Evaluative neutralisation as a means to reduce social desirability in the SDO-7

Abstract: An evaluative neutralisation process, whereby questionnaire items are phrased to be less evaluative, was used to create a new less socially desirable version of the social dominance orientation scale (SDO7). The main goal is to counteract the persistent floor effect which has affected the scale since its inception. The rationale behind the suggested correction is rooted with the problem of social desirability in self-rating inventories. A translated scale (Swedish) was used and it was constructed and tested over three different samples. The final sample consisted of 245 participants (61.5% female). It is argued and demonstrated that simple neutralisation can lessen the floor effect, increase variance and even result in an improved inventory, as measured by the scale’s predictive power compared to 4 criterion variables. Paired samples t-tests indicated that the mean of the new scale was almost 1 point higher on a 7 point Likert-scale compared to the original scale, variance increased by 30%, normality of the SDO scores were better distributed and number of outliers dropped from 13 to 3. Correlations between the new scale and criterion variables were significantly stronger on half of the criterion-variables and stronger on the rest but without reaching significance. All of these improvements suggest that the new scale outperforms the original and that the evaluative neutralisation process was successful. The importance of an evaluative neutralisation for future SDO research is discussed in terms of construct validity and predictive power.