Personality types in software engineering

Reviewer: Gerhard Chroust

The distribution of the Myers-Briggs Type Indicators is studied in a sample of 100 software engineers, and compared to the US average distribution (no source is given). These indicators distinguish four pairs of psychological characteristics: The results of four older investigations and this one are given, indicating usually consistent distributions of characteristics for software engineers. Several interesting observations are shown. For instance, psychological characteristic combinations of ISTJ, ESTJ, and TJ comprise 24 percent, 15 percent, and 50 percent, respectively, of the software engineers studied, while these combinations are 11.6 percent, 8.7 percent, and 10 percent, respectively, in the US population. Conclusions about the relation of these findings to job skills of software engineers are provided. Despite the plausibility of the conclusions, there are some caveats. First, the data are poorly presented in unsorted tables, and graphics would have improved readability and insight. Second, sampling 100 individuals according to four binary alternatives gives an average of six (100/16) subjects per class, causing instability due to small numbers: 1.5 percent of subjects classified as INFS are in the US population. The difference between one or two subjects in the sample results in a deviation of 67 percent, or 130 percent from the average US population. Several other data have similar problems. Finally, the study ignores the approximately 20 different roles in software engineering, their different psychological requirements (such as testing versus requirements engineering), their relation to experience and seniority, and thus their effect on the T-F distribution and others. But all in all, this is an interesting study that provides a good starting point for further comparisons. Online Computing Reviews Service