Establishing personal trust‐based connections in distributed teams

Abstract

Trust is a factor that dramatically contributes to the success or failure of distributed software teams. We present a research model showing that social communication between distant developers enables the affective appraisal of trustworthiness even from a distance, thus increasing project performance. To overcome the limitations of self‐reported data, typically questionnaires, we focus on software projects following a pull request‐based development model and approximate the overall performance of a software project with the history of successful collaborations occurring between developers.

A large-scale, in-depth analysis of developers’ personalities in the Apache ecosystem

“…Furthermore, the results of the logistic regression show that more open developers are more likely (+36%) to contribute commits that are successfully integrated into a project repository. This finding complements the results of our previous work [86,87], where we found that more agreeable integrators are more likely to accept the pull requests during code review sessions. Agreeableness, in fact, is associated with the propensity to trust other, being empathetic, and avoiding harsh confrontations -facets of personality that are 'helpful' during cooperative tasks such as code reviews, where more open/agreeable contributors and integrators are likely to collaborate with less friction.…”

“…On the one hand, these results are not surprising; in fact, they are in line with the results of both RQ4 (i.e., no differences in mean personality traits score among developers when grouped by activity level) and prior work that uncovered the technical antecedents of accepted contributions in OSS projects (e.g., [120,121]). On the other hand, combined with the findings from our previous work on trust [86,87] and RQ5 (i.e., more open developers are more likely to contribute), these results suggest that personality may have an impact on development activities that entail direct communication with others, as in code review tasks. Still, given the marginal fit and the cross-sectional nature of the data fed into the regression models,here we can only hint at possible causal relations, which we reserve to investigate in future work.…”

“…However, analysis at a finer granularity, such as task level, may make it easier to contextualize the relevance and availability of traits in digital traces left by developers. For example, in code reviews, developers performing the inspection of external contributions are likely to behave in ways that make Agreeableness (i.e., cooperation with others) and Conscientiousness (i.e., thoroughness of the inspection) emerge from their comments, as supported by initial evidence reported in our previous work on trust [86,87].…”

“…Calefato et al [87] and Calefato & Lanubile [86] investigated the relationship between project success and propensity to trust, a facet of the Agreeableness trait in the FFM. To avoid subjectivity in the assessment of project success, they approximated the overall performance of two Apache projects with the history of successful collaborations, i.e., code reviews of pull requests in GitHub.…”

“…Calefato et al [87] and Calefato & Lanubile [86] found initial evidence that the propensity to trust -i.e., the facet of Agreeableness representing the individual disposition to perceive the others as trustworthy -is positively correlated with the chances of successfully accepting contributions in code review tasks. Yet, trust is one the many facets in the Big Five model and previous research did not look at the effects of the personality of developers who author those contributions.…”