The validity of handwriting elements in relation to self-report personality trait measures

Research on graphology has provided mixed results regarding its validity in applied settings. Increasing popularity of this method (Edwards & Armitage, 1991) calls for continued evaluation. The current study sought to validate a popular graphoanalysis method proposed by Bunker (1979), which links specific handwriting elements to personality traits. Content-neutral handwriting of 49 college students was analyzed for 30 elements (e.g. length of t-cross) by two trained coders working independently. Participants also completed the Jackson Personality Inventory—Revised (JPI-R; Jackson, 1994), which measures 15 normal personality traits. Inter-rater agreement in handwriting analysis was moderately high (median reliability = 0.80). Both predicted and non-predicted correlations between handwriting elements and JPI-R scales were significant at around chance levels (i.e. 5%). In particular, of the 119 predicted relations, only six (5.0%) were significant in the expected direction and five (4.1%) were significant in the opposite direction. All told, current findings replicate those of previous studies and suggest limited value in hand-writing analysis as a predictive tool.