Quantifying spillovers in open source content production: evidence from Wikipedia

Using detailed edit-level data over eight years across a large number of articles on Wikipedia, we find evidence for a positive spillover effect in editing activity. Cumulative past contributions, embodied by the current article length, lead to significantly more editing activity, while controlling for a host of factors such as popularity of the topic and platform-level growth trends. The magnitude of the externality is significant, and growth in editing activity on the average article would have been about 50% lower in its absence. The spillover operates through an increase in the number of contributing users, whereas the length of contributions remains unchanged. Edits triggered by spillovers involve only marginally more deletion and replacement of content than the average edit, suggesting that past contributions do inspire the creation of new content rather than corrections of past mistakes. Roughly 75% of the spillover is due to new rather than returning users contributing content.