The long and short of asking questions about income: a comparison using data from Hungary

A lot of research on income mobility and income inequality is based on survey questions about income. Various question formats are being used. Researchers seem to assume that the actual format used delivers the best estimate of the ¿true¿ income. However, surprisingly little empirical support is available for this claim. We implemented an experimental design using the short and long versions of income questions in a Hungarian survey. Results show an overall positive difference between the long and short version. The differences are related to the income components (wages and salaries, transfers, and assets), and respondent characteristics, controlling for the effect of the order of the two versions of income questions. Based on the results,we provide some recommendations for implementing income questions in surveys.