J Community Psychol. 2020;48:2086–2107. wileyonlinelibrary.com/journal/jcop 2086 | © 2020 Wiley Periodicals LLC Received: 1 April 2020 | Revised: 8 June 2020 | Accepted: 12 June 2020 DOI: 10.1002/jcop.22408 RESEARCH ARTICLE Improving understanding of how mentoring works: Measuring multiple intervention processes Patrick H. Tolan 1 | Heather L. McDaniel 1 | Malachi Richardson 1 | Nora Arkin 2 | Julia Augenstern 1 | David L. DuBois 3 1 Youth‐Nex Center, Curry School of Education and Human Development, University of Virginia, Charlottesville, Virginia 2 Department of Psychiatry, Harvard Medical School, Harvard University, Boston, Massachusetts 3 Department of Community Health Sciences, School of Public Health, University of Illinois at Chicago, Chicago, Illinois Correspondence Patrick H. Tolan, Youth‐Nex Center, Curry School of Education and Human Development, University of Virginia, Charlottesville, Virginia. Email: pht6t@virginia.edu Abstract The objective is to develop and test dimensionality and validity of a new measure of five mentoring intervention processes drawn from emphases across prior mentoring efforts that might be the basis for effects: identification with the mentor, social and emotional support, teaching and education, and advocacy, and shared time and activity. Partial validation of value was drawn from moderation role in a meta‐analysis of mentoring (Tolan, Henry, Lovegrove, Schoeny, & Nichols, 2014. J. Exp. Criminol., 10, 179–201). Data collected of candidate items completed by 740 mentors and 302 mentees (mean age = 16.59 years) were subjected to confirmatory factor analyses based on the five theorized dimensions. Scales derived from best‐fitting solutions were examined for convergent validity in relation to other indicators of mentoring relationship qualities. A bifactor model with five specific factors was supported for adult mentor reports while a simplified single factor model fit youth reports. Correlations between derived scale scores and validity indicators of mentoring relationship qualities were consistent with expectations. Multiple activities that comprise mentoring can be identified in adult men- tor reports but not youth reports. This scale provides a promising basis for testing how variation in emphasis