CHAPTER 3. SURVEY OF COMMUNITIES
3.5 Conclusion
The gender imbalance results match previous work in identifying a gap in the number
improvement and hopefully can be examined through further study of the socialization process. Worldwide distribution shows that over half of the respondents were from the United States and Europe. These results could be skewed slightly since the solicitation for participation and the survey itself were in English. For those not fluent in reading English they may not have been able to respond to the survey.
Most of the participants in both communities have a reasonably high level of experience contributing to FLOSS communities and possess technical expertise in programming. The amount of time spent each week in various areas such as programming, fixing bugs, commu- nication, and other areas indicates that in general for those who responded, people in Mozilla spend more total time than KDE. This may be reflected by the fact that there were a large number of participants responding that receive monetary compensation to work at Mozilla as opposed to KDE. When examining the relative percentages of time spent in various activities, they are overall roughly similar. Communication takes up a large amount of time for Mozilla participants and to a lesser extent KDE participants. The amount of time participants spend fixing and responding to bugs is consistent with the maturity and stability of both Mozilla and KDE projects.
Many of the participants self-identified as being at the beginner stages of integration into the community for KDE and to a lesser extent Mozilla. The integration level that participants would like to have in the future showed that a large number of participants in both commu- nities would like to be expert and core members of the community. This indicates that these participants do not want to be at the edge of the community but rather be active and high level members after their socialization. Understanding wanted integration levels might help project members focus attention and resource on the appropriate people.
The large variance in the amount of time participants spent watching the mailing list before
posting matches previous research byvon Krogh et al.(2003). This emphasizes the importance
of understanding the social fabric of a community before engaging with it.
In the survey areas related to gauging new members and the socialization process, approxi- mately 50% of respondents for Mozilla indicated some level of difficulty in how easy it is for new members to join the community. This suggests a potential area for improvement. The actual
skills that participants felt new members should possess again show some importance for tech- nical and programming skills however also suggests a high importance in social communication skills. For Mozilla, approximately half of the participants disagreed and felt new members are not easily integrated into the community echoing the results of the previous Likert question. In both communities, participants felt new members are treated with respect.
Participants in both Mozilla and KDE felt that keeping track of various processes in the community such as mailing lists, bug reports, source code changes, and other areas were fairly effective. However, roughly one quarter felt that there could be some improvement in many of these areas. Tools focusing on these particular areas could help in these processes, particularly for new members.
The results from the quantitative and qualitative data seem to paint a primarily positive picture of socialization and inclusion for KDE and perhaps a slightly larger, more complex process of socialization and inclusion for Mozilla because of the daunting size of its community and codebase. Interestingly, in many cases for both communities, the ratings for the community on a number of factors relating to the socialization process and the ability for the community to effectively coordinate tasks were actually higher for the beginner participants as compared to the veterans. This indicates that perhaps the beginners have a slightly rosier view or outlook on the community while those that are veterans know the “true” state of the community. This suggests that new participants may have an over-inflated view of the overall ease of joining, participating, and contributing to the community than veteran members. Ensuring realistic expectations for beginners, while at the same time keeping tasks simple to ease the transition process, may be an important step in this socialization process to ensure new members don’t leave the community early.
The results of this study are subject to the following threats to validity. External validity is defined as how well the findings generalize to other populations or groups. In this case, there appeared to be some overall differences in perception between Mozilla and KDE. This suggests the skewed difference in perception between beginner and veteran community members perhaps may not generalize to other large FLOSS projects. Obviously, projects vary in terms of how they’re organized and developed, especially as their size increases.
The survey data was gathered via convenience sampling. Not all members of the commu- nity participated. The survey may have pulled participants from more technical backgrounds and missed people in other areas of the community such as localization, advocacy, writing, and other areas. A larger sample size pulling participants from more areas of the community would address this issue. In general, the small sample size obtained for each of the two communities means these results should be taken with that under consideration. Increased participation in the survey would increase the validity, however overall, the trends of the data suggest a pre- dominantly positive health of both communities with potential for improvement in socialization of new participants. The results of this initial survey and ideas suggested were considered and evaluated during the brainstorming and development of the algorithms and Jamii website.