Source-linked AI summary
Happy software developers solve problems better: psychological measurements in empirical software engineering
Daniel Graziotin, Xiaofeng Wang, Pekka Abrahamsson
TL;DR
Research has offered limited support for the claim that happier software developers perform better, despite the importance of human factors in software engineering. This study examined affective states, creativity, and analytical problem solving, finding that the happiest developers were significantly better analytical problem solvers, while creativity showed no such result.
Problem
Limited research has supported claims that making software developers satisfied and happy improves productivity and software quality, despite the importance of human factors.
Method
The study used psychological measurements of affective states, creativity, and analytical problem-solving performance, including a Tower of London-based analytical problem-solving score.
Results
APS scores differed significantly between N-POS and POS groups (t(33.45) = -2.82, p = 0.008, d = -0.91), with the happiest developers performing better analytically but not creatively.
Takeaways & Limitations
The findings support multidisciplinary study of affective states and human factors as part of basic software engineering research.
Takeaways & Limitations
The participants were all Computer Science students with limited software development experience compared with professionals.
Abstract
from arXiv · showhide
For more than 30 years, it has been claimed that a way to improve software developers' productivity and software quality is to focus on people and to provide incentives to make developers satisfied and happy. This claim has rarely been verified in software engineering research, which faces an additional challenge in comparison to more traditional engineering fields: software development is an intellectual activity and is dominated by often-neglected human aspects. Among the skills required for software development, developers must possess high analytical problem-solving skills and creativity for the software construction process. According to psychology research, affects-emotions and moods-deeply influence the cognitive processing abilities and performance of workers, including creativity and analytical problem solving. Nonetheless, little research has investigated the correlation between the affective states, creativity, and analytical problem-solving performance of programmers. This article echoes the call to employ psychological measurements in software engineering research. We report a study with 42 participants to investigate the relationship between the affective states, creativity, and analytical problem-solving skills of software developers. The results offer support for the claim that happy developers are indeed better problem solvers in terms of their analytical abilities. The following contributions are made by this study: (1) providing a better understanding of the impact of affective states on the creativity and analytical problem-solving capacities of developers, (2) introducing and validating psychological measurements, theories, and concepts of affective states, creativity, and analytical-problem-solving skills in empirical software engineering, and (3) raising the need for studying the human factors of software engineering by employing a multidisciplinary viewpoint.
Materials and methods
The study measured affective states, creativity, and analytical problem-solving in software developers using SPANE questionnaires and task-based scores. Analytical problem solving was quantified by Tower of London performance relative to planning time, and data were analyzed in R.
- Measures: Analytical Problem Solving (APS) was defined as each Tower of London trial’s progress score divided by the seconds needed to plan its solution.The Tower of London task contained 12 problems, each solvable in a maximum of three trials, for scores ranging from 0 to 36.
- Measures: Planning time (PTS) measured milliseconds from problem presentation to the participant’s first mouse click.PTS served as the time component of the APS score.
- Procedure: Affective states were measured with the SPANE-B questionnaire before and immediately after the creativity task.The second session was included because the first task could change participants’ affective states.
- Analysis: The data were aggregated and analyzed using the open-source R software.Group comparisons used unpaired, two-tailed t-tests; the analytical problem-solving test used Welch’s correction.
- Measures: Creativity was assessed using alternative creativity (ACR), boundary creativity (BCR), and the number of creative responses (NCR).The two creativity scores were compared with the number of generated captions across affective-state groups.
Figures
The figures illustrate the creativity task, the first level of the Tower of London game, and comparisons of analytical problem-solving between the N-POS and POS groups. They also depict the relationship between analytical problem-solving and affect balance.
- Figure 1 shows a photograph used for the creativity task.
- Figure 2 presents the first level of the Tower of London game.
- Figure 3 displays boxplots for analytical problem-solving (APS) in the N-POS and POS groups.
- Figure 4 shows a scatterplot of analytical problem-solving (APS) versus affect balance (SPANE-B) between the N-POS and POS groups.
Tables
The tables section identifies a table reporting task-score means and standard deviations by group, alongside supplemental participant instructions for the experiment.
- Table 1 reports task-score means and standard deviations divided by participant groups.
- The supplemental reference sheet provides experiment instructions and asks participants not to log out, shut down, or reboot the computer.
- Participants receive an anonymous reference code to connect their surveys with experiment data.
1. Survey
The survey instructed participants to access a provided URL, answer all questions about the past four weeks, and submit their responses. It was expected to take less than five minutes, while allowing participants to take their time.
- Participants were instructed to open a browser, visit the provided URL, and answer all survey questions.
- The survey reference period covered the past four weeks, including the present, and participants had to provide a reference code.
- Participants were reminded to submit the survey after completion; it was expected to take less than 5 minutes, but they were told to take their time.
2. Photographs game
In the photographs game, participants write captions for two photographs as though competing in a magazine’s Best Caption of the Year contest. They receive the photographs one at a time after providing a Reference Code to supervisors.
- Photographs game: Participants provide their Reference Code to supervisors before receiving the photographs.The passage specifies the format as “<Reference Number>.”
- Photographs game: Each participant receives two photographs, presented one at a time.The task therefore involves two sequential photograph-caption trials.
- Photographs game: Participants write the best captions possible, imagining a famous magazine’s Best Caption of the Year contest.Winning captions would be published alongside the photographs, and captions may be anything participants choose.
3. Survey
The survey asked participants to answer all questions about the past four weeks, provide a reference code, and submit their responses after completion. It was expected to take less than five minutes, although participants were encouraged to take their time.
- Survey procedure: Participants were instructed to access the survey through a browser, answer all questions, and consider the preceding four weeks, including the present.They were also asked to provide a reference number.
- Survey procedure: Participants were reminded to submit the survey after finishing it.The survey was expected to take less than five minutes, but participants were encouraged to take their time.
4. Tower of London game
The Tower of London game was administered through PEBL using the Shallice Test configuration, which includes 12 problems with three disks and specified pile heights. Participants were instructed to complete the game in about 10 minutes and report to a supervisor without closing the program.
- Administration: Participants entered their reference code in PEBL and launched the TOL.pbl experiment from the battery/tol/ path.They were instructed not to press the “+” button.
- Test configuration: The selected Shallice Test used [1, 2, 3] pile heights, three disks, and Shallice’s 12 problems.Participants selected the test by pressing key 3 when prompted.
- Completion instructions: Participants were told the game should take about 10 minutes, should be completed without rushing, and should be followed by contacting a supervisor.They were also told not to close the program or log out, shut down, or reboot the computer.