Multilevel analysis of a measure of community prevention collaboration

Eric C. Brown, J. David Hawkins, Michael W. Arthur, Robert D. Abbott, M. Lee Van Horn

Research output: Contribution to journalArticlepeer-review

24 Scopus citations


This study assesses a measure of community-wide collaboration on prevention-specific activities (i.e., prevention collaboration) in context of the theory of community change used in the Communities That Care prevention system. Using data from a sample of 599 community leaders across 41 communities, we examined the measure with regard to its factor structure, associations with other concurrent community-level measures, and prediction by individual- and community-level characteristics. Results of multilevel confirmatory factor analysis provide evidence for the construct validity of the measure and indicate significant (p < .05) associations with concurrent validity criteria. Female community leaders reported significantly higher levels of prevention collaboration and community leaders sampled from religious organizations reported lower levels of prevention collaboration than did their respective counterparts. Although no community-level characteristics were associated significantly with prevention collaboration, community clustering accounted for 20-28% of the total variation in the measure. Findings support the use of this measure in assessing the importance of collaboration in community-based prevention initiatives.

Original languageEnglish (US)
Pages (from-to)115-126
Number of pages12
JournalAmerican Journal of Community Psychology
Issue number1-2
StatePublished - Mar 2008
Externally publishedYes


  • Collaboration
  • Communities that care
  • Community-based intervention
  • Prevention

ASJC Scopus subject areas

  • Health(social science)
  • Applied Psychology
  • Public Health, Environmental and Occupational Health


Dive into the research topics of 'Multilevel analysis of a measure of community prevention collaboration'. Together they form a unique fingerprint.

Cite this