Matching Strategies Informed by Participant Characteristics
Evidence Rating for this Practice:
Description of Practice:
Matching strategies informed by participant characteristics involves the intentional use of information about mentor and mentee characteristics to inform the mentor-mentee matching process. Characteristics considered can vary and may include (but are not limited to) gender, race, ethnicity, disability, social class, personality, interests/hobbies, goals, strengths, areas of identified need, and life experiences of the mentor and mentee. Both similarities and differences in these types of characteristics between mentors and mentees may be used to inform the matching process (Pryce et al., 2014). Examples of specific strategies to facilitate or support use of information about mentor and mentee information in the matching process include (but, again, are not limited to) the use of matching criteria or guidelines, surveying mentors and mentees to assess their interests or preferences, soliciting and incorporating youth input into the matching decision, engaging parents in the matching process (e.g. pre-match meetings with parents to gather information on youth characteristics), use of technology (e.g., algorithms based on compatibility scores; Pryce et al., 2014), and staff training related to matching procedures. This practice is consistent with recommended benchmarks for Matching and Initiating standards in the Elements of Effective Practice for MentoringTM (Garringer, Kupersmidt, Rhodes, Stelter, & Tai, 2015). However, it is distinguished from other matching strategies that do not focus on use of information about the characteristics of mentors and mentees (e.g., compatibility of availability schedules, geographic proximity) as well as those that are geared toward initial development of the mentoring relationship (e.g., program-facilitated initial meeting between mentor and mentee). It also does not include ongoing match support practices that involve assessing match compatibility for purposes such as determining the potential need for match closure or re-matching.
The primary goal of the practice is to use information about the characteristics of mentors and mentees to inform matching decisions.
Target Population/Eligibility of Target Sites:
This practice is potentially applicable to all forms of mentoring and the full range of youth who may be served by mentoring programs.
Theory and Background Research:
A variety of theoretical perspectives, such as attachment theory and similarity-attraction theory, suggest that both complementary and similar characteristics between mentors and youth have the potential to contribute to more successful mentoring relationships (Liang, Bogat, & Duffy, 2014; Pryce, Kelly, & Guidone, 2014). Similarity-attraction theory, for example, suggests that the more similar a youth-mentor perceive each other to be, the more they like each other (Ensher & Murphy, 1997).
Research relevant to this practice (excluding the studies that are the focus of this evidence review) includes studies that have examined whether similarity between mentees and mentors along various dimensions is associated with differences in relationship or mentee outcomes. In a study of 45 BBBS affiliates with data in the Agency Information Management system, it was found that 78% of programs considered mentor and mentee characteristics (e.g. shared interests) when making matches (Kupersmidt et al., 2016). However, this matching practice did not predict match longevity. Similarly, research not indicated differential relationship or mentee outcomes in relation to whether mentors and mentees are of the same gender or share the same racial or ethnic group (Liang et al., 2014, Kanchewa et al., 2014; DuBois et al., 2002; Grossman & Rhodes, 2002; Park et al., 2016; Sanchez et al., 2014).
Corresponding Elements of Effective Practice:
This practice is most relevant to the area of Matching within the Elements of Effective Practice.
Evidence Classification: Insufficient Research
DuBois et al (2011) examined the practice of matching informed by participant characteristics in a meta-analysis of 73 evaluations of youth mentoring programs published between 1999 and 2010. (Meta-analysis is a technique for synthesizing and summarizing findings across evaluations of similar, but not identical research studies. One question often addressed in meta-analyses is whether the effects of a certain kind of program, like youth mentoring, differ based on the specific types of practices that are utilized. A correlation between the use of a practice and program effectiveness does not, generally speaking, provide definitive evidence of a causal effect of that practice; one reason for this is that programs that do or do not utilize a particular practice may differ in other important ways, not all of which can be controlled for statistically.) Programs or interventions were categorized as mentoring programs if their goal was to promote positive youth outcomes using “specific non-parental adults (or older youth) who are acting in a nonprofessional helping capacity”; the review thus considered evaluations of programs with a wide variety of formats and settings. Analyses were based on 82 independent samples because some studies contributed more than one sample. To be included, the evaluations needed to utilize a two-group randomized control or quasi-experimental design. By comparing changes in outcomes for mentored youth to non-mentored youth, such designs help to avoid the potential error of attributing changes in outcomes that occur due to normal development to effects of the mentoring program. Evaluations of programs where mentoring was provided in combination with other interventions were not included in the meta-analysis, unless the effect of the mentoring component could be isolated.
Mentoring program effect sizes were estimated for youth outcomes that could fall within any of the six domains: academic/school, attitudinal/motivation, social/relational, psychological/emotions, conduct problems, and physical health. All effect sizes were based on outcomes assessed at the end of the program. Where pretest data were available (53 of the samples), they were subtracted from posttest outcomes to adjust for potential differences between mentoring and comparison groups at baseline. Analyses were conducted under the assumption of a random effects model. Effect sizes were computed as standardized mean differences (specifically Hedge’s g) and were coded so that positive values for outcomes indicated effects in the desired direction (e.g. less delinquent behavior). The meta-analysis included a comparison of effect sizes for programs that used similarity of interests to inform mentor-youth matching (8 samples) and those for which this practice was not evident (74 samples). Prior to testing for differences in effect size in relation to this and other program characteristics, potential effects of study quality on effect sizes were assessed and effect sizes were residualized on those variables to control for their influence. Differences in effect sizes in relation to program characteristics, such as matching based on interests, that reached (p < .05) or approached (p < .10) significance were reported.
Program Effect Size
DuBois et al (2011) found that mentoring programs that used interest information to inform matching of mentors and mentees had larger estimated effects on youth outcomes than those that did not include the practice. Programs that had matching based on interests had an estimated effect size of .41 (95% confidence range of .25 to .57), whereas those without evidence of the practice had an estimated effect size of .20 (95% confidence range of .17 to .23). This difference was found to be statistically significant.
A stepwise regression analysis was used to determine if using interest information to inform matching earned entry into a best-fitting model for predicting estimated effect size in which all program characteristics tested in the meta-analysis were considered. Matching informed by information on mentee and mentor interests earned entry into the best-fitting model, indicating that this program practice was associated with stronger estimated program effects on youth outcomes independent of its overlap with the other variables that also earned entry.
Study 2: Insufficient Research
Ensher & Murphy (1997) examined the practice of matching informed by participant characteristics in an experimental study among youth interns participating in an 8-week summer job training program within a West coast media organization. Participating youth, who were characterized as low opportunity (based on socioeconomic status) and high potential (based on active involvement with a community agency), were recruited from a pool of approximately 16 community service agencies. The goals of the mentoring program were to provide interns with an adult role model who offers friendship and guidance in an informal manner, is empathetic of someone in a first or second job experience, and is willing to listen and be accessible. Mentors were employees in a variety of departments and organizational levels who volunteered to mentor interns and were committed for the 8-week program duration. Mentors were expected to meet with their mentees once a week. Mentors attended a 2-hour training session that provided information regarding their role as a mentor and provided them with a realistic preview of their experience.
One hundred and four interns participated in the program at the time of the study. Interns ranged in age from 16 to 22 and were ethnically and racially diverse (39% Latino, 25% African-American, 24% Asian, 3% Native-American, 4% multi-racial, and 4% ‘other’). Mentors were predominantly Caucasian (45%); 28 percent were African-American, 15 percent were Latino, 4 percent were Asian, 3 percent were multi-racial, 2 percent were Native-American, and 3 percent categorized themselves as ‘‘other.’’ The average mentor was a college graduate and 62% were professionals or managers. Same-gender mentor matching resulted in 43 male and 61 female mentor and intern pairs. Youth were randomly assigned to one of two types of mentor-race pairings – same-race mentor or different-race mentor.
Data were collected from interns and their mentors during their first and last weeks of employment (i.e., baseline and post-intervention, respectively); 79 percent of participating youth and 67 percent of mentors completed both surveys. Twenty-six same-race and 50 different-race pairs completed both assessments. Caucasian mentors with minority interns made up 66 percent (n=33) of the different-race mentoring pairs, whereas Black mentors with different-race interns made up 10 percent (n=5). The racial pairing of mentors and interns resulted in 12 male pairs and 14 female pairs of the same race and 19 male pairs and 31 female pairs of different race. A series of t tests on each of the outcomes of interest revealed no significant differences between interns with Caucasian mentors and those with non-Caucasian mentors.
Surveys collected data on liking between intern and mentor, type of mentoring support, perceived similarity between intern and mentor, likelihood of maintaining relationship, frequency of contact, and satisfaction with mentor. Liking between mentor and mentee was measured using the following two items: ”I like my mentor/protégé very much as a person” and ‘‘I think my mentor/protégé would make a good friend”; responses were scaled from strongly disagree (1) to strongly agree (5) and summed to form a composite score. A modified version of the Mentor Functions Scales (Noe, 1988) was used to assess the extent to which mentors provided psychosocial and instrumental/career support. A 5-item scale was used to measure perceived similarity of mentor/mentee. Example questions are: ”My mentor/protégé and I see things in much the same way,” and ”My mentor/protégé was similar in terms of our outlook, perspective, and values”; responses were scaled from strongly disagree (1) to strongly agree (5) and summed to form a composite score. Frequency of contact was assessed using the following question: ”On average, how many hours a week have you had contact with your mentor/protégé since the first time you met your mentor/protégé?” Possible response categories were: ”Less than 1 hour a week”, ”1–3 hours a week”, ”4–5 hours a week”, ”6–8 hours a week”, and ”More than 8 hours a week”. Likelihood of maintaining the relationship was assessed using the following question: ”How likely do you think is it that you will stay in contact with your mentor after the program is over?” Responses to this item ranged from very unlikely (1) to very likely (5). In addition, mentors and interns were asked to list any reasons that they felt contributed to whether they would stay in touch with one another. Finally, interns also reported on their satisfaction with their mentor using the following three items: ”I effectively utilized my mentor to help me develop,” ”My mentor met my expectations,” and ”I felt satisfied with my mentor”; responses were scaled from strongly disagree (1) to strongly agree (5) and summed to form a composite score.
Analysis of variance (ANOVA) was used to test whether mentor-reported liking between mentor and intern and intern-reported receipt of mentor support functions differed across same-race and different-race pairings of interns and mentors. Furthermore, all mentor-reported outcome variables were entered into a multivariate ANOVA to determine whether there was an overall effect of race-pairing across all of these outcomes; based on a significant effect of race-pairing in this MANOVA, univariate ANOVAs were conducted to test for effects on each mentor-reported outcomes. Because the race-pairing effect for corresponding MANOVA for youth-report measures was not statistically significant, follow-up univariate ANOVAs were not conducted for these outcomes.
Study 3: Insufficient Research
Sowers et al. (2016) examined the practice of matching informed by participant characteristics in a randomized trial of a STEM mentoring program for students with disabilities. Eligible students were those in Grades 9 to 11 in an urban school district and had an Individual Education Plan or a 504 Plan. Students were enrolled across two waves. A total of 78 students were enrolled in the study; 70 completed all data assessments. Students were stratified across 3 disability groups (those with physical and sensory disabilities; specific learning disabilities, attention-deficit/hyperactivity disorder, and cognitive impairment; emotional, mental health, autism spectrum) and across 2 STEM engagement groups (those who participated in at least one STEM extracurricular activity in the past year and those who had not) and randomly assigned to one of three study conditions – mentor with a disability, mentor without a disability, or control. One parent or caregiver for each student also participated in the study.
Students in the two intervention groups (mentor with disability, mentor without disability) were matched with an adult mentor of the same gender. Mentors were recruited from STEM postsecondary education programs and community businesses. Interested individuals completed applications, which asked about the presence and type of disabilities experienced, as well as criminal history background checks and interviews. Matches were informed by student preferences related to STEM interest areas and mentor personality and an attempt was made to match students in the mentor with disability group a mentor who experienced similar disability-related challenges. Mentors were provided with a guidebook that delineated the discussions and activities in which they were expected to engage with their assigned student. Mentors in each intervention group (mentors with disabilities and mentors without disabilities) participated in separate group orientation meetings so as to maintain the integrity of the two mentor groups, during which they reviewed the guidebook and logistics regarding planning mentor-mentee meetings and activities, transportation, safety guidelines, and communication. Mentors also received training on providing psychological support and relationship building as well as disability-related information (e.g., accommodations available at school or work).
Mentors were asked to meet with students twice a month for six months; nine of these meetings were activities planned by the mentor and student and three of the meetings were group workshops planned by project staff and included a presentation about a STEM topic and interactive activity. During individual student-mentor meetings, mentors were expected to engage students in activities and discussions related to the following five topic areas and in doing so to discuss career development strategies related to each topic area at least once:
- How to choose a STEM career you may wish to pursue.
- How to prepare for a STEM career in high school.
- How to get into a STEM college program.
- How to complete a STEM college degree.
- How to get a STEM job.
Mentors also were asked to do each of the following activities with their mentees at least once:
- Arrange for a job shadow at the place of work of the mentor or another STEM professional.
- Visit a local STEM college program.
- Meet with a STEM organization or club and encourage the student to join it and/or others.
- Find STEM internship opportunities, review the application processes, and encourage the mentee to apply for one.
- Spend one meeting doing a relationship building fun activity that was not necessarily focused on STEM.
- Review the mentee’s high school transcript and create a plan for future high school STEM classes.
- Meet with the mentee’s family to share what they had done together, what the mentee had learned from these experiences, and the mentee’s future plans.
Project staff provided telephone support and coaching to mentors prior to and after each activity and collected information on the activities and discussions mentors engaged in with students during these calls using a mentoring fidelity checklist.
Participating students and their parents completed assessments at three time points – before random assignment (baseline) and approximately 6 months later (post-intervention) 10 months (follow-up) after the baseline assessment. Students completed a 5-item STEM Activity Knowledge Questionnaire, which assessed their knowledge of activities they can do to pursue a STEM field in high school, postsecondary education, and a job. Students responded to questions in writing or orally and their responses were scored based on how closely they matched answers on a scoring template; responses could receive a maximum of five points for each question. Students also completed the following measures:
- STEM Self-Efficacy Scale - An 8-item measure assessing the degree to which students believe they will do well in STEM classes in high school, get into college and do well in STEM college courses, get and do well in a STEM job, as well as their ability to successfully deal with difficulties they encounter in these situations (example questions are “I will do well in STEM classes in high school”, and “If I have trouble with an STEM class I will be able to figure out how to deal with it”).
- STEM Career Planning Confidence Scale - A 26-item measure assessing students’ confidence that they can do STEM-related career planning activities (e.g., “Make a list of STEM classes that you need to prepare for the STEM job you want to do”).
- Disability-Related Self-Efficacy Scale - An 8-item measure assessing the extent to which students believe they have the capabilities to achieve desired outcomes (not specific to STEM) made more difficult by their disability (e.g., “I do a good job at getting the help I need”).
- Career Planning Self-Efficacy Scale - A 13-item measure assessing students’ confidence in completing career choice making activities that were not specific to STEM (e.g., “Talk to a person already employed in a field I am interested in”).
Parents rated their confidence in their children’s ability to do STEM-related career planning activities using a parent version of the STEM Career Planning Confidence scale. Both the student and his or her parent reported on the youth’s level of engagement in STEM activities that were not for school credit (e.g. clubs, internships, after-school classes, etc.) between assessments points. Students and parents also completed a survey at post-intervention to assess their perceptions of the utility of and satisfaction with the program.
General linear mixed model regression analyses compared baseline outcome measures with the average of post-intervention and follow-up scores across youth in the intervention and control groups. To assess whether the effects of the intervention diminished after the end of the intervention, analyses also compared post-intervention measures with follow-up measures. Effect sizes were calculated.
Sowers et al. (2016) found no significant differences between students matched with mentors with disabilities and those assigned to mentors without disabilities on any of the outcomes measured.
External Validity Evidence:
Variations in the Practice
Each of the reviewed studies addresses the implications of whether mentors and mentees were paired so as to be similar to one another on a selected characteristic. The characteristics considered vary and include mentor and mentee interests (DuBois et al., 2011), disability status (Sowers et al., 2016), and racial or ethnic background (Ensher & Murphy, 1997). Available studies do not address utilization of information on other types of mentor and mentee characteristics (e.g., personality) or the potential effectiveness of strategies other than those focused on ensuring similarity or compatibility of the mentor and mentee on a single characteristic (e.g., training of staff to utilize information on a range of mentor and mentee characteristics to inform matching decisions). In sum, the studies reviewed examine only a limited portion of the full range of possible strategies that could be used for this practice.
Studies reviewed varied in the characteristics of youth involved in the evaluated programs. One study focused on youth with disabilities (Sowers et al., 2016), whereas another (Ensher & Murphy, 1997) focused on youth who were from low socioeconomic backgrounds but had high levels of community involvement. The DuBois et al. (2011) study had a broader scope and thus included studies of samples of youth that varied along dimensions such as age, gender, race/ethnicity, and socioeconomic status. Other than testing for differences in effects of the practice across male and female mentor/mentee pairings in one study (Ensher & Murphy, 1997), studies did not test for possible differences in effects of the practice across differing subgroups of youth. There are also too few studies available to make informed comparisons of findings across studies with respect to possible implications of youth characteristics for effectiveness of this practice.
Studies investigating the possible effects of matching informed by participant characteristics included mentors with varying characteristics, including age (adult and adolescent), disability status, and career fields or industries (e.g., STEM). Other than testing for differences in effects of the practice across male and female mentor/mentee pairings in one study (Ensher & Murphy, 1997), studies did not test for possible differences in effects of the practice across differing subgroups of mentors. There are also too few studies available to make informed comparisons of findings across studies with respect to possible implications of mentor characteristics for effectiveness of this practice.
Studies included in the DuBois et al. (2011) meta-analysis evaluated mentoring programs with varying structures and that were delivered in a range of different types of settings, including those taking place within the community and those occurring at school as well as those using either a one-on-one or group format. The program evaluated in the Ensher & Murphy (1997) study was delivered within the workplace and the one evaluated in the Sowers et al. (2016) study took place within a community setting. None of the studies tested for evidence of possible differences in effects of the practice on the basis of program characteristics. There are also too few studies available to make informed comparisons of findings across studies with respect to possible implications of program setting or structure.
The DuBois et al. (2011) meta-analysis investigated the potential effects of matching informed by participant characteristics on a range of youth outcomes, including delinquency, aggression, substance use, academic achievement, social skills, self-esteem, and obesity. Estimated effect sizes were collapsed across all categories of outcomes, however, when testing for potential effects of the practice. Sowers et al. (2016) assessed effects of the program on STEM related outcomes (e.g., STEM self-efficacy, career preparation, etc.) and Ensher & Murphy (1997) tested for impact on features of the mentoring relationship (e.g., liking between mentor and mentee, satisfaction with mentor, etc.). Overall, available findings provide only a limited basis for understanding potential effects of this practice on the range of outcomes of potential interest.
Resources Available to Support Implementation:
Resources to support implementation of matching strategies informed by participant characteristics can be found under the Resources section of this website. These include:
The ABCs of School Based Mentoring – This guidebook offers information that can be used at the individual school or district levels to inform the design and implementation of school-based mentoring program, including information on making matches.
Generic Mentoring Program and Procedure Manual – This resource provides a template for a mentoring program to create its own customized manual to guide both policies and day-to-day services, including guidelines related to developing matching policies and procedures.
DuBois, D. L., Portillo, N., Rhodes, J. E., Silverthorn, N., & Valentine, J. C. (2011). How effective are mentoring programs for youth? A systematic assessment of the evidence. Psychological Science in the Public Interest, 12, 57-91. https://doi.org/10.1177/1529100611414806
Ensher, E. A., & Murphy, S. E. (1997). Effects of race, gender, perceived similarity, and contact on mentor relationships. Journal of Vocational Behavior, 50, 460–481. https://doi.org/10.1006/jvbe.1996.1547
Sowers, J., Powers, L., Schmidt, J., Keller, T. E., Turner, A., Salazar, A., & Swank, P. R. (2016). A randomized trial of a science, technology, engineering, and mathematics mentoring program. Career Development and Transitions for Exceptional Individuals, 1-9. https://doi.org/10.1177/2165143416633426
Cavell, T. A. (n.d.). DuBois, D. L., Holloway, B. E., Valentine, J. C., & Cooper, H. (2002). Effectiveness of mentoring programs for youth: A meta-analytic review. American Journal of Community Psychology, 30, 157–197. https://doi.org/10.1023/A:1014628810714
Garringer, M., Kupersmidt, J., Rhodes, J., Stelter, R., & Tai, T. (2015). Elements of effective practice for mentoring (4th ed.). Boston, MA: MENTOR: The National Mentoring Partnership.
Grossman. J. B., & Rhodes, J. E. (2002). The test of time: Predictors and effects of duration in youth mentoring relationships. American Journal of Community Psychology, 30, 199–219. https://doi.org/10.1023/A:1014680827552
Kanchewa, S. S., Rhodes, J. E., Schwartz, S. E. O., & Olsho, L. E. W. (20114). An investigation of same- versus cross-gender matching for boys in formal school-based mentoring programs. Applied Developmental Science, 18, 31–45. http://doi.org/10.1080/10888691.2014.876251
Kupersmidt J. B., Stump K. N., Stelter R. L., & Rhodes J. E. (2017). Mentoring program practices as predictors of match longevity. Journal of Community Psychology, 45, 630–645. https://doi.org/10.1002/jcop.21883
Liang, B., Bogat, A., & Duffy, N. (2014). Gender in mentoring relationships. In D. L. DuBois & M. J. Karcher (Eds.), Handbook of youth mentoring (2nd ed., pp. 159–173). Thousand Oaks, CA: SAGE
Noe, R. A. (1988). An investigation of the determinants of successful assigned mentoring relationships. Personnel Psychology, 41, 457–479. https://doi.org/10.1111/j.1744-6570.1988.tb00638.x
Park, H., Yoon, J., & Crosby, S. D. (2016). A pilot study of Big Brothers Big Sisters programs and youth development: An application of critical race theory. Children and Youth Services Review, 61, 83-89. https://doi.org/10.1016/j.childyouth.2015.12.010
Pryce, J., Kelly, M. S., & Guidone, S. R. (2014). Mentor and youth matching. In D. L. DuBois & M. J. Karcher (Eds.), Handbook of youth mentoring (2nd ed., pp. 427-438). Thousand Oaks, CA: SAGE
Raposa, E. B., Rhodes, J. E., & Herrera, C. (2016). The impact of youth risk on mentoring relationship quality: Do mentor characteristics matter? American Journal of Community Psychology, 57: 320–329 https://doi.org/10.1002/ajcp.12057
Sanchez, B., Colon-Torres, Y., Feuer, R., Roundfield, K. E., & Berardi, L. (2014). Race, ethnicity, and culture in mentoring relationships. In D. L. DuBois & M. J. Karcher (Eds.), Handbook of youth mentoring (2nd ed., pp. 145–158). Thousand Oaks, CA: SAGE.