Controlling social desirability bias

Abstract

Social desirability bias can change the results from marketing experiments and surveys. However, there are few illustrations that show how serious social desirability bias can be. This research starts by reviewing the options for identifying and reducing social desirability bias in experiments and surveys and for controlling its effects. Then two examples that use a social desirability bias scale or a transformation of it (that may improve its utility) as control variables are described. Data from a national panel survey in the United States is used to show that controlling social desirability bias can change the set of demographic variables that are judged to be statistically significant and can have important effects on coefficient sizes. These illustrations will hopefully stimulate more consideration of social desirability bias, more use of bias measures in marketing studies, and more research on the control options.