Stability, Reliability and Validity of Social Value Orientation

Abstract

Social value orientations are often regarded as stable individual differences in otherregarding preferences - the concern for equality and joint outcomes - affecting the propensity to cooperate in experimental social dilemma situations. Despite the widespread use of measures of social value orientations in social dilemma experiments, little is known about their stability, reliability and validity. Study 1 examines the longitudinal stability of social value orientations, showing that the stability coefficient of a linear measure of social value orientations in a panel survey over 19 months is only .22, correcting for measurement errors. Study 2 shows that the degree of altruism displayed in decomposed games and the classification of respondents in types of social value orientations depends on the place of the measurement procedure in a questionnaire, threatening their validity. When social value orientations are measured in a survey after behavioral questions on charitable giving, the proportion of ‘cooperators’ increases from 41% to 57%. Priming effects may be the reason for the low stability of social value orientation. Study 3 shows that priming effects also disturb the relationship of social value orientation with behavioral criteria such as volunteering and donations to charitable causes. Social value orientations are correlated with giving and volunteering behavior in the predicted pattern when social value orientations are measured before the behavioral questions. However, when social value orientations are measured after the behavioral questions, their predictive power declines.