholders suggested that adapting an existing measure may be a more effective and proportionate approach. One recommendation was to build on existing work – the DCMS-commissioned research Evaluation of interventions to tackle loneliness (MacIntyre and Hewings 2023) – by undertaking further qualitative research with practitioners and participants to better understand the barriers and limitations of current loneliness measures and how these relate to specific features of those tools. Insights from this work could guide targeted adaptations, such as refining question wording, adding or removing items, expanding response categories, or adjusting the timeframe for responses. In this view, adjusting existing tools could offer a more feasible and proportionate response to known challenges, balancing the need for better measures with the benefits of maintaining continuity with past research.

In contrast, some supported the idea of developing a new measure, arguing that if current tools are flawed, then aligning with past research may not be a meaningful goal. If a new, well-defined measure were to be introduced and regarded as the most appropriate tool for evaluating loneliness interventions, then transitioning to that standard would be preferable, even at the cost of comparability with past research. For organisations delivering interventions to tackle loneliness, the primary objective is evaluating services at the local level – demonstrating impact to funders and supporting ongoing service improvement, rather than maintaining continuity with previous academic studies. From this perspective, any new measure that offers clearer, more relevant insights into effectiveness would be welcomed.

Some also offered suggestions on how a new measure should be developed, if this path were pursued. A key recommendation was the involvement of statisticians and psychometric experts from the outset, to ensure that any new tool is developed through a rigorous process and meets the necessary standards of validity and reliability. Co-production was also emphasised as essential, ensuring that people with lived experience of loneliness are meaningfully involved from the outset. This would help ensure that the measure is both relevant and sensitive to how loneliness is actually experienced, particularly among groups whose perspectives may be underrepresented in existing tools. Involving practitioners and service providers was also seen as essential to ensure the measure is practical and feasible to use in real-world settings.

Reflections from the measurement of similar sensitive topics

In addition to discussions specifically focused on measuring loneliness, the project also explored what lessons might be drawn from the measurement of similar sensitive topics (e.g., mental wellbeing, self-esteem). To this end, in-depth conversations with experts in the measurement of related topics were conducted, and loneliness experts were invited to share any transferable insights from their broader research experience. The aim was to identify best practices, common pitfalls, and cross-cutting themes that could help inform a more robust and practical approach to loneliness measurement, particularly in the context of intervention evaluation.

Shared insights with loneliness measurement

Many of the perspectives shared during these conversations echoed and reinforced the themes already raised in relation to loneliness, specifically (and discussed above in this chapter). For example:

• Shorter measures are more practical to administer in intervention settings and are less likely to lead to survey fatigue or disengagement among respondents. Brief, clearly focused scales are also more likely to be delivered consistently and appropriately by practitioners or volunteers.
• Timeframes should be clearly defined in either the introduction or response format. Ambiguous or unspecified reference periods introduce unnecessary variation across responses and undermine comparability – particularly problematic when attempting to measure change over time.
• Direct questions about sensitive emotional experiences can risk underreporting, due to discomfort or stigma. While some stakeholders valued the clarity and simplicity of direct items, others cautioned that naming feelings explicitly may inhibit disclosure.
• Frequency alone may not capture the full experience. Stakeholders emphasised the value of also assessing intensity and duration to better reflect meaningful change in how a feeling is experienced over time.
• Unexpected results (e.g., lack of measurable change) should be carefully interpreted. Stakeholders cautioned that this could reflect either a failure of the intervention to deliver impact, or a mismatch between the tool and what it is intended to capture.
• There is no one-size-fits-all solution. While most existing measures of sensitive topics are robust, validated, and widely used with successful outcomes, it is unlikely that a single measure will work equally well across all populations and intervention settings.
• Measures should be co-developed with the target population to ensure they are relevant, appropriate, and accurately interpreted. This principle was seen as essential for ensuring tools resonate with respondents and are feasible to implement in real-world settings.

Additional transferable insights

A few additional insights emerged from stakeholder conversations on the measurement of other sensitive topics (such as mental well-being or self-esteem), which may offer useful guidance when shaping a standardised approach for loneliness interventions.

The importance of comprehensive psychometric properties. Stakeholders stressed that measures should demonstrate strong performance across the full range of psychometric properties – including reliability, internal consistency, test-retest reliability, factor analysis, dimensionality, and measurement invariance – and ideally be validated across different demographic groups and cultural settings. In this view, any measure considered for standardised use should either be a widely tested, established tool, or be subject to extensive testing and validation across diverse contexts if newly developed or adapted.

Clarifying the scope of the measure. Stakeholders emphasised the importance of clearly defining the concept being measured at the outset. Complex psychological experiences such as loneliness are often multi-dimensional, meaning that no single measure can capture every facet. As such, before selecting or designing a tool, it is important to identify which specific dimensions are most relevant to the intervention’s aims – for example, social disconnection, emotional isolation, or perceived belonging – and shape the measurement approach accordingly.

Involving those who will implement the measure. In addition to co-producing a measure with people who have lived experience, stakeholders highlighted the value of involving those who will be responsible for administering the tool in real-world settings. This helps ensure the measure is not only relevant and meaningful but also feasible and practical to implement. For example, frontline workers and practitioners can provide insight into how questions are received, the logistical realities of data collection, and what types of training or support may be needed to use the measure effectively.

4. Approaches to developing loneliness measurement to capture change over time

This research shows that the three most widely used loneliness measures (i.e., the UCLA Loneliness Scale, the De Jong Gierveld Loneliness Scale, and the Three-Item Loneliness Scale) are generally capable of capturing changes in an individual’s loneliness over repeated measurements. These tools have been used in the academic literature, including in robust experimental designs, and have successfully detected meaningful and statistically significant changes.

However, what emerged clearly from conversations with stakeholders is that no single ‘perfect’ loneliness measure currently exists. It is not widely thought that a measure could be developed that is simultaneously sensitive to change, valid across all populations and settings, able to capture multiple dimensions of loneliness, easy to understand, and feasible to administer in intervention contexts. This is not simply a shortcoming of current tools, but a reflection of the complex, subjective nature of loneliness and the inherent challenges of using standardised, quantitative instruments to assess it, particularly in the varied and often informal settings where loneliness interventions take place.

As such, the question of what kind of approach should be recommended as the standard for measuring changes in loneliness over time and evaluating interventions in the UK is ultimately a matter of balancing priorities. The following sections present a set of potential approaches, each with its own benefits and trade-offs, to help inform that decision.

1. Continue with the current ONS-recommended approach

This approach involves maintaining the current ONS-recommended package for measuring loneliness in the context of interventions: the Three-Item Loneliness Scale (TILS), alongside the direct question ‘How often do you feel lonely?’.

Benefits

• Widespread use and familiarity: The TILS is already used extensively in UK national datasets (e.g., Community Life Survey) and local evaluations, which enables comparability across interventions and alignment with broader national data.
• Easy-to-use: With only a few items, this approach is easy to administer, even in informal intervention settings. It requires minimal time commitment from participants and can be implemented by non-specialist staff with limited training.
• Continuity with past research: As the current UK-recommended measure, the TILS (alongside the direct question) is the most widely used tool in loneliness research and evaluation across the UK. Using it enables comparability with the existing body of evidence, supporting consistency in reporting and interpretation across studies and programmes.

Trade-offs

• Limited scope and depth: The TILS includes only three items, all focused on general feelings of social disconnection (e.g., lack of companionship, feeling left out, and feeling isolated). While these are core indicators of loneliness, the scale does not explicitly capture other important aspects. As a result, its ability to reflect the full complexity of loneliness is limited.

• Low sensitivity to change: With just three questions and only three response options per item (typically ‘hardly ever,’ ‘some of the time,’ and ‘often’), the scale has limited granularity. This reduces its sensitivity to detect small but meaningful changes over time. Given the objectives of this research, this represents a notable limitation. Several stakeholders highlighted that the TILS is not ideal for evaluating the impact of interventions, as it could miss subtle but important improvements.

• Potential for discomfort and stigma: All questions are negatively phrased, and the direct question includes the word ‘lonely’. Such phrasing can feel stigmatising, emotionally loaded, or off-putting. This discomfort may discourage engagement or result in less honest responses.

2. Adopt a more comprehensive existing scale (e.g., full UCLA Loneliness Scale)

This approach would involve replacing the current recommended measure with a more comprehensive loneliness scale already available. Based on the literature review and stakeholder input, the most suitable candidate appears to be the full UCLA Loneliness Scale, which is widely used internationally and has been tested across a broad range of populations and settings.

Benefits:

• Highly validated and psychometrically robust: The full UCLA scale has been extensively validated and is consistently regarded as one of the most robust and reliable loneliness measures.
• Widespread use: It is the most commonly used scale globally and has been employed in numerous intervention studies, supporting comparability with a large body of existing evidence.
• More nuanced coverage of loneliness: With 20 items, it captures a broader range of loneliness experiences. This makes it more likely to reflect the complexity of individuals’ experiences of loneliness and to be relevant across different demographic and cultural contexts.
• Greater suitability for tracking change: Its comprehensiveness and granularity make it better suited to detecting meaningful shifts in loneliness over time – an essential feature for evaluating intervention effectiveness.

Trade-offs:

• Longer and more demanding: With 20 items, the scale requires significant time and attention from participants. In community-based or informal settings, this may feel overly clinical or intrusive, potentially affecting participant engagement and increasing the likelihood of incomplete responses.
• Less appropriate for vulnerable or cognitively impaired groups: Stakeholders noted that the complexity and volume of questions may pose particular challenges for groups such as older adults with cognitive decline, individuals with learning difficulties, or those experiencing emotional distress.
• Harder to integrate into broader evaluation tools: In many evaluations, loneliness is only one of several outcomes being measured. The length of the full UCLA scale makes it more difficult to combine with other measures within multi-thematic evaluation tools.

3. Modify and adapt an existing tool

This approach involves selecting an established loneliness measure and making targeted adaptations to address specific challenges in using existing tools to evaluate loneliness interventions. While the evidence gathered through this research does not point to a single best candidate for adaptation, it highlights several practical improvements that could enhance the suitability of existing measures in intervention contexts.

The most widely supported suggestion from stakeholders was to introduce a clearly defined reference period for responses (e.g., ‘in the last week’ or ‘over the past month’), to improve consistency and sensitivity to change. Other potential adjustments – though raised less consistently – included:

1. Expanding response options to capture more nuanced shifts in loneliness

2. Using only positively worded items (or rephrasing negative ones)

3. Refining item wording to ensure accessibility and ease of understanding, and

4. Introducing variations in how the same questions are asked to explicitly capture frequency, intensity, and duration of the feeling.

Implementing this option would require further research and testing to identify the most appropriate existing scale for adaptation and to ensure that proposed changes improve the measure’s usability and responsiveness without compromising its validity.

Benefits:

• Improves relevance and usability: Adaptations allow the measure to be better aligned with the specific needs of different populations and contexts. Minor but strategic changes can improve how well the tool resonates with participants and how effectively it captures meaningful change.
• Builds on existing validation: Modifying an already established and validated scale allows for continuity with existing research while addressing known limitations, offering a middle ground between using an off-the-shelf tool and designing a completely new one.
• May balance sensitivity and feasibility: When adaptations are made in consultation with practitioners and service users, they can help create a measure that is both practically feasible in intervention settings and more sensitive to the changes those interventions are trying to achieve.

Trade-offs:

• Breaks full comparability with past evidence: Once modified, a scale may no longer be directly comparable with earlier studies or datasets that used the original version. This limits opportunities for benchmarking and synthesis across studies.
• Requires some testing and validation work: It is essential to test the revised tool for consistency and reliability. This requires time and resources, though not as intensively as developing a brand-new measure.
• Design trade-offs remain: Even when adapting an existing scale, there are still trade-offs to navigate. For example, efforts to capture frequency, intensity, and duration of loneliness require multiple follow-up questions, which could reduce the tool’s ease of use in practice.

4. Introduce a toolkit approach

The core idea is to provide a selection of validated loneliness scales or items, accompanied by clear, structured guidance to help researchers and practitioners choose the most appropriate tools for their intervention context. This model can take different forms, ranging from lighter-touch guidance to more tailored and modular tools.

One example of this lighter-touch interpretation is the Campaign to End Loneliness’s guidance on measuring the impact of services on loneliness in later life. A more tailored approach to a loneliness measurement toolkit could involve starting with a longer, validated scale with strong internal consistency (i.e., the extent to which all items in the scale measure the same underlying concept of loneliness) – for example, the full UCLA Loneliness Scale, which includes 20 items. The toolkit would then provide structured guidance on selecting a subset of items most relevant to the specific intervention or population, depending on, for example: (in) the intended mechanism of change (e.g., targeting social isolation or emotional loneliness); (ii) the characteristics of the target group (e.g., using only positively worded items for individuals at risk of distress or stigma); and (iii) contextual needs (e.g., time or resource constraints). Additional elements could also be customised, such as defining a suitable reference timeframe (e.g., ‘in the last month’), adapting response formats (e.g., expanding the number of response options to better detect change), or developing and incorporating versions of existing questions that also capture intensity and duration (e.g., by following up core items with prompts about how strongly or how long the feeling persists). This version of the toolkit approach is more resource-intensive to develop, but could provide a stronger balance between rigour, usability, and contextual fit.

Benefits

• Highly flexible and adaptable: The toolkit approach supports more context-specific evaluations by enabling organisations to select or adapt tools based on the goals of the intervention and characteristics of the participants.
• Can improve engagement: By aligning questions with participants’ lived experiences, the toolkit approach may reduce discomfort and improve the relevance of assessments. It can also help to minimise participant burden – especially if practitioners are empowered to use shortened tools without losing relevance.
• Can be co-produced and grounded in practice: A well-designed toolkit could be co-developed with practitioners, programme participants, and people with lived experience of loneliness. Providing guidance, examples, and decision-making support would make it easier for frontline teams to implement measurement in a meaningful and sensitive way, without needing technical expertise.

Trade-offs

• Reduces comparability across interventions: Using different measures across different contexts makes it difficult to aggregate findings or compare outcomes across interventions – potentially undermining efforts to build a unified evidence base about what works to reduce loneliness.
• Relies on practitioner judgment: Without strong, well-structured guidance, there is a risk that tool selection becomes inconsistent, ad hoc, or driven by convenience rather than suitability. This could reduce the quality and rigour of evaluations, particularly in the absence of research expertise.
• More complex and time-consuming to implement: Unlike a single standardised tool, a toolkit requires users to make informed choices each time it is applied. This adds an additional layer of complexity to planning and evaluation processes, particularly for smaller organisations with limited capacity.
• Requires further development and testing: Implementing a new toolkit approach would require further research and validation. This implies additional time, expertise, and investment to ensure the approach is both practical and methodologically sound.

5. Develop a completely new loneliness measure

While most stakeholders advised caution, a few suggested that starting from scratch could be an opportunity – particularly if the new measure is developed through a co-production process involving people with lived experience, practitioners, and researchers. A new tool could be designed to directly address the limitations of existing approaches and tailored specifically for use in tackling loneliness programmes. However, as discussed earlier in this report, a single tool that is both comprehensive and universally applicable across all populations and settings is unlikely to be realistically achievable, even if newly developed.

Benefits:

• Designed with current needs in mind: A new measure could be intentionally built to address known challenges – such as avoiding stigmatising language, improving accessibility across literacy levels, reducing participant burden, and increasing responsiveness to change. This could make it more suitable for diverse populations and better aligned with the realities of intervention delivery.
• Can embed co-production principles: From the outset, the development process could involve people with lived experience of loneliness, as well as practitioners who would be implementing the measure. This would help ensure that the tool reflects real-life experiences and is usable in the kinds of settings where loneliness interventions actually take place.
• Opportunity to unify the field with a modern standard: If well-designed and widely adopted, a new measure could serve as a shared reference point across research, policy, and practice – bridging the gap between academic rigour and practical relevance.

Trade-offs:

• Significant risk of limited added value: There are already several validated loneliness measures in use, each with known strengths and limitations. Against this backdrop, it is unlikely that a newly developed tool would lead to substantial improvements or address all current challenges, making it difficult to justify the investment.
• Time-consuming and expensive: Developing a new tool requires significant investment – not only to design and pilot the measure, but to test its reliability, validity, and responsiveness across multiple populations and contexts. This process could take several years and sustained funding.
• Adoption is uncertain: Even if the tool is well-designed, gaining widespread uptake among researchers, service providers, and funders could be challenging. The field can be cautious about adopting new measures, particularly when existing tools are already embedded in monitoring and reporting systems.

5. Conclusion

Challenges in measuring loneliness over time and assessing the impact of interventions

Measuring loneliness presents several challenges. First, loneliness is a highly subjective experience shaped by cultural and demographic factors, which makes it difficult to capture consistently using standardised tools across diverse population groups. It also fluctuates in response to life events and daily experiences, complicating efforts to distinguish the specific impact of an intervention from natural variability. A lack of measurable change in loneliness scores does not necessarily mean an intervention has been ineffective; it may also reflect limitations in the measurement tools themselves. Additionally, loneliness questions can sometimes feel intrusive or stigmatising, leading to hesitancy from both participants and facilitators when completing assessments, particularly in informal or community-based settings. Finally, practical barriers often arise in real-world implementation, including difficulties with follow-up assessments, inconsistent delivery, response bias, and constraints linked to the mode of delivery.

Limitations of the current landscape of loneliness measurement

While there are reliable and validated tools for measuring loneliness, they also present a number of limitations. Many rely too heavily on frequency-based questions, often overlooking duration and intensity, two dimensions that play a critical role in shaping the depth and impact of the experience. Existing scales may also lack the sensitivity to change needed to detect small but meaningful improvements, in part due to limited and overly broad response categories. Accessibility is another concern, as standardised tools may not be suitable for all populations and demographics. Furthermore, cultural and contextual differences can significantly influence how loneliness is experienced and expressed, which challenges the idea that a single measure can be universally valid and meaningful across different groups and settings.

Recommendations and trade-offs when selecting or designing loneliness measures

Striking the right balance between comprehensiveness and feasibility is a core consideration in selecting or designing loneliness measures. While longer scales tend to capture different dimensions of loneliness more effectively and may be more sensitive to change, they can also place a greater burden on participants and are often more difficult to implement in intervention settings.

The wording of individual items also plays a crucial role. One important distinction is between direct and indirect phrasing: while indirect questions may reduce social desirability bias and discomfort, some participants may find direct questions easier to recognise and interpret. Another issue is the framing of items: positively worded items may feel more supportive and less emotionally confronting, while negatively phrased ones may more directly tap into loneliness as an aversive state. Across all approaches, the use of clear, relatable language, grounded in everyday experience, is key to ensuring that questions are easily understood and appropriately interpreted.

Other design and implementation features can also enhance the quality and relevance of loneliness measurement. Including a specific timeframe (e.g., ‘in the past week’) can improve the consistency and interpretability of responses, though shorter or longer periods each have their own advantages and limitations. Increasing the number of response options (e.g., moving beyond three- or four-point scales) can support more nuanced assessments, though overly granular scales may reduce clarity or overwhelm respondents. Measures can also be improved by going beyond frequency alone – for example, by incorporating questions that explore the intensity and duration of the loneliness experience to more accurately reflect the depth and persistence of loneliness. In addition, it’s important to consider not only the items and response format, but also how the scale is introduced to participants, as this can be a key factor in ensuring clarity, consistency, and meaningful engagement.

Synthesis of findings in response to the research questions

The academic literature does not provide evidence for comparing loneliness measures in terms of their ability to capture change in loneliness over time. Many studies discuss specific psychometric properties^{[footnote 3]} of these tools, but not how well they detect whether someone’s loneliness has changed over time. The UCLA Loneliness Scale, the De Jong Gierveld Scale, and the Three-Item Loneliness Scale (TILS) emerged as the most used in the academic literature reviewed and were also frequently referenced by stakeholders, making them the most suitable existing candidates. However, when considering how well these measures can evaluate the impact of loneliness interventions, additional factors – such as feasibility in the intervention setting and the sensitivity of the question wording – must also be considered.

There is limited support for developing an entirely new loneliness measure. While existing tools have limitations, creating a new one is unlikely to solve the main problems – especially how to design something that works well for everyone, in all types of settings. A more realistic approach may involve adapting existing measures to improve clarity, sensitivity, and relevance, or developing a flexible toolkit model that can be adapted to different contexts while drawing from validated tools.

Implementing loneliness measures in intervention settings requires balancing accuracy with ease of use. The measure should align with the intervention’s goals, focusing on the specific aspects of loneliness it aims to influence. While comprehensive tools offer greater depth and sensitivity to change, they may not be feasible for all groups or settings; trialling them early can help determine suitability. Measures should be clearly worded, resonate with participants, and use defined timeframes to improve consistency. Adequate training for staff administering the tool is essential, as is adapting formats for specific populations and delivery modes. In informal or short-term interventions, alternative follow-up methods may be needed. Finally, using empathetic, non-stigmatising language and being aware of how sensitive loneliness questions can be, helps make sure the process is safe for everyone involved.

Recommended pathways for developing the measurement of loneliness to better capture change over time

This research identified 5 options for establishing a standardised approach to measuring changes in loneliness in the context of intervention evaluations, including:

1. Continuing with the current ONS-recommended approach

2. Adopting a more comprehensive existing scale

3. Adapting an existing scale, such as refining item wording, expanding response options, or specifying clearer timeframes

4. Developing a flexible ‘toolkit of validated items or measures, and guidance for selection; and

5. Designing an entirely new loneliness measure.

While each option has benefits and limitations, it is important to highlight that there is no single best approach a priori. The most appropriate choice ultimately depends on balancing strategic priorities, such as whether greater weight should be given to robustness and sensitivity in capturing change, or to feasibility and ease of implementation within intervention settings, as well as practical considerations, including the urgency of establishing a standardised approach and the resources available to support development and implementation.

Nevertheless, the research findings recommend focusing on either adapting and refining an existing validated scale or developing a flexible ‘toolkit’ approach.

In summary, adapting and refining an existing tool or introducing a more flexible toolkit approach offer the greatest potential to balance robustness (i.e., the reliability and sensitivity of the measure in detecting meaningful changes over time), relevance (i.e., ensuring the measure captures experiences that matter to participants and aligns with the aims of loneliness interventions), and feasibility (i.e., being practical and accessible to implement across a wide range of intervention settings and populations). The first two approaches (maintaining the current ONS measure or simply switching to a more comprehensive scale) represent quicker and less resource-intensive options but would leave some of the most well-known limitations unaddressed. The development of a completely new tool remains a high-risk and resource-heavy endeavour with uncertain added value and is therefore not recommended as a priority for future development.

Appendix 1: Loneliness measure items

UCLA Loneliness Scale

1. How often do you feel that you are ‘in tune’ with the people around you?

2. How often do you feel that you lack companionship?

3. How often do you feel that there is no one you can turn to?

4. How often do you feel alone?

5. How often do you feel part of a group of friends?

6. How often do you feel that you have a lot in common with the people around you?

7. How often do you feel that you are no longer close to anyone?

8. How often do you feel that your interests and ideas are not shared by those around you?

9. How often do you feel outgoing and friendly?

10. How often do you feel close to people?

11. How often do you feel left out?

12. How often do you feel that your relationships with others are not meaningful?

13. How often do you feel that no one really knows you well?

14. How often do you feel isolated from others?

15. How often do you feel that you can find companionship when you want it?

16. How often do you feel that there are people who really understand you?

17. How often do you feel shy?

18. How often do you feel that people are around you but not with you?

19. How often do you feel that there are people you can talk to?

20. How often do you feel that there are people you can turn to?

De Jong Gierveld Loneliness Scale

Emotional loneliness subscale:

1. I experience a general sense of emptiness

2. I miss having people around

3. I often feel rejected

4. I miss having a really close friend

5. I miss the pleasure of the company of others

6. I find my circle of friends and acquaintances too limited

Social loneliness subscale:

1. There are plenty of people I can rely on when I have problems *

2. There are many people I can trust completely *

3. There are enough people I feel close to *

4. There is always someone I can talk to about my day-to-day problems

5. I can call on my friends whenever I need them

Note: Items marked with an asterisk (*) are also included in the 6-item version of the scale.

The Three Items Loneliness Scale (TILS)

1. How often do you feel that you lack companionship?

2. How often do you feel left out?

3. How often do you feel isolated from others?

The Social and Emotional Loneliness Scale for Adults (SELSA)

Romantic loneliness subscale:

1. I am an important part of someone else’s life

2. I have a romantic partner with whom I share my most intimate thoughts and feelings

3. There is someone who wants to share their life with me

4. I have a romantic or marital partner who gives me the support and encouragement I need

5. I have an unmet need for a close romantic relationship

6. I wish I could tell someone who I am in love with, that I love them

7. I find myself wishing for someone with whom to share my life

8. I’m in love with someone who is in love with me

9. I wish I had a more satisfying romantic relationship

10. I have someone who fulfils my needs for intimacy

11. I have someone who fulfils my emotional needs

12. I have a romantic partner to whose happiness I contribute

Family loneliness subscale:

13. I feel alone when I’m with my family

14. No one in my family really cares about me

15. There is no one in my family I can depend upon for support and encouragement, but I wish there were

16. I really care about my family

17. I really belong in my family

18. I wish my family was more concerned about my welfare

19. I feel part of my family

20. My family really cares about me

21. There is no one in my family I feel close to, but I wish there were

22. My family is important to me

23. I feel close to my family

Social loneliness subscale:

24. What’s important to me doesn’t seem important to the people I know

25. I don’t have a friend(s) who shares my views, but I wish I did

26. I feel part of a group of friends

27. My friends understand my motives and reasoning

28. I feel ‘in tune’ with others

29. I have a lot in common with others

30. I have friends that I can turn to for information

31. I like the people I hang out with

32. I can depend upon my friends for help

33. I have friends to whom I can talk about the pressures in my life

34. I don’t have a friend(s) who understands me, but I wish I did

35. I do not feel satisfied with the friends that I have

36. I have a friend(s) with whom I can share my views

37. I’m not part of a group of friends and I wish I were

The Loneliness and Social Dissatisfaction Questionnaire

1. It’s easy for me to make new friends at school.

2. I like to read.

3. I have nobody to talk to.

4. I’m good at working with other children.

5. I watch TV a lot.

6. It’s hard for me to make friends.

7. I like school.

8. I have lots of friends.

9. I feel alone.

10. I can find a friend when I need one.

11. I play sports a lot.

12. It’s hard to get other kids to like me.

13. I like science.

14. I don’t have anyone to play with.

15. I like music.

16. I get along with other kids.

17. I feel left out of things.

18. There’s nobody I can go to when I need help.

19. I like to paint and draw.

20. I don’t get along with other children.

21. I’m lonely.

22. I am well-liked by the kids in my class.

23. I like playing board games a lot.

24. I don’t have any friends.

The Campaign to End Loneliness Measurement Tool

1. I am content with my friendships and relationships.

2. I have enough people I feel comfortable asking for help at any time.

3. My relationships are as satisfying as I would want them to be.

Appendix 2: Methodology

Phase 1 – Evidence review and assessment of measures

Search strategy and selection of relevant studies

We conducted a systematic search of published academic literature using the following databases: Science Direct, PubMed, Sage Journals, Cochrane Library, Taylor and Francis, SpringerLink, and JSTOR. To this aim, a structured search strategy was developed, combining a list of keywords into search strings using Boolean operators (AND/OR/NOT) and other database-specific search operators.

The resulting studies were screened using the following inclusion criteria:

• Geographic scope: Studies conducted in the UK, US, Australia, New Zealand, or Canada. Studies from comparable EU/EEA countries were also included where relevant.
• Focus: Studies measuring loneliness at two or more points in time from the same sample of individuals, with the aim of capturing potential changes in their loneliness levels to evaluate the impact of an intervention, programme, or activity.
• Publication date: Studies published between 2014 and 2024.
• Study type: Peer-reviewed journal articles, non-peer-reviewed academic outputs, government publications or government-commissioned research, publications by research organisations, evidence by providers of interventions/support, and book chapters.

Studies that did not meet these criteria were excluded. However, additional academic publications identified through snowballing – particularly those where loneliness measures were developed, refined, or validated – were reviewed and incorporated into the analysis and findings presented in the following chapters.

In addition to database searches, a manual search of grey literature was conducted, reviewing the websites of relevant sector organisations to identify additional resources. Among the materials found, the most useful for the research were guidance documents published by advocacy organisations such as Campaign to End Loneliness, What Works Wellbeing, and Ending Loneliness Together. These publications provide practical frameworks for charities, social enterprises, and grassroots organisations on how to select and implement loneliness measures to evaluate their programmes and interventions.

Evidence analysis and synthesis

The final stage of the REA involved synthesising insights from the literature to address the research questions. This process resulted in a comprehensive review of loneliness measures identified in the REA sample, detailing their development, structure, key characteristics, and use in intervention evaluations. For each loneliness measure, the following aspects were analysed and discussed:

• Background and conceptual foundation: A brief history of the measure’s development, its theoretical grounding, and the conceptual framework underpinning its approach to loneliness.
• Versions of the scale: A description of different versions of the measure, including any shortened adaptations designed for specific research requirements.
• Dimensional structure: An explanation of whether the measure distinguishes between different dimensions of loneliness (e.g., social and emotional loneliness) and how its items are structured accordingly.
• Wording of the items (direct vs. indirect, positive vs. negative): A discussion of whether loneliness is explicitly mentioned in the items (direct vs. indirect measurement) and whether items are positively or negatively worded.
• Response format: An overview of how participants provide their answers, including the type of response scale used (e.g., binary, Likert scales), and any commonly adopted variations.
• Scoring method and categorisation of results: An explanation of how responses are scored and aggregated, along with any established thresholds or categories for interpreting loneliness levels.
• Frequency, duration, and intensity of loneliness: A consideration of whether the measure captures not only the presence of loneliness (and related constructs) but also its frequency, duration, or intensity.
• Population origin and subsequent applications: An outline of the populations for which the measure was originally developed, its subsequent adaptations, and the demographic groups in which it has been most frequently used.
• Applications in the REA sample of intervention evaluations: A summary of studies within the REA that have used the measure to assess the impact of interventions on loneliness, highlighting the types of programmes evaluated and the study populations involved.

Through this structured approach, a comprehensive, yet comparative overview of the measures identified in the REA was provided, facilitating an understanding of their strengths, limitations, and suitability for different research contexts.

Review of measures of similar sensitive topics

Alongside the REA on loneliness measurement, a high-level review was conducted of tools used to measure similar sensitive topics, such as mental well-being, life satisfaction, quality of life, and psychological distress. The aim was to identify transferable lessons that could inform the selection, adaptation, or development of loneliness measures.

The measurement of these topics often faces challenges similar to those encountered when measuring loneliness, including ensuring validity, addressing awareness and sensitivity issues, and mitigating the potential impact of external factors on responses. As such, this review aimed to supplement rather than replicate the REA on loneliness measures by collating transferable insights on best practices for assessing individual-level change over time.

To achieve this, academic and reputable grey literature was reviewed to identify and assess widely used measures of similar sensitive topics. Given the exploratory nature of this review, the inclusion criteria were broader than the REA and included:

• Domain: Measures which assess individual-level well-being or mental health states.
• Geographic scope: Studies conducted in the UK, US, Australia, New Zealand, or Canada. Studies from comparable EU/EEA countries were also included where relevant.
• Language: Measures which could be administered in English.
• Study type: Peer-reviewed journal articles, non-peer-reviewed academic outputs, government publications or government-commissioned research, publications by research organisations, evidence by providers of interventions/support, and book chapters.

We shortlisted 10 measures from an initial longlist (n=34) which were indicated as the most robust for measuring similar sensitive topics. We extracted best practices and other transferable insights from this shortlist, which were used to inform the subsequent stakeholder engagement and wider research recommendations for loneliness measurement.

Phase 2 – Stakeholder engagement

In the second phase of the project, three focus groups and ten interviews were conducted with relevant UK-based experts, targeting the following three stakeholder groups:

1. Academics and researchers with expertise in loneliness (n=8)

2. Evaluation experts from sector organisations with expertise in loneliness (n=6)

3. Academics and researchers with expertise in similar sensitive topics (n=4)

Although only a subset of stakeholders were selected for their expertise in similar sensitive topics, discussions on this theme also featured across other interviews and focus groups, as many stakeholders had relevant cross-cutting experience.

Recruitment

Recruitment of academics and researchers was led by Alma Economics. The approach to identify relevant stakeholders included reviewing authors of studies identified through the REA, examining key loneliness research beyond the REA sample, scanning relevant university department profiles, and applying snowballing techniques. This process generated a longlist of 32 academics and researchers with expertise in loneliness, and 26 with expertise in similar, sensitive topics. These individuals were invited to participate either directly by Alma or via DCMS where an existing relationship existed, to help facilitate engagement. Recruitment of evaluation experts from sector organisations was led by DCMS, where there were existing relationships. To accommodate stakeholder availability and preferences, a choice was offered between joining a focus group or taking part in a one-on-one interview. These sessions were conducted over a five-week period between February and March 2025.

Appendix 3: In-depth discussion of loneliness measures

Sources of evidence for the assessment of measures

The information provided on each measure is drawn from multiple sources, including (1) the studies in the REA, which offer insight into how each scale has been applied in practice, including the intervention context and target population; (2) the original publications where the scales (and any of their adaptations or subsequent versions) were developed and validated; (3) relevant pieces of grey literature; and, for some measures, (4) a recent key comparative study by Maes et al. (2022).

The latter is based on the MASLO (‘Meta-Analytic Study of Loneliness’) project – a large database compiling studies that have used one of eight loneliness questionnaires – and offers two key insights for the analysis. First, it provides data on how frequently each measure has been employed across different age groups and continents. Second, it presents an independent coding of all items in the eight scales, categorising them as measuring emotional or social loneliness – even when a scale was not originally built with this distinction in mind. However, not all the measures identified in the REA are included in the MASLO database, meaning this information will not be available for every measure discussed in the following sections.

The University of Los Angeles, California (UCLA) Loneliness Scale

Background and conceptual foundation

The UCLA Loneliness Scale (UCLA-LS) was first created by D. Russell, Peplau, and Ferguson (1978) as a response to the need for a simple, valid and reliable general loneliness scale. It has since undergone multiple revisions, with updated versions released in 1980 and 1996 (D. Russell, Peplau, and Cutrona 1980; Daniel W. Russell 1996). It represents one of the most psychometrically sound and widely used self-report measures of loneliness, as it has demonstrated strong validity and reliability, with high internal consistency and good construct validity (Penning, Liu, and Chou 2014; Alsubheen et al. 2023).

This measure is grounded in a subjective conceptualisation of loneliness, defining it as the distressing experience that results from perceived deficiencies in one’s social relationships, leading to dissatisfaction (D. Russell, Peplau, and Cutrona 1980). In this sense, it focuses on capturing an individual’s subjective feelings of loneliness and social isolation (D. W. Russell 1996).

Versions of the scale

The original version of the UCLA Loneliness Scale (UCLA-LS), developed in 1978, comprised 20 self-report items, all phrased in a negative direction,^{[footnote 4]} with high scores reflecting feelings of social dissatisfaction (D. Russell, Peplau, and Ferguson 1978). Concerns about potential response bias due to the negative stigma surrounding loneliness led to a revision of the scale in 1980. This resulted in the Revised UCLA Loneliness Scale (R-UCLA-LS), which incorporated ten (10) positively reworded statements alongside ten (10) negatively worded items retained from the original version (D. Russell, Peplau, and Cutrona 1980).

As the scale began to be used across a diverse range of populations beyond its original target group (i.e., young adults), concerns arose regarding the clarity of the statements. To address this, the UCLA Loneliness Scale Version 3 (UCLA-LS3) was developed by D. W. Russell (1996), featuring simplified item content and response formats to enhance usability across broader groups (Alsubheen et al. 2023).

Several shortened versions of the UCLA Loneliness Scale (UCLA-LS) have been developed to suit research contexts that require more concise assessments and diverse demographic samples. Among the most frequently cited in the literature are the UCLA-16, developed for a Portuguese sample of older adults (Faustino et al. 2019; Jesus et al. 2024); the UCLA-8 (Hays and DiMatteo 1987), used in a wide range of interventions and samples, such as Australian community-dwelling adults (Dingle et al. 2024), German adults (Liu et al. 2023), and US college students (Bruehlman-Senecal et al. 2020); the UCLA-6, developed for a sample of Portuguese older adults (Neto 2014) and later implemented in Norway to assess loneliness in high school students (Urke, Larsen, and Kristensen 2023); and the UCLA-4, a four-item survey consisting of two positively and two negatively worded items (D. Russell, Peplau, and Cutrona 1980). However, the most widely used shortened version is the Three-Item Loneliness Scale (TILS), also known as the UCLA-3, which will be discussed in greater detail in the following section (Hughes et al. 2004).

Dimensional structure

The UCLA-LS is a unidimensional measure that consists of 20 items designed to assess subjective feelings of loneliness and social isolation (Daniel W. Russell 1996). Some studies have suggested the presence of underlying subdimensions, for instance, Maes et al. (2022) coded the 20 items of the scale and concluded that seven items measure emotional loneliness, eleven items measure social loneliness, and the remaining two are unrelated to loneliness. Nonetheless, there is no general consensus in the literature regarding the specific types of loneliness potentially captured by the scale.

Wording of the items (direct vs. indirect, positive vs. negative)

The UCLA Loneliness Scale (UCLA-LS) incorporates a blend of indirect statements to measure loneliness and direct statements to assess feelings of social connectedness and isolation.

To minimise response bias, the scale features both positively and negatively worded items. It consists of 11 negatively worded items and 9 positively worded items. The item format was revised, transforming the statements into questions by adding the phrase ‘How often do you feel…?’. For instance, the item ‘I lack companionship’ was rephrased to ‘How often do you feel that you lack companionship?’(D. Russell, Peplau, and Cutrona 1980; Daniel W. Russell 1996).

Response format

The scale uses a four-point Likert-type response format ranging from 1 (‘Never’), 2 (‘Rarely’), 3 (‘Sometimes’) and 4 (‘Always’) for the direct-scored, or negatively worded, items. For the reversed score, or positively worded items, the format ranges from 4 (‘Never’), 3 (‘Rarely’), 2 (‘Sometimes’) and 1 (‘Always’) (Daniel W. Russell 1996).

Scoring method and categorisation of results

Since the scale is intended to capture a unidimensional construct of loneliness, a single overall loneliness score is calculated by summing responses across all items. Considering each item is scored from 1-4, the score of the 20-item scale ranges from 20 to 80, with higher scores indicating higher perceived loneliness (Daniel W. Russell 1996). Short-form versions maintain this scoring approach but with adjusted ranges based on item count.

A universally accepted categorical classification of loneliness based on the UCLA-LS scores does not exist. However, some studies have applied percentile-based cut-off points to classify respondents into low, moderate or high loneliness groups (Theeke et al. 2015; Käll et al. 2020; Berger et al. 2024; Patapoff et al. 2024; Yan, Johnson, and Jones 2024).

Frequency, duration and intensity of loneliness

The UCLA Loneliness Scale is designed to assess how frequently individuals experience loneliness, rather than its intensity or duration (Maes et al. 2022). The wording of its items focuses on recurring feelings of social isolation or disconnection, rather than the depth or persistence of these emotions, with response categories ranging from 1 (never) to 4 (often).

Population origin and applications

The scale was originally developed and validated in English using a sample of 239 young adults recruited at the University of California, Los Angeles (D. Russell, Peplau, and Ferguson 1978). Among the studies included in the MASLO database, 64.24% utilised the UCLA Loneliness Scale, making it the most frequently used loneliness measure. Studies in the MASLO database indicate its broad applicability, with substantial usage among college students (35.11%), adults (33.70%), adolescents (17.28%), and older people (13.30%). The scale has been translated into multiple languages and is widely used in North America (59.69%), Europe (20.06%), and Asia (17.04%), with increasing applications in Latin America and Africa (Maes et al. 2022).

In the REA, the UCLA scale has been used across a variety of populations, such as adolescents (Urke, Larsen, and Kristensen 2023; P. Iyer et al. 2024; 2024), adults ( et al. 2017; A Käll et al. 2021), older adults (Theeke et al. 2015; Deol et al. 2022; Richards et al. 2024; Lim et al. 2024), college students (Bruehlman-Senecal et al. 2020; Calderon Leon et al. 2024), cancer patients (Adams et al. 2017), and those with cognitive impairment (L. A. Theeke, Mallow, and Theeke 2021).

De Jong Gierveld Loneliness Scale

Background and conceptual foundation

First developed by de Jong Gierveld and colleagues in 1985 (De Jong-Gierveld and Kamphuls 1985) and later refined in a manual published in 1999 (De Jong-Gierveld and Van Tilburg 1999), the De Jong Gierveld Loneliness Scale is among the most commonly used measures of loneliness (Maes et al. 2022) This scale exhibits good psychometric properties, which demonstrate construct and convergent validity, as well as acceptable internal consistency (De Jong-Gierveld and Kamphuls 1985).

The scale is grounded in the cognitive theoretical approach to loneliness, which emphasises the discrepancy between an individual’s desired level of interpersonal affection and intimacy and their actual social reality. In this view, loneliness is understood as a subjective experience rather than a direct consequence of situational factors. It is defined as an unpleasant or unacceptable lack of certain relationships, either in terms of quantity or quality. Loneliness, in this framework, encompasses both a deficit in the number of relationships an individual considers adequate and a failure to achieve the desired level of intimacy (De Jong-Gierveld 1989).

Versions of the scale

The original De Jong Gierveld Loneliness Scale consists of 11 items. However, a shortened 6-item version was later introduced to accommodate settings that require a more efficient tool, such as large-scale surveys (De Jong-Gierveld 2006). Both the original and shortened versions include distinct social and emotional loneliness subscales, which can also be used independently, as discussed in the following section.

Dimensional structure

The scale was developed with Weiss’s distinction between social and emotional loneliness in mind. In the 11-item version, six items assess emotional loneliness, while five measure social loneliness. The shortened version maintains this dual structure, with three items dedicated to each dimension. Whilst these two sub-dimensions can be scored separately, researchers can opt for a unidimensional total loneliness score, when a single broad indicator of loneliness is sufficient for the research question.

Maes et al. (2022) reassessed the coding of the original 11-item scale, identifying six items as measuring emotional loneliness and four as measuring social loneliness, while one item (‘I experience a general sense of emptiness’) was not considered to measure any type of loneliness. However, their categorisation does not always align with the original classification. For instance, the original emotional loneliness subscale includes the item ‘I miss having people around,’ which Maes et al. (2022) instead classify as measuring social loneliness.

Wording of the items (direct vs. indirect, positive vs. negative)

None of the items in De Jong Gierveld scale explicitly mention loneliness or feeling lonely explicitly, making the scale a fully ‘indirect’ measure of loneliness.

Both the 11-item and the 6-item versions contain a mix of positively and negatively worded statements. Notably, for both the original and shortened versions of the scale, the emotional loneliness subscale consists exclusively of negatively worded items, whereas the social loneliness subscale includes only positively worded items.

Response format

Both versions of the scale typically employ a five-point response format – ‘Yes!’, ‘Yes’, ‘More or Less’, ‘No’, and ‘No!’ – or a simplified three-point format – just including ‘Yes’, ‘More or Less’, and ‘No’ – suggested in the 1999 manual, especially for face-to-face or telephone interviews.

Scoring method and categorisation of results

Responses to each item are dichotomised: answers indicating loneliness (e.g., ‘Yes!’, ‘Yes’, or ‘More or Less’ for negatively worded items, and ‘No’ or ‘No!’ for positively worded items) are scored as 1, while all other responses receive a score of 0. Therefore, in the 11-item version, total scores range from 0 to 11, while in the 6-item version, they range from 0 to 6.

According to the 1999 manual for the use of the scale, scores of 0-2 are classified as ‘not lonely,’ 3-8 as ‘quite lonely,’ and 9-11 as ‘severely lonely.’ No equivalent categorisation has been explicitly defined for the shortened version.

Frequency, duration and intensity of loneliness

Based on the wording of the items and the response format, the De Jong Gierveld scale is not designed to capture the frequency or duration of loneliness. Whether the scale effectively measures loneliness intensity is less clear-cut.

The scale’s most common response format (Yes / More or Less / No) primarily measures agreement with statements related to experiencing loneliness. While ‘More or less’ might hint at a milder or moderate experience of loneliness, there is no direct measure of how intense or severe the feeling is.

When an aggregated score is calculated from individual item responses, it can be interpreted as an indicator of loneliness intensity, as discussed in the previous section. However, the dichotomisation of responses limits this function. In the 11-item version, the total score ranges from 0 to 11, providing a relatively broad scale for assessing intensity. However, when, for instance, the social or emotional subscales of the shortened 6-item version are used independently, aggregated scores only range from 0 to 3, reducing the sensitivity of the scale in capturing different levels of loneliness intensity.

Population origin and applications

The scale was originally developed for older adults in the Netherlands. However, subsequent translations and validations have expanded its use. Considering all the studies in the MASLO database that used the De Jong Gierveld scale, the latter was mostly used with older adults (58.7%) and adults (36.6%), and rarely with adolescents (2.4%) and college students (2.4%). The scale has been translated into multiple languages and is used worldwide, though predominantly in Europe (78.5%), with fewer additional applications identified in North America, Asia, and Australia (Maes et al. 2022).

In the REA, the De Jong Gierveld scale was implemented mainly on the population of older adults (Roberts and Windle 2020; Boekhout et al. 2021; Ehsan et al. 2021; Hernández-Ascanio et al. 2023), with one study focusing on immigrants (Lai et al. 2020) and another on older adults with long-term conditions (Gilbody et al. 2021). Finally, one study focuses on a representative sample of the Australian population (Astell-Burt, Navakatikyan, and Feng 2024).

The Three Items Loneliness Scale (TILS)

Background and conceptual foundation

The Three-Item Loneliness Scale (TILS), also known as UCLA-3, was developed by Hughes et al. (2004) as a concise version of the UCLA Loneliness Scale. Grounded in the same conceptual framework as the UCLA-LS, the TILS was designed to address the need for a reliable and valid short measure suitable for large-scale population surveys where brevity is essential. These three items were selected based on their strong alignment with the original scale and their ability to effectively represent overall loneliness. The TILS is a reliable measure of loneliness that shows good internal consistency and strong validity, with correlations to broader loneliness scales and health-related outcomes (Hughes et al. 2004).

Initially developed by the authors specifically for use on a telephone survey, it has since been applied also in online assessments as well as person evaluations, as some of the studies in the REA sample demonstrate (Magid et al. 2024; Perkins, Spiro, and Waddell 2023, Mays et al. 2021; Cawthorne et al. 2023).

Versions of the scale

There is only one version of this scale used, which includes the following items:

1. ‘How often do you feel that you lack companionship?’

2. ‘How often do you feel left out?’; and

3. ‘How often do you feel isolated?’

In the UK, these three items are sometimes accompanied by a single direct measure asking, ‘How often do you feel lonely?’, as recommended by the Office for National Statistics (2018b).

Dimensional structure

The TILS is a unidimensional measure of loneliness and perceived social isolation. This scale does not assess different types of loneliness and rather provides a general assessment of it.

Wording of the items (direct vs. indirect, positive vs. negative)

This scale consists of three indirect statements about feelings of loneliness, which means respondents are not directly asked about their perceived loneliness. All three items are negatively worded. Some researchers (Perkins, Spiro, and Waddell 2023) have highlighted acquiescence bias as a potential limitation of this measure, as participants may be inclined to agree with the statements regardless of their true feelings.

Response format

This scale employs a three-point Likert-type response format: 1 (‘Hardly ever’), 2 (‘Some of the time’), and 3 (‘Often’). This simple format is aimed at ensuring accessibility for a wide variety of populations and facilitating rapid administration in research studies (Hughes et al. 2004).

Scoring method and categorisation of results

The scale is scored by summing responses across all three items, resulting in a total score ranging from 3 to 9. Higher scores indicate greater levels of perceived loneliness and social isolation.

Frequency, duration and intensity of loneliness

The TILS, like the UCLA-LS, primarily measures the frequency of loneliness experiences rather than their intensity or duration. The items focus on how often individuals feel they lack companionship, feel left out, or feel isolated. It does not explicitly assess the severity of loneliness or specify a particular time frame.

Population origin and applications

The scale was originally developed and validated in two different samples of older adults from two distinct studies: the Health and Retirement Study (HRS), a national, longitudinal study of individuals born in 1947 or earlier; and data from the first year of the Chicago Health, Aging, and Social Relations Study (CHASRS), a longitudinal, population-based study of individuals born between 1935 and 1952 (Hughes et al. 2004).

Since then, the TILS has been used to measure loneliness across a variety of populations, as demonstrated by the studies in the REA sample. These include adults (Mays et al. 2021; Bravata et al. 2023; Magid et al. 2024; Soh et al. 2024), children (Cawthorne et al. 2023), women suffering post-natal depression (Perkins, Spiro, and Waddell 2023), and older adults (Thomas, Akobundu, and Dosa 2016; Balta et al. 2023; Mierzwicki et al. 2024).

Other loneliness measures

The following sections provide a concise review of the key aspects of the three measures that were each used in only one study within the REA sample.

The Social and Emotional Loneliness Scale for Adults (SELSA)

Background and conceptual foundation

Developed by Di Tommaso and Spinner (1993), the SELSA was designed to measure loneliness as a multidimensional experience, distinguishing between social, romantic, and family loneliness. It is grounded in Weiss’s (1973) distinction between emotional and social loneliness, further refining emotional loneliness into romantic and familial dimensions.

Versions of the scale

The original SELSA consists of 37 items, while a short-form version (SELSA-S) with 15 items was later developed to provide a more efficient assessment (Di Tommaso, Brannen, and Best 2004)

Dimensional structure

SELSA assesses three types of loneliness: social, romantic, and family loneliness. The 37-item version consists of 14 social, 12 romantic, and 11 family loneliness items, while the short version (SELSA-S) includes five items per subscale. Maes et al. (2022) re-examined the scale and noted discrepancies in how items map onto different loneliness dimensions.

Wording of the items

Primarily indirectly worded, with one direct item asking about feeling alone when surrounded by family. Mix of positively and negatively worded statements to reduce response bias (23 positively framed, 14 negatively framed).

Response format

Uses a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Reverse-scored items are scored inversely to maintain consistency.

Scoring method and categorisation of results

Scores are summed within each subscale, allowing for independent assessment of social, romantic, and family loneliness. In the 37-item version, subscale scores range from 14-98 (social), 12-84 (romantic), and 11-77 (family), with an overall loneliness score from 37-25.

Frequency, duration, and intensity of loneliness

The SELSA does not measure frequency or duration but can indirectly capture the intensity of loneliness in different social domains.

Population origin and applications

Originally developed and validated in English with university students in Canada. Based on the MASLO database (Maes et al. 2022), it is primarily used with adults (49%) and university students (31%), but also applied to adolescents (17%) and older adults (3%).

The Loneliness and Social Dissatisfaction Questionnaire

Background and conceptual foundation

Also known as the Children’s Loneliness Scale (CLS), it was developed by Asher, Hymel, and Renshaw (1984) to assess children’s feelings of loneliness and social dissatisfaction with social relationships. It was created in response to the need for a psychometrically sound measure to capture subjective feelings of loneliness in school-age children.

Versions of the scale

There is only one version of the scale, consisting of 24 items. However, 16 of these specifically assess loneliness, while the remaining eight focus on children’s hobbies and are considered ‘filler’ items. These eight ‘filler’ items are included to help children feel more at ease when responding to the questionnaire.

Dimensional structure

This scale is unidimensional as it focuses on the overall feeling of loneliness and social dissatisfaction, with no subscales for different types of loneliness.

Wording of the items

Primarily indirectly worded, with two direct items stating, ‘I feel alone’ and ‘I’m lonely’. Mix of positively and negatively worded statements to reduce response bias (14 positively framed, 10 negatively framed).

Response format

Uses a 5-point scale ranging from 1 (always true) to 5 (not true at all). Reverse-scored items are those negatively framed, so they are scored inversely to maintain consistency.

Scoring method and categorisation of results

Higher scores indicate greater loneliness and social dissatisfaction, while lower scores suggest stronger peer relationships and social integration.

Frequency, duration, and intensity of loneliness

The scale captures frequency and agreement of feelings of loneliness and social dissatisfaction in children (Maes et al. 2022).

Population origin and applications

It was first developed and validated in a sample of children in grades 3-6 (ages 8-12) from two schools of a Midwestern city in the United States. Based on the MASLO database (Maes et al. 2022), it is primarily used with children (65.73%) and adolescents (34.09%).

The Campaign to End Loneliness Measurement Tool

Background and conceptual foundation

Developed in 2014 by the Campaign to End Loneliness, this tool was designed to assess loneliness in older adults, particularly those at risk of social isolation. It is based on the cognitive discrepancy theory of loneliness, which defines loneliness as the mismatch between desired and actual social connections (Campaign to End Loneliness 2015). The tool was developed through focus groups and workshops with older adults, service providers, and housing associations, followed by statistical validation.

Versions of the scale

Only one version of the scale exists, comprising three items: ‘I am content with my friendships and relationships’; ‘I have enough people I feel comfortable asking for help at any time’; and ‘My relationships are as satisfying as I would want them to be’.

Dimensional structure

The scale does not distinguish between different types of loneliness (e.g., social or emotional) but instead captures overall loneliness through three questions about relationship satisfaction, social support, and companionship.

Wording of the items

Measures loneliness indirectly, avoiding explicit references to loneliness. All items are positively framed to provide a non-intrusive approach and minimise distress for respondents.

Response format

Uses a five-point Likert scale, ranging from 0 (Strongly agree) to 4 (Strongly disagree).

Scoring method and categorisation of results

Responses are summed, producing a total score ranging from 0 to 12. A score of 0-3 suggests an individual is unlikely to experience loneliness, while 10-12 indicates high levels of loneliness. Scores in between represent varying levels of loneliness, though intervals are not evenly distributed in terms of perceived intensity.

Frequency, duration, and intensity of loneliness

The scale does not directly measure the frequency or duration of loneliness. The aggregated score may serve as a proxy for intensity.

Population origin and applications

Developed for older adults and designed as a tool for service providers to evaluate loneliness interventions, it’s primarily used in public health initiatives and interventions in the UK.

One-item measures

Eight of the studies in the REA included a single-item measure to assess loneliness. Four of these used it alongside other scales, such as the De Jong Gierveld, UCLA, and TILS scales, while the other four relied solely on the single-item question. Most of these measures employed direct wording, as is typically the case for one-item measures.

These one-item measures are often used in combination with other scales for various purposes. For example, Bouwman et al. (2017) employed a single-item measure to assess loneliness on a daily basis, while the primary loneliness scale was administered only four times throughout the study, allowing for a more frequent assessment of loneliness within a specific time frame (i.e., that day).

Appendix 4: Measures of similar topics

Warwick-Edinburgh Mental Well-Being Scale

The Warwick-Edinburgh Mental Well-being scale (WEMWBS) assesses feeling and functioning aspects of mental well-being in the general population (‘The Warwick-Edinburgh Mental Wellbeing Scales - WEMWBS’ 2025). It was developed to monitor mental well-being in the general population and for evaluating interventions aimed at improving mental well-being (ibid.). WEMWBS is available in a 14-item (WEMWBS) and 7-item (SWEMWBS) scale, both validated in clinical and non-clinical settings as well as a variety of geographical locations, languages and cultural contexts (ibid.). While WEMWBS has been primarily validated in adult populations, it has also been tested with specific groups such as young people (Clarke et al. 2011) and minority populations (Taggart et al. 2013).

Warwick-Edinburgh Mental Well-being scale: Factsheet

Attribute	Description
Domain	Mental well-being.
Number of items	14 (or 7) statements.
Response format	5-point scale from 1 (‘None of the time’) to 5 (‘All of the time’).
Scoring	The total score is a sum of item responses, with higher scores indicating greater mental well-being (total scores need transforming for the 7-item scale).
Time dimension	Asked to report experience ‘over the last 2 weeks’.

ONS4

The Office for National Statistics (ONS) developed ONS4 to measure three types of personal well-being: evaluative, eudemonic and affective (Office for National Statistics 2025). ONS4 includes four subjective well-being measures concerning life satisfaction, happiness, worthwhileness, and anxiety (Child Outcomes Research Consortium, n.d.-a). These questions should not be combined and can be used individually or alongside additional non-ONS4 measures (What Works Wellbeing 2020). ONS4 was primarily developed in English for use by adults (16+) in the general population, however it has also been used by children and young people both in national surveys and evaluation settings (Office for National Statistics 2025).

ONS4: Factsheet

Attribute	Description
Domain	Personal well-being.
Number of items	4 statements.
Response format	11-point scale from 0 (‘Not at all’) to 10 (‘Completely’).
Scoring	Scores of 0-4 in positively worded questions and 6-10 in negatively worded questions indicate low well-being.
Time dimension	Asked to report experience ‘yesterday’.

World Health Organization-Five Well-Being Index

The World Health Organization-Five Well-Being Index (WHO-5) is a self-reported measure of subjective mental well-being. It includes five statements concerning positive mood, vitality, and general interest (World Health Organization 2024). The measure was first developed in Europe in 1998 and has now been translated and tested in over 30 languages (Child Outcomes Research Consortium, n.d.-b). The WHO-5 measure is suitable for use on adults, older adults, and children aged 9 and above (Topp et al. 2015). It can be used both as a screening tool and for measuring outcomes in interventions.

World Health Organization-Five Well-Being Index: Factsheet

Attribute	Description
Domain	Mental well-being.
Number of items	5 statements.
Response format	6-point scale from 5 (‘All the time’) to 0 (‘At no time’).
Scoring	Total raw scores are the sum of each response ranging from 0 to 25, with higher scores indicating greater mental well-being.
Time dimension	Asked to report experience ‘over the last two weeks’.

General Health Questionnaire

The General Health Questionnaire (GHQ) is a screening tool to identify minor psychiatric conditions. There are four versions available with 12 (GHQ-12), 28 (GHQ-28), 30 (GHQ-30), and 60 (original) items. Questions assess the respondent’s current state and whether it differs from ‘normal’, focusing on aspects such as somatic symptoms, anxiety and sleeplessness, social dysfunction, and depression (Jackson 2007). It has demonstrated validity for the general population aged 17 and above (Tait, Hulse, and Robertson 2002). GHQ-12 is the most extensively used version to screen for common mental disorders and psychological well-being due to its brevity (del Pilar Sánchez-López and Dresch 2008).

General Health Questionnaire: Factsheet

Attribute	Description
Domain	Psychological well-being.
Number of items	Four versions available, consisting of 12 (GHQ-12), 28 (GHQ-28), 30 (GHQ-30), or 60 (original) statements.
Response format	4-point scale from 1 (‘Better than usual’) to 4 (‘Much worse than usual’).
Scoring	Likert (0-1-2-3) or binary (0-0-1-1) scoring system. Higher scores indicate a greater possibility of psychological distress.
Time dimension	Asked to report how you have been feeling ‘recently’.

WHOQOL-BREF

The WHOQOL-BREF is an abbreviated 26-item quality of life assessment, adapted from the longer WHOQOL-100. It measures both subjective well-being and functional status across four health domains: physical health, psychological health, social relationships, and environmental health (Vahedi 2010). WHOQOL-BREF can be used in a variety of settings, including research, policymaking, and evaluation, and has been validated different cultural contexts, languages and sub-groups (World Health Organization 2012).

WHOQOL-BREF: Factsheet

Attribute	Description
Domain	Quality of life.
Number of items	26 statements.
Response format	5-point scale from 1 (‘Never’ or similar) to 5 (‘Always’ or similar).
Scoring	The raw scores for each domain are transformed into a score from 0 to 100, with higher scores indicating better quality of life.
Time dimension	Asked to report experience ‘over the last two weeks’ (but can be adapted to suit uses e.g. for chronic conditions).

Psychological Well-being Scale

Ryff’s Psychological Well-being (PWB) scale was developed as a multidimensional measure of well-being in the general population, assessing six key dimensions: autonomy, environmental mastery, personal growth, positive relationships, purpose in life, and self-acceptance (Ryff 2013). It is available in multiple versions, including 18-item, 42-item, 54-item and 84-item formats which combine positively and negatively worded questions. While the longer versions are considered more statistically robust, the abbreviated 18-item version is most commonly used in intervention and research settings for its brevity (Stanford University, n.d.).

Psychological Well-being Scale: Factsheet

Attribute	Description
Domain	Psychological well-being.
Number of items	84, 54, 42 or 18 statements.
Response format	6-point scale from 1 (‘Strongly agree’) to 6 (‘Strongly disagree’).
Scoring	Negatively worded items are reverse-scored and all values are summed, with higher scores indicating greater wellbeing.
Time dimension	No time dimension mentioned.

Mental Health Continuum Short Form

The Mental Health Continuum Short Form (MHC-SF) was developed from the long form version (MHC-LF) and measures emotional, psychological and social well-being. The MHC-SF consists of 14-items, chosen as the most prototypical items representing each dimension of well-being from the long form’s original 40-items (Keyes 2022). The short form’s response format measures the frequency with which respondents experience each symptom of mental health. Studies have shown that the MHC-SF is valid for both adolescents and adults for use in multiple cultural contexts and in over eight languages.

Mental Health Continuum Short Form: Factsheet

Attribute	Description
Domain	Emotional, psychological and social well-being.
Number of items	14 statements.
Response format	6-point scale from 0 (‘Never’) to 5 (‘Every day’).
Scoring	Scores are summed to produce a continuous score (reflecting overall degree of well-being) or a categorical score (reflecting whether individuals’ mental health is ‘moderate’, ‘flourishing’ or ‘languishing’). Higher scores indicate better well-being.
Time dimension	Asked about frequency of experience, from ‘never’ to ‘every day’.

Personal Well-being Index

The Personal Well-being Index (PWI) is a 9-item measure, adapted from the Comprehensive Quality of Life Scale (ComQol). It assesses satisfaction with seven life domains, namely: standard of living, health, achieving in life, relationships, safety, community connectedness, and future security (International Wellbeing Group 2024). Versions of the PWI have been developed for adults, young people aged over 12 years old, and people with intellectual or cognitive disabilities, and have been translated for use into multiple languages (Australian Centre on Quality of Life, n.d.).

Personal Well-being Index: Factsheet

Attribute	Description
Domain	Subjective well-being.
Number of items	7 core and 2 optional statements.
Response format	11-point scale, from 0 (‘No satisfaction’) to 10 ‘Complete satisfaction’).
Scoring	Scores are presented as a single score, with higher scores indicating higher levels of personal well-being.
Time dimension	No time dimension mentioned.

Rosenberg Self-Esteem Scale

The Rosenberg Self-Esteem Scale (RSES) is a 10-item measure for assessing self-esteem. It uses a combination of positively and negatively worded items to evaluate two facets of self-esteem: self-competence and self-liking (Rosenberg 1965). The measure is widely used to measure self-esteem in adults and adolescents across multiple clinical, research and intervention settings, and has been validated in multiple languages and cultural contexts.

Rosenberg Self-Esteem Scale: Factsheet

Attribute	Description
Domain	Self-esteem.
Number of items	10 statements.
Response format	4-point scale, from 0 (‘Strongly disagree’) to 3 (‘Strongly agree’).
Scoring	Negatively worded items are reverse scored and all values are summed, with higher scores representing higher self-esteem.
Time dimension	No time dimension mentioned.

Acknowledgements

We would like to acknowledge and thank the following people and organisation for their valuable contributions to this research:

• Andrew Steptoe (University College London)
• Christina Victor (Brunel University London)
• David McDaid (London School of Economics)
• Helayna Jenkins (London Borough of Bromley)
• Jennifer Lau (Queen Mary University)
• Joelle Bradly (National Academy for Social Prescribing)
• John Ratcliffe (University of Leeds)
• Julie Barnett (University of Bath)
• Keming Yang (Durham University)
• Kirsty Veitch-Sorsby (Bassetlaw Community and Voluntary Service)
• Linda Monckton (Historic England)
• Louise Arseneault (King’s College London)
• Louise Mansfield (Brunel University London)
• Manuela Barreto (University of Exeter)
• Peter Kinderman (University of Liverpool)
• Sarah Jewell (Age UK)
• Silia Vitoratou (King’s College London)

Bibliographic references

Adams, Rebecca N., Catherine E. Mosher, Kevin L. Rand, Adam T. Hirsh, Patrick O. Monahan, Rafat Abonour, and Kurt Kroenke. 2017. ‘The Cancer Loneliness Scale and Cancer-Related Negative Social Expectations Scale: Development and Validation’. Quality of Life Research 26 (7): 1901–13.

Alsubheen, Sanaa A., Ana Oliveira, Razanne Habash, Roger Goldstein, and Dina Brooks. 2023. ‘Systematic Review of Psychometric Properties and Cross-Cultural Adaptation of the University of California and Los Angeles Loneliness Scale in Adults’. Current Psychology 42 (14): 11819–33. https://doi.org/10.1007/s12144-021-02494-w.

Applebaum, Jeremy, Matthew Hoyer, Joshua Betz, Frank R. Lin, and Adele M. Goman. 2019. ‘Long-Term Subjective Loneliness in Adults after Hearing Loss Treatment’. International Journal of Audiology 58 (8): 464–67. https://doi.org/10.1080/14992027.2019.1593523.

Asher, Steven R., Shelley Hymel, and Peter D. Renshaw. 1984. ‘Loneliness in Children’. Child Development 55 (4): 1456. https://doi.org/10.2307/1130015.

Astell-Burt, Thomas, Michael A. Navakatikyan, and Xiaoqi Feng. 2024. ‘Contact with Nature May Be a Remedy for Loneliness: A Nationally Representative Longitudinal Cohort Study’. Environmental Research 263 (December):120016. https://doi.org/10.1016/j.envres.2024.120016.

Australian Centre on Quality of Life. n.d. ‘Personal Wellbeing Index’. Measures. Accessed 17 March 2025. https://www.acqol.com.au/instruments.

Balta, Marianna, Konstantinos Katsas, Chrysoula Grigoropoulou, Dimitrios V. Diamantis, Dimitrios Kalogiannis, Nikolaos Drougos, Eleni Fagogeni, Afroditi Veloudaki, Demosthenes Panagiotakos, and Athena Linos. 2023. ‘Combating Loneliness in Older Adults during the COVID-19 Pandemic: Findings from a Volunteer-Based Program in Greece’. Behavioral Sciences (Basel, Switzerland) 13 (10): 804. https://doi.org/10.3390/bs13100804.

Berger, Ruben, Rene Hurlemann, Simone Shamay-Tsoory, Alisa Kanterman, Maura Brauser, Jessica Gorni, Maike Luhmann, et al. 2024. ‘Oxytocin-Augmented Modular-Based Group Intervention for Loneliness: A Proof-Of-Concept Randomized Controlled Trial’. Psychotherapy and Psychosomatics 93 (3): 169–80. https://doi.org/10.1159/000538752.

Boekhout, Janet M., Esmee Volders, Catherine A. W. Bolman, Renate H. M. de Groot, and Lilian Lechner. 2021. ‘Long-Term Effects on Loneliness of a Computer-Tailored Intervention for Older Adults With Chronic Diseases: A Randomized Controlled Trial’. Journal of Aging and Health 33 (10): 865–76. https://doi.org/10.1177/08982643211015027.

Bouwman, Tamara E., Marja J. Aartsen, Theo G. van Tilburg, and Nan L. Stevens. 2017. ‘Does Stimulating Various Coping Strategies Alleviate Loneliness? Results from an Online Friendship Enrichment Program’. Journal of Social and Personal Relationships 34 (6): 793–811. https://doi.org/10.1177/0265407516659158.

Bravata, Dena M, Joseph Kim, Daniel W Russell, Ron Goldman, and Elizabeth Pace. 2023. ‘Digitally Enabled Peer Support Intervention to Address Loneliness and Mental Health: Prospective Cohort Analysis’. JMIR Formative Research 7 (January). https://doi.org/10.2196/48864.

Bruehlman-Senecal, E, CJ Hook, JH Pfeifer, C FitzGerald, B Davis, KL Delucchi, J Haritatos, and DE Ramo. 2020. ‘Smartphone App to Address Loneliness among College Students: Pilot Randomized Controlled Trial’. JMIR Mental Health 7 (10): e21496. https://doi.org/10.2196/21496.

Cacioppo, John T., and Louise C. Hawkley. 2009. ‘Perceived Social Isolation and Cognition’. Trends in Cognitive Sciences 13 (10): 447–54. https://doi.org/10.1016/j.tics.2009.06.005.

Calderon Leon, Maria D., João F. Guassi Moreira, Natalie M. Saragosa-Harris, Yael H. Waizman, Anna Sedykin, Tara S. Peris, and Jennifer A. Silvers. 2024. ‘Parent and Friend Relationship Quality and Links to Trajectories of Loneliness During the First Year of College’. Child Psychiatry & Human Development 55 (3): 680–94. https://doi.org/10.1007/s10578-022-01416-6.

Campaign to End Loneliness. 2015. ‘Measuring Your Impact on Loneliness in Later Life’.

Cawthorne, Tom, Anton Käll, Sophie Bennett, Elena Baker, Gerhard Andersson, and Roz Shafran. 2023. ‘The Development and Preliminary Evaluation of Cognitive Behavioural Therapy (CBT) for Chronic Loneliness in Young People’. Behavioural and Cognitive Psychotherapy 51 (5): 414–31. https://doi.org/10.1017/S1352465823000231.

Child Outcomes Research Consortium. n.d.-a. ‘Office of National Statistics Personal Wellbeing Domain for Children & Young People’. Child Outcomes Research Consortium. Accessed 17 March 2025. https://www.corc.uk.net/outcome-experience-measures/directory-of-outcome-measures/office-of-national-statistics-personal-wellbeing-domain-for-children-young-people/.

———. n.d.-b. ‘The World Health Organisation- Five Well-Being Index (WHO-5)’. Child Outcomes Research Consortium. Accessed 17 March 2025. https://www.corc.uk.net/outcome-experience-measures/directory-of-outcome-measures/the-world-health-organisation-five-well-being-index-who-5/.

Clarke, Aileen, Tim Friede, Rebecca Pultz, Jacquie Ashdown, Steven Martin, Amy Blake, Yaser Adi, et al. 2011. ‘Warwick-Edinburgh Mental Well-Being Scale (WEMWBS): Validated for Teenage School Students in England and Scotland. A Mixed Methods Assessment’. BMC Public Health 11 (487). https://doi.org/10.1186/1471-2458-11-487.

Coll-Planas, Laura, Gabriela Del Valle Gómez, Petra Bonilla, Teresa Masat, Teresa Puig, and Rosa Monteserin. 2017. ‘Promoting Social Capital to Alleviate Loneliness and Improve Health among Older People in Spain’. Health & Social Care in the Community 25 (1): 145–57. https://doi.org/10.1111/hsc.12284.

De Jong Gierveld, Jenny. 1987. ‘Developing and Testing a Model of Loneliness’. Journal of Personality and Social Psychology 53 (1): 119–28. https://doi.org/10.1037/0022-3514.53.1.119.

De Jong-Gierveld, Jenny. 1989. ‘Personal Relationships, Social Support, and Loneliness’. Journal of Social and Personal Relationships 6 (2): 197–221.

———. 2006. ‘A 6-Item Scale for Overall, Emotional, and Social Loneliness: Confirmatory Tests on Survey Data’. Research on Aging 28 (5): 582–98.

De Jong-Gierveld, Jenny, and Frans Kamphuls. 1985. ‘The Development of a Rasch-Type Loneliness Scale’. Applied Psychological Measurement 9 (3): 289–99. https://doi.org/10.1177/014662168500900307.

De Jong-Gierveld, Jenny, and Theo Van Tilburg. 1999. ‘Manual of the Loneliness Scale’.

Deol, Ekamjit Singh, K. Yamashita, S. Elliott, T. K. Malmstorm, and J. E. Morley. 2022. ‘Validation of the ALONE Scale: A Clinical Measure of Loneliness’. The Journal of Nutrition, Health & Aging 26 (5): 421–24. https://doi.org/10.1007/s12603-022-1794-8.

Department for Digital, Culture, Media and Sport. 2018. ‘A Connected Society: A Strategy for Tackling Loneliness – Laying the Foundations for Change’, October. https://assets.publishing.service.gov.uk/media/5fb66cf98fa8f54aafb3c333/6.4882_DCMS_Loneliness_Strategy_web_Update_V2.pdf.

Dingle, Genevieve A., Leah S. Sharman, Shaun Hayes, Catherine Haslam, Tegan Cruwys, Jolanda Jetten, S. Alexander Haslam, et al. 2024. ‘A Controlled Evaluation of Social Prescribing on Loneliness for Adults in Queensland: 8-Week Outcomes’. Frontiers in Psychology 15:1359855. https://doi.org/10.3389/fpsyg.2024.1359855.

DiTommaso, Enrico, Cyndi Brannen, and Lisa A. Best. 2004. ‘Measurement and Validity Characteristics of the Short Version of the Social an Emotional Loneliness Scale for Adults’. Educational and Psychological Measurement 64 (1): 99–119. https://doi.org/10.1177/0013164403258450.

DiTommaso, Enrico, and Barry Spinner. 1993. ‘The Development and Initial Validation of the Social and Emotional Loneliness Scale for Adults (SELSA)’. Personality and Individual Differences 14 (1): 127–34. https://doi.org/10.1016/0191-8869(93)90182-3.

Ehsan, Annahita, Danilo Bolano, Sylvie Guillaume-Boeckle, and Dario Spini. 2021. ‘Reducing Loneliness in Older Adults: Looking at Locals and Migrants in a Swiss Case Study’. European Journal of Ageing 18 (3): 333–44. https://doi.org/10.1007/s10433-020-00577-4.

Ellaway, A., S. Wood, and S. Macintyre. 1999. ‘Someone to Talk to? The Role of Loneliness as a Factor in the Frequency of GP Consultations’. The British Journal of General Practice: The Journal of the Royal College of General Practitioners 49 (442): 363–67.

Faustino, Bruno, Paulo Lopes, Jorge Oliveira, Giulia Campaioli, Maria Rondinone, Helena Bomfim, and Lindanuza Germano. 2019. ‘Psychometric and Rash Analysis of the UCLA Loneliness Scale-16 in a Portuguese Sample of Older Adults’. Psychological Studies 64 (2): 140–46. https://doi.org/10.1007/s12646-019-00483-5.

Follmann, Andreas, Franziska Schollemann, Andrea Arnolds, Pauline Weismann, Thea Laurentius, Rolf Rossaint, and Michael Czaplik. 2021. ‘Reducing Loneliness in Stationary Geriatric Care with Robots and Virtual Encounters-A Contribution to the COVID-19 Pandemic’. International Journal of Environmental Research and Public Health 18 (9): 4846. https://doi.org/10.3390/ijerph18094846.

Fong, Polly, Tegan Cruwys, Sam L. Robinson, S. Alexander Haslam, Catherine Haslam, Paula L. Mance, and Claire L. Fisher. 2021. ‘Evidence That Loneliness Can Be Reduced by a Whole-of-Community Intervention to Increase Neighbourhood Identification’. Social Science & Medicine (1982) 277 (May):113909. https://doi.org/10.1016/j.socscimed.2021.113909.

Franke, Thea, Joanie Sims-Gould, Lindsay Nettlefold, Callista Ottoni, and Heather A. McKay. 2021. ‘’It Makes Me Feel Not so Alone’: Features of the Choose to Move Physical Activity Intervention That Reduce Loneliness in Older Adults’. BMC Public Health 21 (1): 312. https://doi.org/10.1186/s12889-021-10363-1.

Geller, J., P. Janson, E. McGovern, and A. Valdini. 1999. ‘Loneliness as a Predictor of Hospital Emergency Department Use’. The Journal of Family Practice 48 (10): 801–4.

Gilbody, Simon, Elizabeth Littlewood, Dean McMillan, Carolyn A. Chew-Graham, Della Bailey, Samantha Gascoyne, Claire Sloan, et al. 2021. ‘Behavioural Activation to Prevent Depression and Loneliness among Socially Isolated Older People with Long-Term Conditions: The BASIL COVID-19 Pilot Randomised Controlled Trial’. PLoS Medicine 18 (10): e1003779. https://doi.org/10.1371/journal.pmed.1003779.

Gould, Christine E., Chalise Carlson, Ana Jessica Alfaro, Christina F. Chick, Martha L. Bruce, and Valerie L. Forman-Hoffman. 2021. ‘Changes in Quality of Life and Loneliness Among Middle-Aged and Older Adults Participating in Therapist-Guided Digital Mental Health Intervention’. Frontiers in Public Health 9:746904. https://doi.org/10.3389/fpubh.2021.746904.

Hays, Ron, and M. Robin DiMatteo. 1987. ‘A Short-Form Measure of Loneliness’. Journal of Personality Assessment 51 (1): 69–81. https://doi.org/10.1207/s15327752jpa5101_6.

Hernández-Ascanio, J, LÁ Perula-de Torres, M Rich-Ruiz, J González-Santos, J Mielgo-Ayuso, and J González-Bernal. 2023. ‘Effectiveness of a Multicomponent Intervention to Reduce Social Isolation and Loneliness in Community-Dwelling Elders: A Randomized Clinical Trial’. Nursing Open 10 (1): 48‐60. https://doi.org/10.1002/nop2.1277.

Holt-Lunstad, Julianne, Timothy B. Smith, Mark Baker, Tyler Harris, and David Stephenson. 2015. ‘Loneliness and Social Isolation as Risk Factors for Mortality: A Meta-Analytic Review’. Perspectives on Psychological Science 10 (2): 227–37.

Hughes, Mary Elizabeth, Linda J. Waite, Louise C. Hawkley, and John T. Cacioppo. 2004. ‘A Short Scale for Measuring Loneliness in Large Surveys: Results From Two Population-Based Studies’. Research on Aging 26 (6): 655–72. https://doi.org/10.1177/0164027504268574.

International Wellbeing Group. 2024. ‘Personal Wellbeing Index Manual’. International Wellbeing Group. https://www.acqol.com.au/uploads/pwi-a/pwi-a-english.pdf.

Iyer, P, L Iyer, N Carter, R Iyer, A Stirling, L Priya, and U Sriraman. 2024. ‘Self-Care Program as a Tool for Alleviating Anxiety and Loneliness and Promoting Satisfaction With Life in High School Students and Staff: Randomized Survey Study’. JMIR Formative Research 8:e56355. https://doi.org/10.2196/56355.

Iyer, Ranjani B., Sumedha Vadlapudi, Lina Iyer, Vetriliaa Kumar, Laya Iyer, Priya Sriram, Roma Tandon, et al. 2023. ‘Impact of the Heartfulness Program on Loneliness in High Schoolers: Randomized Survey Study’. Applied Psychology. Health and Well-Being 15 (1): 66–79. https://doi.org/10.1111/aphw.12360.

Jackson, Craig. 2007. ‘The General Health Questionnaire’. Occupational Medicine 57 (1): 79. https://doi.org/10.1093/occmed/kql169.

Jesus, Andreia G., Margarida Pedroso de Lima, Manuela Vilar, and Nancy A. Pachana. 2024. ‘Homebound Elderly People Psychotherapeutic Intervention (HEPPI): Exploring the Mediating Role of Loneliness’. Archives of Gerontology and Geriatrics 118 (March):105308. https://doi.org/10.1016/j.archger.2023.105308.

Kahlon, Maninder K., Nazan Aksan, Rhonda Aubrey, Nicole Clark, Maria Cowley-Morillo, Elizabeth A. Jacobs, Rhonda Mundhenk, Katherine R. Sebastian, and Steven Tomlinson. 2021. ‘Effect of Layperson-Delivered, Empathy-Focused Program of Telephone Calls on Loneliness, Depression, and Anxiety Among Adults During the COVID-19 Pandemic: A Randomized Clinical Trial’. JAMA Psychiatry 78 (6): 616–22. https://doi.org/10.1001/jamapsychiatry.2021.0113.

Käll, A, M Bäck, C Welin, H Åman, R Bjerkander, M Wänman, T Lindegaard, et al. 2021. ‘Therapist-Guided Internet-Based Treatments for Loneliness: A Randomized Controlled Three-Arm Trial Comparing Cognitive Behavioral Therapy and Interpersonal Psychotherapy’. Psychotherapy and Psychosomatics 90 (5): 351‐358. https://doi.org/10.1159/000516989.

Käll, Anton, Ulrika Backlund, Roz Shafran, and Gerhard Andersson. 2020. ‘Lonesome No More? A Two-Year Follow-up of Internet-Administered Cognitive Behavioral Therapy for Loneliness’. Internet Interventions 19 (March):100301. https://doi.org/10.1016/j.invent.2019.100301.

Keyes, Corey L M. 2022. ‘The Mental Health Continuum: From Languishing to Flourishing in Life’. J Health Soc Behav 43 (2): 207–22.

Kramer, Lean L., Lex van Velsen, Jenna L. Clark, Bob C. Mulder, and Emely de Vet. 2022. ‘Use and Effect of Embodied Conversational Agents for Improving Eating Behavior and Decreasing Loneliness Among Community-Dwelling Older Adults: Randomized Controlled Trial’. JMIR Formative Research 6 (4): e33974. https://doi.org/10.2196/33974.

Lai, Daniel W. L., Jia Li, Xiaoting Ou, and Celia Y. P. Li. 2020. ‘Effectiveness of a Peer-Based Intervention on Loneliness and Social Isolation of Older Chinese Immigrants in Canada: A Randomized Controlled Trial’. BMC Geriatrics 20 (1): 356. https://doi.org/10.1186/s12877-020-01756-9.

Lim, Michelle H., Alexandra Hennessey, Pamela Qualter, Ben J. Smith, Lily Thurston, Robert Eres, and Julianne Holt-Lunstad. 2024. ‘The KIND Challenge Community Intervention to Reduce Loneliness and Social Isolation, Improve Mental Health, and Neighbourhood Relationships: An International Randomized Controlled Trial’. Social Psychiatry and Psychiatric Epidemiology, August. https://doi.org/10.1007/s00127-024-02740-z.

Lindsay, Emily K., Shinzen Young, Kirk Warren Brown, Joshua M. Smyth, and J. David Creswell. 2019. ‘Mindfulness Training Reduces Loneliness and Increases Social Contact in a Randomized Controlled Trial’. Proceedings of the National Academy of Sciences of the United States of America 116 (9): 3488–93.

Lippke, Sonia, Tiara Ratz, Franziska Maria Keller, Dennis Juljugin, Manuela Peters, Claudia Pischke, and Claudia Voelcker-Rehage. 2022. ‘Mitigating Feelings of Loneliness and Depression by Means of Web-Based or Print-Based Physical Activity Interventions: Pooled Analysis of 2 Community-Based Intervention Trials’. JMIR Aging 5 (3): e36515. https://doi.org/10.2196/36515.

Liu, Shuyan, Luisa Wegner, Matthias Haucke, Jennifer Gates, Maya Adam, and Till Bärnighausen. 2023. ‘An Entertainment-Education Video and Written Messages to Alleviate Loneliness in Germany: Pilot Randomized Controlled Study’. JMIR Formative Research 7 (January). https://doi.org/10.2196/43036.

MacIntyre, Helen, and Robin Hewings. 2023. ‘Evaluation of Interventions to Tackle Loneliness’. GOV.UK. 2023. https://www.gov.uk/government/publications/exploring-interventions-to-tackle-loneliness/evaluation-of-interventions-to-tackle-loneliness.

Maes, Marlies, Pamela Qualter, Gerine M. A. Lodder, and Marcus Mund. 2022. ‘How (Not) to Measure Loneliness: A Review of the Eight Most Commonly Used Scales’. International Journal of Environmental Research and Public Health 19 (17): 10816. https://doi.org/10.3390/ijerph191710816.

Magid, Kirby, Sara J. Sagui-Henson, Cynthia Castro Sweet, Brooke J. Smith, Camille E. Welcome Chamberlain, and Sara M. Levens. 2024. ‘The Impact of Digital Mental Health Services on Loneliness and Mental Health: Results from a Prospective, Observational Study’. International Journal of Behavioral Medicine 31 (3): 468–78. https://doi.org/10.1007/s12529-023-10204-y.

Mayers, Aviva M., Siek-Toon Khoo, and Martin Svartberg. 2002. ‘The Existential Loneliness Questionnaire: Background, Development, and Preliminary Findings’. Journal of Clinical Psychology 58 (9): 1183–93. https://doi.org/10.1002/jclp.10038.

Mays, Allison Moser, Sungjin Kim, Katrina Rosales, Tam Au, and Sonja Rosen. 2021. ‘The Leveraging Exercise to Age in Place (LEAP) Study: Engaging Older Adults in Community-Based Exercise Classes to Impact Loneliness and Social Isolation’. The American Journal of Geriatric Psychiatry: Official Journal of the American Association for Geriatric Psychiatry 29 (8): 777–88. https://doi.org/10.1016/j.jagp.2020.10.006.

Mierzwicki, Justin T, Joshua Kline, Brooke Schach, and Lauren Vandenberg. 2024. ‘Loneliness Improved by Either Telephone or Video Communications in Community-Dwelling Older Adults’. Archives of Gerontology and Geriatrics Plus 1 (2): 100011. https://doi.org/10.1016/j.aggp.2024.100011.

Molloy, Gerard J., Hannah M. McGee, Desmond O’Neill, and Ronan M. Conroy. 2010. ‘Loneliness and Emergency and Planned Hospitalizations in a Community Sample of Older Adults’. Journal of the American Geriatrics Society 58 (8): 1538–41. https://doi.org/10.1111/j.1532-5415.2010.02960.x.

Neil-Sztramko, Sarah E., Giulia Coletta, Maureen Dobbins, and Sharon Marr. 2020. ‘Impact of the AGE-ON Tablet Training Program on Social Isolation, Loneliness, and Attitudes Toward Technology in Older Adults: Single-Group Pre-Post Study’. JMIR Aging 3 (1): e18398. https://doi.org/10.2196/18398.

Neto, Félix. 2014. ‘Psychometric Analysis of the Short-Form UCLA Loneliness Scale (ULS-6) in Older Adults’. European Journal of Ageing 11 (4): 313–19. https://doi.org/10.1007/s10433-014-0312-1.

Office for National Statistics. 2018a. ‘National Measurement of Loneliness: 2018’.

———. 2018b. ‘Recommended National Indicators of Loneliness’. 5 December 2018. https://www.ons.gov.uk/peoplepopulationandcommunity/wellbeing/compendium/nationalmeasurementofloneliness/2018/recommendednationalindicatorsofloneliness.

———. 2025. ‘Personal Well-Being User Guidance’. 14 January 2025. https://www.ons.gov.uk/peoplepopulationandcommunity/wellbeing/methodologies/personalwellbeingsurveyuserguide.

Parenteau, China I., Jessica Floyd, Katy Ankenman, Tara Glavin, Julia Charalel, Enjey Lin, Whitney Ence, Young Shin Kim, Somer Bishop, and Shuting Zheng. 2024. ‘Efficacy of Community-Delivered PEERS® for Adolescents: Increases in Social Skills and Decreases in Social Anxiety and Loneliness’. Journal of Autism and Developmental Disorders, June. https://doi.org/10.1007/s10803-024-06433-z.

Patapoff, Molly A., Dylan J. Jester, Rebecca E. Daly, Brent T. Mausbach, Colin A. Depp, and Danielle K. Glorioso. 2024. ‘Remotely-Administered Resilience and Self-Compassion Intervention Targeting Loneliness and Stress in Older Adults: A Single-Case Experimental Design’. Aging & Mental Health 28 (2): 369–76. https://doi.org/10.1080/13607863.2023.2262411.

Penning, Margaret J., Guiping Liu, and Pak Hei Benedito Chou. 2014. ‘Measuring Loneliness Among Middle-Aged and Older Adults: The UCLA and de Jong Gierveld Loneliness Scales’. Social Indicators Research 118 (3): 1147–66.

Perkins, R., N. Spiro, and G. Waddell. 2023. ‘Online Songwriting Reduces Loneliness and Postnatal Depression and Enhances Social Connectedness in Women with Young Babies: Randomised Controlled Trial’. Public Health 220 (July):72–79. https://doi.org/10.1016/j.puhe.2023.04.017.

Perlman, Daniel, and Letitia Anne Peplau. 1981. ‘Toward a Social Psychology of Loneliness’. In R. Gilmour, & S. Duck (Eds.), Personal Relationships: 3. Relationships in Disorder, 31–56.

Pilar Sánchez-López, María del, and Virginia Dresch. 2008. ‘The 12-Item General Health Questionnaire (GHQ-12): Reliability, External Validity and Factor Structure in the Spanish Population’. Psicothema 20 (4): 839–43.

Richards, T. Anne, Doug Oman, Angela DiMartino, and Ashwin A. Kotwal. 2024. ‘A Technology-Based Intervention Impacts Quality of Life for Low-Income Older Adults by Reducing Loneliness and Improving Healthcare Self-Efficacy and Self-Rated Health’. Journal of Applied Gerontology: The Official Journal of the Southern Gerontological Society, October, 7334648241289928. https://doi.org/10.1177/07334648241289928.

Roberts, J. R., and G. Windle. 2020. ‘Evaluation of an Intervention Targeting Loneliness and Isolation for Older People in North Wales’. Perspectives in Public Health 140 (3): 153–61. https://doi.org/10.1177/1757913919868752.

Rosenberg, M. 1965. ‘Rosenberg Self-Esteem Scale’. Acceptance and Commitment Therapy, 61–62. https://doi.org/10.1037/t01038-000.

Routasalo, PE, RS Tilvis, H Kautiainen, and KH Pitkala. 2009. ‘Effects of Psychosocial Group Rehabilitation on Social Functioning, Loneliness and Well-Being of Lonely, Older People: Randomized Controlled Trial’. Journal of Advanced Nursing 65 (2): 297‐305. https://doi.org/10.1111/j.1365-2648.2008.04837.x.

Russell, D. W., C. E. Cutrona, A. de la Mora, and R. B. Wallace. 1997. ‘Loneliness and Nursing Home Admission among Rural Older Adults’. Psychology and Aging 12 (4): 574–89. https://doi.org/10.1037//0882-7974.12.4.574.

Russell, Dan, Letitia A. Peplau, and Carolyn E. Cutrona. 1980. ‘The Revised UCLA Loneliness Scale: Concurrent and Discriminant Validity Evidence.’ Journal of Personality and Social Psychology 39 (3): 472–80. https://doi.org/10.1037/0022-3514.39.3.472.

Russell, Dan, Letitia Anne Peplau, and Mary Lund Ferguson. 1978. ‘Developing a Measure of Loneliness’. Journal of Personality Assessment 42 (3): 290–94. https://doi.org/10.1207/s15327752jpa4203_11.

Russell, Daniel W. 1996. ‘UCLA Loneliness Scale (Version 3): Reliability, Validity, and Factor Structure’. Journal of Personality Assessment 66 (1): 20–40. https://doi.org/10.1207/s15327752jpa6601_2.

Ryff, Carol D. 2013. ‘Psychological Well-Being Revisited: Advances in Science and Practice’. Psychother Psychosom 83 (1): 10–28. https://doi.org/10.1159/000353263.

Shankar, Aparna, Anne McMunn, James Banks, and Andrew Steptoe. 2011. ‘Loneliness, Social Isolation, and Behavioral and Biological Health Indicators in Older Adults’. Health Psychology: Official Journal of the Division of Health Psychology, American Psychological Association 30 (4): 377–85. https://doi.org/10.1037/a0022826.

Sheridan, Ann J., Jonathan Drennan, Barbara Coughlan, Donal O’Keeffe, Kate Frazer, Mary Kemple, Denise Alexander, et al. 2015. ‘Improving Social Functioning and Reducing Social Isolation and Loneliness among People with Enduring Mental Illness: Report of a Randomised Controlled Trial of Supported Socialisation’. The International Journal of Social Psychiatry 61 (3): 241–50. https://doi.org/10.1177/0020764014540150.

Smith, Matthew Lee, Eunice Chen, Christy Ann Lau, Dianne Davis, June W. Simmons, and Ashley L. Merianos. 2023. ‘Effectiveness of Chronic Disease Self-Management Education (CDSME) Programmes to Reduce Loneliness’. Chronic Illness 19 (3): 646–64. https://doi.org/10.1177/17423953221113604.

Smith, Matthew Lee, and Gang Han. 2024. ‘Effectiveness of Evidence-Based Fall Prevention Programmes to Reduce Loneliness in the United States’. Frontiers in Public Health 12:1459225. https://doi.org/10.3389/fpubh.2024.1459225.

Soh, Yenee, Ichiro Kawachi, Laura D. Kubzansky, Lisa F. Berkman, and Henning Tiemeier. 2024. ‘Chronic Loneliness and the Risk of Incident Stroke in Middle and Late Adulthood: A Longitudinal Cohort Study of U.S. Older Adults’. eClinicalMedicine 73 (July):102639. https://doi.org/10.1016/j.eclinm.2024.102639.

Southward, Matthew W., Douglas R. Terrill, and Shannon Sauer-Zavala. 2022. ‘The Effects of the Unified Protocol and Unified Protocol Skills on Loneliness in the COVID-19 Pandemic’. Depression and Anxiety 39 (12): 913–21. https://doi.org/10.1002/da.23297.

Stanford University. n.d. ‘Psychological Wellbeing Scale’. Stanford SPARQ. Accessed 17 March 2025. https://sparqtools.org/mobility-measure/psychological-wellbeing-scale/.

Taggart, Frances, Tim Friede, Scott Weich, Aileen Clarke, Mark Johnson, and Sarah Stewart-Brown. 2013. ‘Cross Cultural Evaluation of the Warwick-Edinburgh Mental Well-Being Scale (WEMWBS) -a Mixed Methods Study’. Health and Quality of Life Outcomes 11 (27). https://doi.org/10.1186/1477-7525-11-27.

Tait, Robert J, Gary K Hulse, and Suzanne I Robertson. 2002. ‘A Review of the Validity of the General Health Questionnaire in Adolescent Populations’. Aust N Z J Psychiatry 36 (4): 550–57. https://doi.org/10.1046/j.1440-1614.2002.01028.x.

‘The Warwick-Edinburgh Mental Wellbeing Scales - WEMWBS’. 2025. The University of Warwick. 26 March 2025. https://warwick.ac.uk/fac/sci/med/research/platform/wemwbs/.

Theeke, LA, JA Mallow, J Moore, A McBurney, S Rellick, and R VanGilder. 2016. ‘Effectiveness of LISTEN on Loneliness, Neuroimmunological Stress Response, Psychosocial Functioning, Quality of Life, and Physical Health Measures of Chronic Illness’. International Journal of Nursing Sciences 3 (3): 242‐251. https://doi.org/10.1016/j.ijnss.2016.08.004.

Theeke, Laurie A., Jennifer A. Mallow, Emily R. Barnes, and Elliott Theeke. 2015. ‘The Feasibility and Acceptability of LISTEN for Loneliness’. Open Journal of Nursing 5 (5): 416–25. https://doi.org/10.4236/ojn.2015.55045.

Theeke, Laurie A., Jennifer A. Mallow, and Elliott Theeke. 2021. ‘A Pilot One Group Feasibility, Acceptability, and Initial Efficacy Trial of LISTEN for Loneliness in Lonely Stroke Survivors’. SAGE Open Nursing 7:23779608211015154. https://doi.org/10.1177/23779608211015154.

Thomas, Kali S., Ucheoma Akobundu, and David Dosa. 2016. ‘More Than A Meal? A Randomized Control Trial Comparing the Effects of Home-Delivered Meals Programmes on Participants’ Feelings of Loneliness’. The Journals of Gerontology. Series B, Psychological Sciences and Social Sciences 71 (6): 1049–58. https://doi.org/10.1093/geronb/gbv111.

Tomás, Jose M., Sacramento Pinazo-Hernandis, Mónica Donio-Bellegarde, and Pedro M. Hontangas. 2017. ‘Validity of the de Jong Gierveld Loneliness Scale in Spanish Older Population: Competitive Structural Models and Item Response Theory’. European Journal of Ageing 14 (4): 429–37. https://doi.org/10.1007/s10433-017-0417-4.

Topp, Christian Winther, Søren Dinesen Østergaard, Susan Søndergaard, and Per Bech. 2015. ‘The WHO-5 Well-Being Index: A Systematic Review of the Literature’. Psychother Pschosom 84 (3): 167–76. https://doi.org/10.1159/000376585.

Urke, Helga Bjørnøy, Torill Bogsnes Larsen, and Sara Madeleine Eriksen Kristensen. 2023. ‘Preventing Loneliness and Reducing Dropout: Results from the COMPLETE Intervention Study in Upper Secondary Schools in Norway’. International Journal of Environmental Research and Public Health 20 (13): 6299. https://doi.org/10.3390/ijerph20136299.

Vahedi, Shahrum. 2010. ‘World Health Organization Quality-of-Life Scale (WHOQOL-BREF): Analyses of Their Item Response Theory Properties Based on the Graded Responses Model’. Iran J Psychiatry 5 (4): 140–53.

Valtorta, Nicole K., Mona Kanaan, Simon Gilbody, Sara Ronzi, and Barbara Hanratty. 2016. ‘Loneliness and Social Isolation as Risk Factors for Coronary Heart Disease and Stroke: Systematic Review and Meta-Analysis of Longitudinal Observational Studies’. Heart (British Cardiac Society) 102 (13): 1009–16. https://doi.org/10.1136/heartjnl-2015-308790.

What Works Wellbeing. 2020. ‘What Works to Improve Wellbeing across Different Interventions’. What Works Wellbeing. October 2020. https://whatworkswellbeing.org/resources/ons-personal-wellbeing-findings-in-impact-evaluations/.

World Health Organization. 2012. ‘WHOQOL User Manual’. World Health Organization. https://iris.who.int/bitstream/handle/10665/77932/WHO_HIS_HSI_Rev.2012.03_eng.pdf?sequence=1.

———. 2024. ‘The World Health Organization-Five Well-Being Index (WHO-5)’. Publication. World Health Organization. 2 October 2024. https://www.who.int/publications/m/item/WHO-UCN-MSD-MHE-2024.01.

Yan, Changmin, Kate Johnson, and Valerie K. Jones. 2024. ‘The Impact of Interaction Time and Verbal Engagement with Personal Voice Assistants on Alleviating Loneliness among Older Adults: An Exploratory Study’. International Journal of Environmental Research and Public Health 21 (1): 100. https://doi.org/10.3390/ijerph21010100.

Yang, Keming. 2019. ‘Loneliness: A Social Problem’, January. https://doi.org/10.4324/9781315148410.

Yap, Keong, Kiara R. Timpano, Norman B. Schmidt, and Jessica R. Grisham. 2024. ‘An Online Experimental Test of the Compensatory Process in Hoarding Disorder: Reducing Loneliness and Its Effects on Object Attachment’. Current Psychology 43 (23): 20626–39. https://doi.org/10.1007/s12144-024-05895-9.

1. The Three-Item Loneliness Scale forms part of the ONS-recommended approach for measuring loneliness in the UK, alongside a single direct question, asking ‘How often do you feel lonely?’. ↩

2. Additional stakeholder insights on the practical challenges of evaluating loneliness interventions can be found in the DCMS-commissioned report Evaluation of information to tackle loneliness, available at: https://www.gov.uk/government/publications/exploring-interventions-to-tackle-loneliness/evaluation-of-interventions-to-tackle-loneliness ↩

3. Psychometric properties refer to the technical qualities of a measurement tool that determine how well it works. This includes, for example, how reliable it is (does it give consistent results?) and how valid it is (does it measure what it claims to?). ↩

4. Negatively phrased items are questions or statements that reflect experiences, feelings, or circumstances generally associated with distress or absence (e.g., ‘How often do you feel left out?’, or ‘I miss having really close friends’), whereas positively phrased items describe the presence of supportive or favourable conditions (e.g., ‘How often do you feel that you have a lot in common with the people around you?’ or ‘I often feel rejected’). ↩