Exploring the Relationship between Membership Turnover and Productivity in Online Communities

One of the more disruptive reforms associated with the modern Internet is the emergence of online communities working together on knowledge artefacts such as Wikipedia and OpenStreetMap. Recently it has become clear that these initiatives are vulnerable because of problems with membership turnover. This study presents a longitudinal analysis of 891 WikiProjects where we model the impact of member turnover and social capital losses on project productivity. By examining social capital losses we attempt to provide a more nuanced analysis of member turnover. In this context social capital is modelled from a social network perspective where the loss of more central members has more impact. We find that only a small proportion of WikiProjects are in a relatively healthy state with low levels of membership turnover and social capital losses. The results show that the relationship between social capital losses and project performance is U-shaped, and that member withdrawal has significant negative effect on project outcomes. The results also support the mediation of turnover rate and network density on the curvilinear relationship.

💡 Research Summary

This paper investigates how member turnover and the associated loss of social capital affect the productivity of online collaborative groups, using Wikipedia WikiProjects as a case study. The authors assembled a longitudinal dataset covering 891 WikiProjects from 2001 to 2013, dividing the observation period into 90‑day quarters. For each quarter they recorded member activity, edit counts, and edit longevity (a quality‑adjusted contribution metric computed with WikiTrust). Productivity is thus measured as the aggregate edit longevity of all contributions made by project members during a quarter.

Two primary independent variables are defined. “Social capital losses” (SC Losses) are quantified as the ratio of the average betweenness centrality of members who left the project in a given quarter to the average betweenness of all members. Betweenness captures the extent to which a member bridges structural holes; therefore, the departure of high‑betweenness users represents a larger erosion of network‑level resources. “Turnover rate” is the proportion of members who were active in the previous quarter but made no edits in the current quarter.

Network density is included as a moderating variable because a dense collaboration graph can provide redundant ties that mitigate the impact of losing central actors. The model also controls for project age, scope (number of articles tagged), size (active members), discussion activity (number of talk‑page topics), average tenure, controversy level (percentage of reverted edits), and tenure diversity (coefficient of variation of member tenure).

Three hypotheses guide the analysis: (Hₐ) higher turnover rates negatively affect productivity; (H_b) the relationship between social‑capital loss and productivity follows a U‑shaped (curvilinear) pattern; and (H_c) turnover rate and network density mediate the curvilinear relationship.

The authors employ mixed‑effects panel regression, incorporating a quadratic term for SC Losses to capture curvature. Results partially confirm Hₐ: turnover is generally associated with lower productivity, though the effect diminishes at very low turnover levels. H_b receives strong support: when social‑capital loss is low, productivity is high; as loss rises to a moderate level, productivity drops sharply; beyond a high loss threshold, productivity begins to recover, forming a U‑shaped curve. This suggests that the departure of core members initially disrupts coordination, but the community can adapt by re‑forming ties or attracting new contributors.

H_c is also validated. The negative slope of the SC Losses‑productivity curve is attenuated when turnover rates are higher, and especially when network density is high. Dense networks appear to buffer the detrimental effects of losing central actors by offering alternative pathways for information flow.

The study contributes several practical insights for managing online communities. First, retaining members who occupy structurally important positions is more critical than merely reducing overall churn. Second, fostering a dense interaction network—through design features that encourage frequent communication—can increase resilience to turnover. Third, the operationalization of social‑capital loss provides a quantitative health indicator that can be monitored over time.

Limitations include reliance on edit longevity as a proxy for quality, which may not capture all dimensions of contribution value, and the focus on WikiProjects, which may limit generalizability to other types of online collectives. Future work is suggested to explore alternative centrality measures, examine the role of departing members’ expertise, and model onboarding processes of new members to better understand how communities recover after substantial social‑capital loss.