The halo effect: Is it a unitary concept?

Four alternative techniques are commonly used for measuring the halo effect in performance appraisal: dimension intercorrelations, factor analysis, raters x ratees interaction, and rating dispersion. The study tested the equivalence of these techniques by measuring possible differences in their respective estimates of halo when two variables were manipulated: the global impression held by the ratee (poor vs. excellent) and status relative to the rater (superior, colleague, subordinate). Each of 125 Israeli managers was asked to assess on a seven‐dimension form one outstanding and one poor superior, colleague and subordinate with whom they had worked. Findings indicate that the different measures of halo were differentially influenced by the manipulated variables. The intercorrelation and factor analysis measures yielded no significant differences among ratee groups, while both the standard deviation and the interaction measures were affected by the two manipulated variables. These findings suggest that the halo effect may be divided into two distinguishable types: covariance, reflected by intercorrelation and factor analysis methods, and co‐occurrence, measured by the interaction and dispersion methods. The article discusses the conceptual meaning of these types of halo and the factors to which they are differentially sensitive. 1983 The British Psychological Society