We hypothesised that the question-behaviour effect, referred to as the influence of questioning about a given behaviour on its subsequent performance, is a relevant issue when exploring the external validity of intergroup attitudes. In a pair of studies, we have corroborated that merely expressing attitudes towards the Jewish minority affects people's relevant behaviour towards this group. In an Internet study, participants who first completed verbal attitude measures were more likely to donate to a Jewish organisation compared to those who completed the measures after making the decision to donate. Moreover, responses to attitude measures of various types and donating to the Jewish organisation were correlated when attitudes had been expressed in the first step. When attitudes were measured after the decision to donate, only the responses to the traditional anti-Semitic scale were correlated with this behaviour. In the field study, in which the time interval between attitude and behaviour measures was introduced, no question-behaviour effect was observed. We explain the results with reference to cognitive dissonance and attitude accessibility mechanisms and discuss them in a broader context of attitude-behaviour research.