This brief paper discusses the relevance of conducting surveys that measure individuals’ attitudes for understanding fiscal behaviour. While many surveys assess individuals’ attitudes towards paying taxes (e.g. by asking them to what extent they believe tax evasion is ever justified), it is less clear whether individuals’ responses to such survey questions are indicative of the way they would behave in reality. The paper presents a discussion of the way attitudes have been assessed in tax surveys and assesses existing evidence to support a link between these attitude measures and actual compliance behaviour. The paper suggests several avenues to improve the predictive value of attitude measures, such as increasing the specificity of measures, using evaluation scales or mitigating social desirability biases. A series of recommendations are made for measuring attitudes and interpreting attitude surveys for the use of researchers planning to conduct survey work, as well as for the use of findings from taxpayer surveys in the design of tax policy and administration.