Do divergences between stated and actual willingness to pay signify the existence of bias in contingent valuation surveys?