An exploratory survey of Microsoft employees had respondents complete a personality test as well as answer questions about various work practices and various beliefs about software development. Questions about work practices came from discussion with and observation of developers. Questions about software development were selected from a list of controversial topics at Stack Overflow. The number of respondents was 797, which represents a 26 percent response rate.
Around one in five respondents reported that they did not use Stack Overflow. Around three in five respondents reported that they listened to music while working. Around half of the respondents reported that they disagreed that open workspaces are more productive. A statistical analysis was undertaken to determine if any relationships existed between the five measured personality traits and demographics, work practices, and beliefs. After adjusting for multiple comparisons, six relationships were found to remain statistically significant. For example, managers were found to be more conscientious. Those who had built tools on their own initiative were also found to be more conscientious. No difference in conscientiousness was found between developers and testers, a result that contradicts a previous study. No attempt is made to explain this contradiction.
The discussion throughout this paper sets aside the adjustment for multiple comparisons. Given that questionnaires are weak instruments, many readers will feel the focus should have been solely on those findings that survived the adjustment to p-values. Despite this criticism, this short paper provides some useful insights and is recommended to the software engineering community.