In this paper we examine the presence of panel conditioning in a payment diary survey studying the payment behavior of Dutch consumers. We analyze whether the reporting behavior of frequent participants in the payment survey systematically differs from that of less-frequent survey participants. We introduce refreshment groups that allow us to compare the reporting behavior of ˜trained and ˜fresh respondents, where the differences between reported values are used as a proxy for the panel conditioning effect. We find no consistent significant differences for the number and value of cash and debit card payments between trained and fresh respondents, for any demographic group, sector or transaction size. Likewise we find no consistent significant differences for the value of cash withdrawals between trained and fresh respondents. However we do find some signs of panel conditioning in the reported number of cash withdrawals.
Keywords: payment behavior, diary studies, survey design, panel conditioning.
JEL classifications: C81, C93, D12. Working paper no. 546