Traditional validation studies of handwriting analysis have two major problems: the first one is that the procedure of personality trait evaluation used was mostly manual and subjective, although often the experts demonstrated statistically good agreement. The second problem is that the researchers took for practical reasons just small subsets of the handwriting variables and selected a very restricted set of personality traits. The improved validation procedure proposed in this study and discussed in the current article is based on the computer-aided handwriting analysis HSDetect and solves both problems.
HSDetect is based on the following principles:
- Statistically based consolidation and harmonization of different handwriting analysis methods and schools to avoid biased results.
- Formal unambiguous presentation of all handwriting signs, personality traits, and the relations between them.
- Quantitative registration of handwriting signs and evaluation of personality traits.
- Assurance of evaluation objectivity and reliability.
In HSDetect, the handwriting signs are evaluated manually by experts, and the corresponding psychological traits are calculated algorithmically. Besides the algorithms of the handwriting analysis, HSDetect includes two databases: the database of the handwriting signs and psychological traits with connections between them (the handwriting analysis model) and the database of the evaluation of subjects’ handwriting samples—statistical results that serve as norming data and as a basis for different statistical studies.
Both the handwriting signs and the values of the psychological traits are presented on a continuous scale from 0 to 1. Each psychological trait is mathematically modelled as a function of several handwriting signs. The relations are complex, many-to-many, which means that a handwriting sign relates to several traits, and a typical trait is a function of several signs.