An investigation into effects of superior-subordinate similarity on performance ratings and rating discrepancy

The effect on performance ratings of superiorsubordinate similarity in background, interests, occupational values, job needs and personality was examined. The sample included 434 dyads from a total of 84 professional employees, 16 of whom were in supervisory positions. Three primary research questions posed by the present study included: (1) Is the similarity effect an artifact of simpler rater and ratee effects? (2) What is a more appropriate dependent measure—performance rating or deviation of rating from the mean of non- dyadic raters, i.e., discrepancy? (3) Does superior familiarity with subordinate performance moderate similarity effects? Results from hierarchical multiple regression analyses provided compelling evidence that the direct relationship between similarity and performance criteria could be an artifact of rater leniency and ratee scores on similarity scales. This finding brings into question the conclusions of previous researchers. With regard to the second research question, data from this study did not support an argument for preference of either dependent measure over the other. Finally, it was found that familiarity moderated the similarity effect for values. It appears that as familiarity increases, similarity in values becomes more predictive of performance criteria.