Answering for yourself versus others : direct versus indirect estimates of charitable donations
It is common for researchers in marketing and other social sciences interested in ethical behavior such as propensity to donate to a charity to ask “indirect” questions about others (e.g., “what would another student donate?”) in order to measure respondents’ own propensity to donate. The idea is that people project their own desires onto their responses about others and that they are more likely to admit a lower level of generosity when they are under the lessened social pressure of the indirect question. In these four studies, we measure estimates respondents make about self (self-estimates) and others (other-estimates) as well as their actual donations to charity, doing so at an individual level to challenge the prevailing wisdom that indirect responses are more accurate and useful for research than direct responses. I conceptualize accuracy in terms of both the mean and correlations. I show that although mean-level results sometimes show other-estimates to be closer to actual behavior, they are not consistently so, and explain this inconsistency; and further, that correlations show self-estimates to always better reflect actual donations than other-estimates. These results support the use of self-estimates in the ethical domain and argue against the existence of projection in marketing research donation responses.