Levels and subject

When does reference values or targeted subject influence positional preferences?

Abstract

In this study, we use an experimental survey approach to if the degree of positionality is sensitive to variations in reference levels and targeted subject. Based on previous research in economics and psychology, our hypotheses are that 1) people are more positional when they choose between alternatives with relatively high consumption levels, and 2) people are more positional when they choose for a hypothetical grandchild, than for themselves. We measure positional preferences in five domains – Income, housing, vacation and SAT-score, and test our hypotheses on a large representative sample from the US (N=1300).  As social demographic indicators, we include information about gender, birth year, children or grandchildren, individual income, vacation days, size of home and reported SAT-score. Our results suggest that the instruments commonly used to elicit positional preferences are relatively insensitive to variations in consumption levels and targeted subject, with a few important exceptions. First, we find that positional preferences for income and SAT scores depend on the reference level used in the hypothetical choice scenarios. Second, our results suggest that people are significantly more likely to choose the positional option for housing when they choose for a hypothetical grandchild than when they choose for themselves.