Stack Overflow Answerers and Visitors Survey

We support our findings of toxic code snippets on Stack Overflow by asking Stack Overflow users to take online surveys. The surveys were used for assessing awareness of the developers on the two issues of outdated code and license-violating code snippets.

There were two surveys: The Stack Overflow Answerers’ Survey and Stack Overflow Visitors’ survey.

The Answerers’ Survey

The participants were invited to take the survey via email addresses available on their public Stack Overflow and GitHub profiles. We selected the answerers based on the all-time reputation ranking. The invited participants had a reputation from 963,731 (the highest) to 6,999 and we sent out 607 emails (excluding undelivered ones, e.g. due to illegal email addresses).

The survey was open for participation for two months, from 25 July to 25 September 2017, before we collected and analysed the responses.

The Visitors’ Survey

We adopted non-probability convenient sampling to invite participants for this survey. Participation in the survey requires experience of visiting Stack Overflow for solving programming tasks at least once. The participants were invited to take the survey via five channels: social media post (Facebook), blognone.com (a popular technol- ogy news and media community in Thailand), the Univer- sity of Molise in Italy where the third author works, the comp.lang.java.programmer group, and the Software Engineering Facebook page. The survey was open for par- ticipation for 2 months from 25 July 2017 to 25 September 2017.

The complete survey results can be found in our UCL CS Research Note.