Can graphology predict occupational success? Two empirical studies and some methodological ruminations.

Abstract

Two empirical studies for testing the validity of graphological predictions are reported. In the first, the graphologists rated bank employees on several job relevant traits, based on handwritten biographies. The scripts were also rated on the same traits by a clinical psychologist with no knowledge of graphology. The criterion was the ratings on the same traits by the employees' supervisors. The graphologists' and the clinician's correlations with the criterion were typically between 0.2 and 0.3. To test whether these validities might be attributable to the scripts' content, we developed a third method of prediction. The information in the texts (e.g., education) was systematically extracted and combined in a linear model. This model outperformed the human judges. In the second study, graphologists were asked only to judge the profession, out of 8 possibilities, of 40 successful professionals. This was done on the basis of rich (e.g., containing numbers and Latin script as well as Hebrew text), though uniform, scripts. The graphologists did not perform significantly better than a chance model. The measurement and prediction of personality traits present a major obstacle for personnel selection. Traits such as honesty, responsibility, independence, sociability, and so forth, seem to be desirable and even necessary for many occupations, yet traditional psychological testing devices typically fail to predict associated job behavior with anything approaching satisfactory rigor. The increasing demand for better personnel selection, combined with the weakness of standard personality tests, has led many firms to turn to alternative prediction methods— most notably, graphology. Levy (1979) reported that graphology is routinely used in the hiring of personnel by 85% of firms in Europe. Rafaeli and Klimoski (1983) estimated that 3,000 American firms use this tool, and the number appears to be growing. In Israel, graphology is more widespread than any other single personality test. In view of this trend, it is surprising to note the paucity of serious research efforts to assess the validity of graphology in predicting job performance. Such research as is available typi