We present the results of a survey-experiment -using a representative sample of the Dutch population -in which we relate respondents' opinion about the tax deductibility of mortgages to their estimates about other people's opinion. The experiment employs three treatment variables: monetary incentives, the provision of arguments pro and contra, and ambiguity of the question posed. We find that respondents are characterized by a significant consensus effect. Respondents' estimates of others'

more »

... ons are strongly related to their own opinion. The size of the effect, however, is not affected by ambiguity of the question posed. Information by means of the provision of arguments pro and contra the tax provision does reduce the consensus effect significantly, though. Monetary incentives appear to have only a weak effect. We also find a strong effect of house ownership. Not only are house owners more in favor of the tax provision, they are also characterized by a significantly stronger consensus effect. These results suggest that both cognitive factors and motivational factors are responsible for the consensus effect. JEL-codes: A13, C99, H29