Understanding Fairness in Software Engineering: Insights from Stack Exchange Sites

Data DisclaimerThe ethical side of the study has been evaluated by the corresponding ethical committee and it has been concluded that the study is ethically appropriate.The data in this repository comprises contributions from users on the Stack Exchange network, shared under the cc-by-sa 4.0 license encouraging sharing and remixing with proper attribution. Users' registration with Stack Exchange is subject to their explicit agreement to make all content available under the Creative Commons license terms. This data analysis, characterized by its intentionally permissive nature and implicit consent, carries no foreseeable risks or discomfort that could negatively affect users associated with the research (see here). However, we acknowledge the importance of user’s privacy and ethical consideration.To adhere to EU General Data Protection Regulation (GDPR), we have undertaken measures to limit the spread of personal user data, ensuring that the current dataset is pseudonymized. Detailed information on these measures is only available upon request for those seeking to replicate or extend this research.We did not intentionally analyze posts made by minor users due to limitations in identifying such users. However, the Stack Exchange Terms of Service specifies that users must be at least 13 years old to access or use the Network or Services, with a minimum age requirement of 16 within the European Union, as outlined in the age eligibility section here.If you believe your information is included in this dataset and would like it removed, please do not hesitate to contact us.Files<b>Script.ipynb. </b>In our Methodology, we created a query to find posts that may involve fairness discussion. To accomplish this, we compiled a list of keywords and phrases that are likely to be related with concerns about fairness. The query containing list of keywords and phrase is available in this file. This file also contains script to obtain information regarding fairness post owners.<b>Fairness Posts on Stack Exhange.xlsx</b> contains results for RQ1, RQ2.1, and RQ2.2. The spreadsheet contained six sheets:<i>Statistics</i>: sheet containing statistics of the data<i>WP StEx Posts</i>: fairness posts from Workplace Stack Exchange<i>SE StEx Posts</i>: fairness posts from Software Engineering Stack Exchange<i>PM StEx Posts</i>: fairness posts from Project Management Stack Exchange<i>OS StEx Posts</i>: fairness posts from Open Source Stack Exchange<i>DO StEx Posts</i>: fairness posts from DevOps Stack ExchangeIn the sheet containing posts from the sites, columns B-G are for questions, H-J are for answers. The rest of the columns refer to the post's context (<i>Context</i>), protected attribute mentioned (<i>Protected Attributes</i>), and <i>Fairness Dimensions</i> discussed. All fairness posts are color-coded as green. If an answer is color coded as green, its corresponding question is available on the corresponding row. <i>Notes</i> we made during analysis of fairness dimension is available in the last column.To answer RQ2.3, we inferred post owners' demographics using manual analysis of user information available in <b>Fairness on Stack Exchange_User Information.xlsx</b>. The spreadsheet contained six sheets:<i>Summary</i>: summary of post owner's information<i>WP StEx Fairness Posts Users</i>: user informations who write fairness post on Workplace Stack Exchange<i>SE StEx Fairness Posts Users</i>: user informations who write fairness post on Software Engineering Stack Exchange<i>PM StEx Fairness Posts Users</i>: user informations who write fairness post on Project Management Stack Exchange<i>OS StEx Fairness Posts Users</i>: user informations who write fairness post on Open Source Stack Exchange<i>DO StEx Fairness Posts Users</i>: user informations who write fairness post on DevOps Stack ExchangeWe looked at their information about geographical location by looking at any inferable location from any content in <i>Location</i>, <i>Website</i>, and <i>AboutMe </i>that correspond to the <i>UserId</i>s. We also attempted to infer users' roles from their <i>AboutMe</i> and <i>Website</i>.We explored the potential impact of regional differences on the occurrence of fairness problems by inferring a possible relationship between the location of fairness post owners and the geographic distribution of users on each site. We ran a query <b>Query_Proportion of Users with Location Available.pdf</b> on all 5 sites to obtain the geographic distribution of the users.<b>Site Population Distribution.xlsx </b>contains the <i>L</i><i>ocation</i> and the <i>Count</i> of the overall users in the 5 StEx sites. The spreadsheet contained 5 sheets:<i>WP StEx Users Locations</i>: Location and count of users on Workplace Stack Exchange<i>SE StEx Users Locations</i>: Location and count of users on Software Engineering Stack Exchange<i>PM StEx Users Locations</i>: Location and count of users on Project Management Stack Exchange<i>OS StEx Users Locations</i>: Location and count of users on Open Source Stack Exchange<i>DO StEx Users Locations</i>: Location and count of users on DevOps Stack Exchange<br>