Organizing without formal organization:

Organizing without formal organization: Group Identification, Goal Setting and Social Modeling in Directing Online Production • Haiyi Zhu • Robert E. Kraut • Aniket Kittur

Challenges for online communities

Managing volunteers in online communities How to direct volunteers to work on unattractive but important tasks?

Managing volunteers in online communities Conventional techniques are not effective in online communities • lack of employment contracts • weak external incentives • weak interpersonal bonds • impoverished communication • large size • high turnover

Other means to manage volunteers • Perception of belonging to a group • Motivate people to work for the group • Does not specify which particular tasks to work on

Other means to manage volunteers • Specify which particular tasks need work • Is not sufficient to motivate people

Managing volunteers in online communities • Group identification can motivate people to work for the group • Direction setting can channel their group-oriented motivations toward specific tasks.

Direction setting & group identification

Hypotheses H1. Highlighting tasks important to the group through goal setting directs people’ efforts towards these tasks. H2. The effect is stronger for self-identified group members. H3. Exposure to prototypical group members should lead people to perform more group-valued behaviors that prototypical members engage in. H4. The effect is stronger for self-identified group members.

Study platform: Wikiprojects • Wikiprojects • Self-organized subgroups in Wikipedia around topics

Collaborations of the Week (COTW) A project activity which designates one or two articles to improve in a defined period (e.g., a week or a month)

Collaborations of the Week (COTW) A COTW announcement in a project page

Collaborations of the Week (COTW) A COTW announcement in a project page An example template identifying an article as a COTW

Data collection • 618 COTWs until Mar.2008 • 26 projects • On average, each project consisted of 26,553 articles (median = 4,632) and 471 members (median = 255.5)

Test hypotheses H1. Highlighting tasks important to the group through goal setting directs people’ efforts towards these tasks. H2. The effect is stronger for self-identified group members. H3. Exposure to prototypical group members should lead people to perform more group-valued behaviors that prototypical members engage in. H4. The effect is stronger for self-identified group members.

Group goal setting Collaborations of the Week is a group goal setting mechanism that highlights tasks crucial for the Wikiproject • Some projects explicitly claim that the goal of collaborations is to "fill the gap" of the Wikiproject. • Collaboration targets are typically articles rated as high importance but having low quality.

Group identification Operationalized people who identified with the group as those who edited the project member lists.

Analysis • Included in our sample editors who were aware of the event. • Examined whether these editors’ contributions increased during the goal period and whether the increase was larger for editors who self-identified as group members. • Conducted analysis on the editor-period level.

The revision counts on the COTW target articles Group goal + group identification Non self-identified editors Self-identified group members Pre-collaboration period Collaborationperiod Post-collaboration period The interaction effect of goal periods and group identification is highly significant, p < 0.001

Test hypotheses H1. Highlighting tasks important to the group through goal setting directs people’ efforts towards these tasks. H2. The effect is stronger for self-identified group members. H3. Exposure to prototypical group members should lead people to perform more group-valued behaviors that prototypical members engage in. H4. The effect is stronger for self-identified group members.

Social modeling • Collaborations of the Week is a social event. • Provides opportunities for volunteers to discuss with each other, and potentially to influence each other.

Prototypical members • We operationalized prototypical members as those who were the heaviest contributors in project pages and at the same time participated in COTWs. • We selected regular editors as non-prototypical members who also participated in COTWs at least once in the history.

Group valued behaviors Assessment Adding or changing the rating of an article on its Wikiproject template. Anti-vandalism Reverting the vandalized version to a previous state.

Analysis • Calculated the correlation between regular members’ behaviors with the behaviors of prototypical members, considering 1) Group identification: whether the regular editors identified themselves as project members or not 2) COTW participation: whether the regular editors participated in COTWs in the given period or not.

Social modeling + group identification The correlation of regular editors’assessment with the prototypical members’assessment The interaction effect of COTW participation and group identification is significant, p < 0.01

Social modeling + group identification The correlation of regular editors’anti-vandalism with the prototypical members’anti-vandalism The interaction effect of COTW participation and group identification is not significant

Group valued behaviors Assessment Adding or changing the rating of an article on its project template. (Specific to the particular group) Anti-vandalism Reverting the vandalized version to a previous state. (Not strongly identified by any group)

Findings 1. Highlighting tasks important to the group through goal setting directs people’ efforts towards these tasks. 2. The effect is stronger for self-identified group members. 3. Exposure to prototypical group members should lead people to perform more group-valued behaviors that prototypical members engage in. 4. The effect is stronger for self-identified group members when the behaviors are specific to the group.

Managerial implications • Generalize to other kinds of online communities and offline organizations. • Organizations emphasizing creative work • Open source projects

Managerial implications • Limits to the applicability of group goal • setting if these tasks involve high coordination costs. • “Adding manpower to a late software project makes it later.”- Brooks aptly

Managerial implications • Social modeling is effective in drawing in peripheral members and training them in a wide range of subtle behaviors. • Encourage desired behaviors from prototypicalmembers. • Providesocial opportunities (e.g., communication channels or collaboration tasks).

Thank you! • Haiyi Zhu • haiyiz@cs.cmu.edu • Robert E. Kraut • robert.kraut@cmu.edu • AniketKittur • nkittur@cs.cmu.edu

Organizing without formal organization: