Abstract
Faced with more than 100M open source projects, a more manageable small subset is needed for most empirical investigations. More than half of the research papers in leading venues investigated filtering projects by some measure of popularity with explicit or implicit arguments that unpopular projects are not of interest, may not even represent "real" software projects, or that less popular projects are not worthy of study. However, such filtering may have enormous effects on the results of the studies if and precisely because the sought-out response or prediction is in any way related to the filtering criteria.
This paper exemplifies the impact of this common practice on research outcomes, specifically how filtering of software projects on GitHub based on inherent characteristics affects the assessment of their popularity. Using a dataset of over 100,000 repositories, we used multiple regression to model the number of stars -a commonly used proxy for popularity- based on factors such as the number of commits, the duration of the project, the number of authors and the number of core developers. Our control model included the entire dataset, while a second filtered model considered only projects with ten or more authors. The results indicated that while certain characteristics of the repository consistently predict popularity, the filtering process significantly alters the relationships between these characteristics and the response. We found that the number of commits exhibited a positive correlation with popularity in the control sample but showed a negative correlation in the filtered sample. These findings highlight the potential biases introduced by data filtering and emphasize the need for careful sample selection in empirical research of mining software repositories. We recommend that empirical work should either analyze complete datasets such as World of Code, or employ stratified random sampling from a complete dataset to ensure that filtering is not biasing the results.