Research and analysis

Service Modernisation Customer Experience Survey: Technical report

Published 29 January 2026

Applies to England, Scotland and Wales

DWP research report no. 1120

A report of research carried out by IFF Research on behalf of the Department for Work and Pensions.

Crown copyright 2026. 

You may re-use this information (not including logos) free of charge in any format or medium, under the terms of the Open Government Licence. View this licence or write to:

Information Policy Team
The National Archives
Kew
London
TW9 4DU

or email psi@nationalarchives.gov.uk. 

This document/publication is also available on the GOV.UK website 

If you would like to know more about DWP research email socialresearch@dwp.gov.uk. 

First published January 2026. 

ISBN 978-1-78659-936-0

Views expressed in this report are not necessarily those of the Department for Work and Pensions or any other government department.

Introduction

Technical report overview

This report provides the technical and methodological details for the Service Modernisation Programme Customer Experience Survey commissioned by the Department for Work and Pensions (DWP) and conducted by IFF Research. It covers two large-scale customer surveys conducted in spring 2024 (Wave 1) and spring 2025 (Wave 2), and one additional survey of customers claiming more than one service line, conducted in spring 2025.

The report covers details on sampling, research materials and fieldwork for Wave 1 and Wave 2 surveys first, followed by the additional survey of multiple service line users. The final section on data processing, analysis and weighting covers all three surveys together.

Wave 1 and Wave 2 Surveys

Sampling

Overall approach

At the outset of the project, a target of 10,000 interviews per wave of fieldwork was set. This was split into equal targets for the nine service lines included in this research: Attendance Allowance (AA), Access to Work (AtW), Carer’s Allowance (CA), Child Maintenance Service (CMS), Disability Living Allowance for children (DLAc), Disputes Resolution Service (DRS), Maternity Allowance (MA), Pension Credit (PC) and State Pension (SP). Each service line had an overall target of 1,111 interviews per wave.

The aim across each wave of fieldwork was to achieve 30% of the interviews online (3,000 interviews) and 70% through Computer Assisted Telephone Interviews (CATI) (7,000 interviews).

A random selection of the DWP customer population was drawn, including a variety of both new and existing customers across the nine service lines. This sampling plan was set through discussion with DWP at Wave 1, with the method then repeated at Wave 2.

As the DWP customer population changes over time, the two waves of the survey represent ‘snapshots’ of customer perceptions at two different times, rather than presenting longitudinal findings of customer experience between waves. While changes in the population were small between waves, differences in customer experience discussed in the main findings report could be impacted by changes in the DWP customer profile as well as changing experiences or perceptions.

The sample was drawn by DWP using a specification provided by IFF research.

Service Line definitions

The population for each of the 9 service areas were defined as follows:

1. Carer’s Allowance: Customers in England and Wales with an open award, managed and paid to them by the Carer’s Allowance benefit system. Customers who were receiving support through a Carer Premium within their Universal Credit claim were out of scope.

2. Attendance Allowance: Customers in England and Wales with a claim that was open and awarded.

3. Disability Living Allowance for children: All child claimants aged 15 years old or under in England and Wales with a claim that was open and awarded.

4. State Pension: Customers in Great Britain with a claim that was open and awarded.

5. Pension Credit: Customers in Great Britain with a claim that was open and awarded.

6. Maternity Allowance: Customers in Great Britain with a claim that was open and awarded.

7. Child Maintenance Service (CMS): Split between Paying and Receiving Parents associated with any active CMS case in Great Britain. Any parent attached to more than one open case was selected only once, and for the case they had been attached to the longest.

8. Access to Work: Customers in Great Britain who had been awarded any element of Access to Work support that was either ongoing and still in payment, or had resulted in a one-off payment received within the previous 6 months.

9. Disputes Resolution Service: Customers who had requested a Mandatory Reconsideration or Appeal related to their Universal Credit or Personal Independence Payment claim, and this had either been decided in the previous 3 months or was still in progress (very recently opened cases were not used). Any customer with more than one of these events was selected for the one with the most recent action on it.

Where customers fit criteria for more than one service, they were sampled to one service line based on the most recent action date.

Stratification and sample draw

To ensure accurate representation of the population and make-up of customers within each service line, each service line was split into ‘New’, ’Contact’ or ‘Stock’ stratifications based off the status or length of their claim. The definitions of each are listed in Table 1 for Wave 1 and Table 2 for Wave 2. The definitions were used to be consistent with the DWP Customer Experience Survey definitions, which is a separate survey run by DWP.

Table 1 - Definition of stratifications for each service line at Wave 1:

Service Line	New Claim / Open dispute	Contact	Stock claim or Closed dispute
Attendance Allowance	Made a claim in October to December 2023	Made a new claim or notified a change of circumstance on an existing claim in July to September 2023	Claims made before October 2023, and not in ‘Contact’ group
Carer’s Allowance	Made a claim in October to December 2023	Made a new claim or notified a change of circumstance on an existing claim in July to September 2023	Claims made before October 2023 and not in ‘Contact’ group
Child Maintenance Service	Made a claim in October to December 2023	Made a new claim or notified a change of circumstance on an existing claim in July to September 2023	Claims made before October 2023 and not in ‘Contact’ group
State Pension	Made a claim in October to December 2023	Made a new claim or notified a change of circumstance on an existing claim in July to September 2023	Claims made before October 2023 and not in ‘Contact’ group
Pension Credit	Made a claim in October to December 2023	Made a new claim or notified a change of circumstance on an existing claim in July to September 2023	Claims made before October 2023 and not in ‘Contact’ group
Disability Living Allowance for children	Made a claim in July to December 2023	Made a new claim or notified a change of circumstance on an existing claim in July to September 2023	Claims made before July 2023 and not in ‘Contact’ group
Maternity Allowance	Made a claim in October to December 2023	This stratification was not used for Maternity Allowance	Claims made before October 2023
Disputes Resolution Service	Dispute open	This stratification did not apply to Disputes Resolution Service	Dispute closed since September 2023
Access to Work	Not split by length of claim due to nature of service line	Not split by length of claim due to nature of service line	Not split by length of claim due to nature of service line

Table 2 - Definition of stratifications for each service line at Wave 2:

Service Line	New claim /Open dispute	Contact	Stock claim or Closed dispute
Attendance Allowance	Made a claim in October to December 2024	Made a new claim or notified a change of circumstance on an existing claim in July to September 2024	Claims made before October 2024 and not in ‘Contact’ group
Carer’s Allowance	Made a claim in October to December 2024	Made a new claim or notified a change of circumstance on an existing claim in July to September 2024	Claims made before October 2024 and not in ‘Contact’ group
Child Maintenance Service	Made a claim in October to December 2024	Made a new claim or notified a change of circumstance on an existing claim in July to September 2024	Claims made before October 2024 and not in ‘Contact’ group
State Pension	Made a claim in October to December 2024	Made a new claim or notified a change of circumstance on an existing claim in July to September 2024	Claims made before October 2024 and not in ‘Contact’ group
Pension Credit	Made a claim in October to December 2024	Made a new claim or notified a change of circumstance on an existing claim in July to September 2024	Claims made before October 2024 and not in ‘Contact’ group
Disability Living Allowance for children	Made a claim in July to December 2024	Made a new claim or notified a change of circumstance on an existing claim in July to September 2024	Claims made before July 2024 and not in ‘Contact’ group
Maternity Allowance	Made a claim in October to December 2024	This stratification was not used for Maternity Allowance	Claims made before October 2024
Disputes Resolution Service	Dispute open	This stratification did not apply to the Disputes Resolution Service	Dispute closed since September 2024
Access to Work	Not split by length of claim due to nature of service line	Not split by length of claim due to nature of service line	Not split by length of claim due to nature of service line

At both waves, the sample was also stratified by the availability of email and phone number contact information. Any customers that had neither a phone or email address were excluded for the purposes of drawing the sample because of the methods being used to contact them for the survey. ‘Phone only’ sample, defined as customers with a phone number but no email address, were intended to be targeted for CATI surveys. ‘Email’ sample, which was customers with both a phone number and an email, were targeted for online completes.

At Wave 1, CATI surveys were set with an aim of 20% response rate for ‘phone only’, apart from CMS Paying Parents who had a target of 12%. CMS had a lower target due to historic lower response rates in other DWP surveys.

‘Email’ sample was drawn with the aim of achieving a 10% response rate.

After calculating population counts and split of email and phone sample, the final sample was drawn based off these estimated response rates. The sample draw for Wave 1 is shown in Table 3 below.

Table 3 - Wave 1 sample draw:

Email draw	Phone only draw	Total draw	Estimated target completes
26,393	38,925	65,318	10,000

At Wave 2, target response rates for both email and phone sample were updated based on the actual response rates from the Wave 1 survey. ‘Phone only’ sample remained at roughly 20% response rate, so the draw remained similar. However, overall online response rate was lower than the target 10% rate with variations across the different stratification groups. Therefore the ‘email’ sample draw was increased to mitigate this at Wave 2 as shown in Table 4 below.

Table 4 - Wave 2 sample draw:

Email draw	Phone draw	Total draw	Estimated target completes
36,990	35,361	72,351	10,000

Research materials development

Development workshops

The questionnaire was developed collaboratively between IFF and DWP between November and December 2023 for Wave 1 and then reviewed and updated for Wave 2 in October to December 2024.

Cognitive testing

A total of 20 cognitive testing interviews were conducted ahead of the launch of the Wave 1 survey, with at least 1 customer from each service line. These interviews aimed to determine the comprehension, relevance, and appropriateness of the survey questions. The interviews were carried out between 18 January and 6 February 2024. Following Wave 1 cognitive testing, the survey was amended to reflect customer feedback.

Once again, ahead of the launch of the Wave 2 survey, 20 cognitive interviews were conducted with at least one customer from each service line. This was to test out sections of the survey that had been updated between Wave 1 and Wave 2. The interviews were carried out between 15 and 30 January 2025. All these respondents had completed the pilot survey and had opted to take part in further research. Following Wave 2 cognitive testing, some minor amends were made to the survey to reflect customer feedback.

Pilot testing

At both waves, the survey was also piloted before launching mainstage fieldwork. The Wave 1 pilot was conducted between 16 and 21 February 2024 to test the survey length and content. A total of 84 interviews were completed, 78 over the phone and 6 online, with a variety of customers from all service lines. The Wave 2 pilot was conducted between 9 and 16 January 2025. A total of 137 interviews were completed, 100 over the phone and 37 online.

Topic coverage

Both surveys comprise of 8 sections, which were: the screener, overall ratings of the service, customer journey touchpoints, attributes and the Customer Experience Drivers, customer expectations, service feedback, classification questions and permission to recontact.

Between waves, the questionnaire remained largely unchanged to allow for comparison across waves. This included keeping question wording the same where possible to allow for direct comparisons. However, there were some differences between the Wave 1 and Wave 2 surveys, including additional questions on modernisation experience and some questions that were removed to maintain survey length.

As the research was mixed mode (both online and telephone), the survey design incorporated elements known to reduce mode effects. For example, longer coded questions always used ‘read out’ lists, rather than ‘prompted/unprompted’ in the telephone script, which better reflects how the lists are viewed online.

Fieldwork

Communications to participants

In both waves, an invite letter was sent to participants who had no email, including a URL and QR code to complete the survey online and contact details of the research team to opt-out of further contact. Those who had an email received an email version of this.

The invite letter and email included:

• information about the research
• a link to complete the survey online (and QR code for physical letters)
• information about potentially being contacted by the telephone team
• how to opt-out of the research
• a link to the survey privacy policy

Respondents also received reminder emails to encourage them to complete the survey. Up to four reminder emails were sent to customers for each wave of the survey.

Fieldwork management and monitoring

At both waves, the telephone interviewers received a detailed briefing ahead of calling potential respondents. They were provided with details about the project background, research objectives and how the project would be carried out.

A series of screening questions meant that respondents could only take part if they were the named contact (or their appointees) who had contact with DWP in the last 12 months.

Survey completes were monitored on a regular basis throughout the fieldwork period and reported to DWP in regular weekly updates. The research team monitored the number of CATI completes and online completes and asked the telephone unit to focus the CATI interviews on groups with proportionately lower responses compared with the sample frame.

Wave 1 survey completes

In total, 10,172 interviews were conducted with customers across all service lines. Table 5 shows the number of interviews completed by method and service line.

Overall, 3,235 interviews were conducted online, and 6,937 interviews were conducted by telephone. Fieldwork took place between 5 March and 6 May 2024.

Table 5 - The total number of interviews completed, by method and service line in Wave 1:

Service line	Total number of interviews completed	Number of online interviews completed	Number of telephone interviews completed
Attendance Allowance	1,129	941	188
Carer’s Allowance	1,130	694	436
Child Maintenance Service	1,129	672	457
State Pension	1,127	835	292
Pension Credit	1,129	914	215
Access to Work	1,140	603	537
Disability Living Allowance for children	1,125	999	126
Maternity Allowance	1,134	696	438
Disputes Resolution Service – UC MR	239	113	126
Disputes Resolution Service – UC Appeal	239	110	129
Disputes Resolution Service – PIP MR	353	197	156
Disputes Resolution Service – PIP Appeal	298	163	135
Total	10,172	3,235	6,937

Fieldwork was monitored regularly to ensure the profile of completes was roughly aligned to the population proportions, with soft quotas applied to CATI fieldwork to target respondents from groups which had proportionately lower response rates.

At Wave 1, the adjusted CATI response rate was 25%, the email response rate was 7%, and the push to web response rate was 3%.

Wave 2 survey completes

In total, 10,178 interviews were conducted with customers across all service lines. Table 6 shows the number of interviews completed by method and service line.

Overall, 3,270 interviews were conducted online, and 6,908 interviews were conducted by telephone. Fieldwork took place between 13 February and 14 April 2025.

Table 6 - The total number of interviews completed, by method and service line in Wave 2:

Service line	Total number of interviews completed	Number of online interviews completed	Number of telephone interviews completed
Attendance Allowance	1,127	156	971
Carer’s Allowance	1,111	395	716
Child Maintenance Service	1,111	477	634
State Pension	1,113	273	840
Pension Credit	1,112	240	872
Access to Work	1,147	443	704
Disability Living Allowance for children	1,117	118	999
Maternity Allowance	1,200	593	607
Disputes Resolution Service – UC MR	245	97	148
Disputes Resolution Service – UC Appeal	285	182	103
Disputes Resolution Service – PIP MR	306	140	166
Disputes Resolution Service – PIP Appeal	304	156	148
Total	10,178	3,270	6,908

At Wave 2, the adjusted CATI response rate was 22%, the email response rate was 6%, and the push to web response rate was 3%.

Survey of multiple service line users

An additional survey was commissioned to complement the Wave 1 and Wave 2 Service Modernisation Customer Experience Surveys. This focused on the experiences of customers who use more than one DWP service line.

Sampling

The customer sample was grouped into four mutually exclusive service line clusters outlined in Table 7.

Table 7 - Definition of service line clusters:

Service line clusters	Definition
Later Life	Using 2 or more service lines, with at least one of these being State Pension or Pension Credit.
Care	Using 2 or more DWP services, with at least one of these being Disability Living Allowance for children, Carer’s Allowance or Attendance Allowance. Those in this group cannot fall into in the Later Life or Children clusters.
Children	Claiming both Maternity Allowance and Universal Credit, with the potential of using other services additionally.
Irregular Payments	Using 2 or more service lines, where at least 1 of these is Funeral Expenses Payment, Winter Fuel Payment or Cold Weather Payment. Customers in this grouping cannot meet the criteria for Later Life, Care or Children clusters.

The sample was stratified by a combination of service line cluster and presence of email (stratification group) as seen in Table 8 below

Table 8 - Sample stratification and draw:

Stratification group	Sample draw
Later Life - with email	359
Later Life - no email	2320
Care - with email	2962
Care - no email	140
Children - with email	2934
Children - no email	77
Irregular - with email	778
Irregular - no email	21
Total	9,591

Research materials development

The questionnaire largely mirrored the Wave 1 and Wave 2 Service Modernisation Customer Experience Surveys but with more focus on the experience of managing multiple claims. The questionnaire covered:

• overall customer experience

• perceptions of modernisation

• experience of modernisation

• demographics

The questionnaire was cognitively tested and piloted amongst samples of relevant DWP customers, with refinements made at each stage; 10 cognitive interviews and 46 pilot interviews were carried out, with respondents across the four service line clusters, as mentioned above.

Fieldwork

Mainstage fieldwork was carried out between 26 March and 29 April 2025. Communication to participants, as well as fieldwork management and monitoring, was conducted in a similar way to Wave 1 and 2, as described above. In total, 1,123 interviews were conducted with customers who used multiple service lines. Table 9 below shows the number of interviews completed by method and service line.

Table 9 - The total number of interviews completed, by method and service line cluster:

Service line cluster	Total number of interviews completed	Number of online interviews completed	Number of telephone interviews completed
Later Life	389	8	381
Care	350	83	267
Children	312	124	188
Irregular Payments	72	10	62
Total	1,123	225	898

Total number of interviews completed for Irregular Payments: The initial sample definition for Irregular Payments changed after the sample was drawn, resulting in fewer responses in this cluster, and more allocated to the other clusters. Originally it contained any respondent that received Funeral Expenses Payment, Winter Fuel Payment or Cold Weather Payment regardless of their other claims. The final definition meant that respondents whose other claims made them eligible for any of the other clusters were moved there instead.

In total, 898 CATI interviews were completed, yielding an adjusted response rate of 26%. The email response rate was 4%.

All Surveys: data processing, analysis and weighting

Analysis

Upon fieldwork completion, the data was processed into Excel tables and an SPSS dataset based on an agreed specification. Summary variables were defined to summarise scale questions or add banding to numeric questions. Excel tables were produced with z and t tests run at the 95% confidence level, and all differences included in the report are significant at a 95% confidence level.

Coding process

Coded responses were then added to the data. Any verbatim responses that matched pre-codes in the questionnaire were back coded without being added to the base. Verbatim responses from open questions and questions with “other specify” were coded by IFF Research’s in-house coding team.

The data outputs underwent comprehensive checks against the raw data collected to ensure the accuracy of elements such as labelling, derived variables and base sizes. Excel tables and SPSS datasets were also internally checked by the DWP research team to ensure full accuracy. Development of the datasets proceeded iteratively until finalised and agreed by both IFF and DWP.

Regression analysis

Regression analysis was performed on Wave 2 survey data to provide more evidence of the relationship between the extent to which customers have experienced a modernised service, and how this influences overall customer experience.

To achieve this, a series of multiple regression analyses were conducted. A regression model was run for each of the nine service lines, using multiple linear regression, with overall customer experience as the dependent variable, and a selection of independent and control variables from the survey. The models were run separately for each service line, given their substantial differences in demographic characteristics.

Multiple regression models the extent to which different factors predict overall customer experience, while controlling for factors like the type of interaction with DWP (for example, raising a dispute, receiving a payment), and customer demographics (for example, gender, health, employment status). Controlling for these factors increases confidence that the effects seen from the modernisation variables represent the true impact of those specific experiences on overall customer experience.

Independent and control variables were selected based on theoretical relevance to overall customer experience, prior research, and practical considerations around data availability and interpretability. Control variables were included to account for demographic and contextual differences across service lines and varied depending on the specific characteristics of each service (for example, CMS parent status was only relevant for the CMS model).

The key outputs from the regression models were:

• Adjusted R-squared (R2): This value shows how much of the variation in the overall customer experience can be explained by all the factors included in each model. A higher percentage means the model is a better fit for the data.

• Relative Importance (Johnson’s Importance ^{[footnote 1]}): This metric shows how much each factor contributes to the model’s overall explanatory power. It allows us to identify which drivers have the most influence on overall customer experience, both positive and negative.

  • For this project, importance scores were reported as percentages (0 to 100%) rather than raw regression coefficients. This approach was chosen for consistency and ease of interpretation, particularly given the mix of scales sizes across predictors and the dependent variable. While effect sizes were also calculated, they were less intuitive for some variables (for example, scale mismatches), so Johnson’s Importance was used as the primary reporting metric.

  • Johnson’s Importance is calculated by assessing the relative contribution of each predictor variable to the total variance explained by the model, considering both its unique contribution and the variance it shares with other predictors. By proportioning the shared variance, this method provides a more accurate picture of each variable’s true importance to the model’s predictive power.

  • A score of 100% would reflect a scenario in which all predictors together account for the full variance explained by the model. Individual scores represent the proportionate contribution of each variable to that total. For example, a variable with a Johnson’s Importance score of 20% contributes one-fifth of the model’s explanatory power.

Summary tables of the most important drivers of overall customer experience are included in the report. Factors that were found to be statistically significant at the 95% level are reported.

Weighting

Wave 1 and Wave 2

To ensure the findings from Wave 1 and Wave 2 of the Service Modernisation Customer Experience surveys could be generalised to the population of those in contact with DWP for the nine service lines, weights were constructed to reduce potential bias caused by differences between groups and response rates.

The weighting was carried out in two stages. The initial weight applied across each service line was a cell weighting to the population profile split by New vs. Stock, as shown in Tables 1 and 2 earlier on in the report, and then split by ‘email sample’ vs. ‘phone only’ sample, as previously defined (email sample are customers with both phone number and email). For Disputes Resolution Service, rather than New vs. Stock, the stratification was by whether they were PIP or UC customers, whether their claim was an appeal or a mandatory reconsideration, and whether their claim was open or closed.

A rim weight was then applied within each service line separately. The weighting variables were region, and age cross-tabulated with gender. For CMS specifically, this was also weighted to service type (maintenance direct or calculation and collection).

The combined dataset was not weighted to the overall population of DWP customers as this would have resulted in the largest service lines dominating (for example, State Pension) and the contribution to overall results for smaller service lines becoming insignificant. Unweighted and weighted bases by service line for Wave 1 and Wave 2 can be seen in Table 10 and 11 below, respectively.

Table 10 - Unweighted and weighted base by service line at Wave 1:

Service Line	Unweighted Base	Weighted Base
Attendance Allowance	1129	1115
Carer’s Allowance	1130	1186
Child Maintenance Service	1129	1217
State Pension	1127	978
Pension Credit	1129	1048
Access to Work	1140	1153
Disability Living Allowance for children	1125	1188
Maternity Allowance	1134	1176
Disputes Resolution Service	1129	1112

Table 11 - Unweighted and weighted base by service line at Wave 2:

Service Line	Unweighted Base	Weighted Base
Attendance Allowance	1127	1129
Carer’s Allowance	1111	1113
Child Maintenance Service	1111	1113
State Pension	1113	1097
Pension Credit	1112	1114
Access to Work	1147	1149
Disability Living Allowance for children	1117	1119
Maternity Allowance	1200	1202
Disputes Resolution Service	1140	1142

Multiple service line users

The multiple service line users survey was rim weighted by stratification group (clusters broken down by presence of email or not), age, gender and region. There were large differences in the sizes of populations of the different clusters, with over 1.4 million customers in the Later Life cluster compared to around 8,000 customers in the Children cluster. Due to this distribution, customers in the Children cluster make up a very small proportion of the population, therefore the weighting conducted reduced the overall effective sample size of the multiple service lines users to a greater extent than Wave 1 or Wave 2 results. The unweighted and weighted bases can be seen in Table 12 below.

Table 12 - Unweighted and weighted base by service line cluster:

Service line cluster	Unweighted Base	Weighted Base
Later Life	389	705
Care	350	293
Children	312	5
Irregular Payments	72	119

Reporting

Wave 1 results were analysed independently in the summer 2024. Following completion of Wave 2 fieldwork, Wave 2 data was analysed in comparison to Wave 1 data to produce a published report, Service Modernisation Customer Experience Survey: Wave 1 and Wave 2.

The data from the multiple service line users survey was analysed in summer 2025 and a standalone report was produced as an annex to the main report.

1. Johnson, J. W. (2000). A heuristic method for estimating the relative weight of predictor variables in multiple regression. Multivariate Behavioral Research, 35(1), 1–19. ↩