Pre-Match Mentor Training

Evidence Classification

Insufficient Evidence (Inadequate Design Quality for Assessing Effects of Practice; Limited or Inconsistent Outcome Evidence)

Evaluation Methodology

Bernstein and colleagues (2009) examined correlates of pre-match training for mentors as part of a randomized control evaluation of Education Department (ED) funded school-based mentoring programs. This evaluation included 1,272 students from 32 purposively selected ED funded school-based mentoring programs. A majority of the programs were operated by non-profit/community-based organizations (66 percent), with an average of 6 years of experience with school-based mentoring programs. Both the 32 program sites and 974 mentors from these programs were surveyed about various characteristics of program delivery, including training and support for mentors, characteristics of mentors, and mentor/student relationship duration and activities. Almost three quarters (72 percent) of the mentors were female. More than half (56 percent) were between the ages of 23 and 64 years. An additional 23 percent were between 19 and 22 years. Among the remaining, 17 percent were 18 years and younger and 3 percent were above 65 years old. Mentors came from a variety of racial and ethnic backgrounds with approximately two thirds being white and twenty-nine percent identifying Black or African American. Approximately one-half (49 percent) of the mentors were employed; 39 percent were in school full-time, primarily in college. Eighty-two percent of the mentors had completed at least some college or other form of post-secondary training.

On the surveys that programs completed, they responded to the following 3 items related to pre match training: “Are mentors in your school-based mentoring program required to participate in pre-match training/orientation?”; “If yes, what does that pre-match training/orientation for mentors include?” (A list was provided and they could check all that apply); and “In your school-based mentoring program how many hours of pre-match training/orientation must each mentor receive before he or she is able to meet with his or her mentee?” Mentors answered the following survey questions relevant to pre match training: “What kind of orientation and training did you receive before you were matched with your student?” (A list was provided and they could check all that apply); and “About how many hours of training or orientation did you receive before you met with your student for the first time?” Associations between average hours of pre-match mentor training provided to mentors (based on average responses of mentors in the program) and estimates of program effects from the random assignment portion of the evaluation on student outcomes were examined using both bivariate and multivariate specifications. Student outcomes were assessed in the domains of domains of interpersonal relationships and personal responsibility, academic achievement and engagement, and high-risk or delinquent behavior and were derived from both student self-report and school records.

Evaluation Outcomes

Scholastic Efficacy and School Bonding
Bernstein et al. (2009) found that mentor participation in pre-match training was not significantly associated with student-reported level of scholastic efficacy and school bonding.

Future Orientation
Bernstein et al. (2009) found that mentor participation in pre-match training was not significantly associated with students’ self reported future orientation.

Pro-social Behaviors
Bernstein et al. (2009) found that mentor participation in pre-match training was not significantly correlated with student-reported level of involvement in pro-social behaviors (interpersonal relationships, personal responsibility and community involvement).

Delinquency (Student Survey)
Bernstein et al. (2009) found that mentor participation in pre-match training was not significantly correlated with students’ self reported engagement in delinquent behaviors.

Misconduct (Student Survey)
Bernstein et al. (2009) found that mentor participation in pre-match training was not significantly correlated with student-reported level of misconduct.

Math Grades
Bernstein et al. (2009) found that mentor participation in pre-match training was not significantly correlated with students’ math grades obtained from school records.

English Language Arts Grades
Bernstein et al. (2009) found that mentor participation in pre-match training was not significantly correlated with students’ English language arts grades obtained from school records.

Any Delinquency (School Records)
Bernstein et al. (2009) found that mentor participation in pre-match training was not significantly correlated with students’ involvement in delinquent behaviors (violence, drug related infractions and truancy) obtained from school records.

Any Misconduct (School Records)
Bernstein et al. (2009) found that mentor participation in pre-match training was not significantly correlated with students’ engagement in misconduct obtained from school records.

Overall Absenteeism Rates
Bernstein et al. (2009) found that mentor participation in pre-match training was not significantly correlated with students’ attendance rates (including excused and unexcused absences) obtained from school records.

Evidence Classification

Insufficient Evidence (Inadequate Design for Assessing Effects of the Practice; Lack of Sufficient Information on Fidelity of Practice Implementation)

Evaluation Methodology

Herrera and colleagues (2000) examined correlates of pre-match training for mentors as part of a comparative study of mentoring relationships in one-to-one school-based (SB) and community-based (CB) mentoring programs. Mentors were randomly selected to participate in the study from lists of current mentors provided by 98 programs. Twenty-five minute phone interviews were conducted to administer a survey to 669 mentors who were involved either in a CB mentoring program (346 mentors from 29 programs) or SB mentoring program (323 mentors from 35 programs). The survey included questions on mentor demographic characteristics, youth demographic and risk characteristics, program infrastructure and matching variables. Mentors in the SB programs were equally divided among 3 age groups – 21 or under, 22 to 49 and older adults, whereas mentors in the CB programs were overwhelmingly between the ages of 22 and 49. Although mentors from both types of programs were primarily White, those volunteering in SB programs were more likely to have ethnic minority backgrounds than those in CB programs.

Mentors were asked how much mentor orientation or training they had received from their programs before they started meeting with their mentees and were give 4 response options: none, less than 2 hours, between 2 and 6 hours, and more than 6 hours. They also provided ratings of mentoring relationship quality. Regression analyses were conducted to assess whether amount of reported pre match training was related to different measures of mentoring relationship quality. Analyses controlled statistically for demographic and other measured characteristics of youth and mentors, characteristics of mentors and youth (e.g., mentor/youth gender and mentor age and ethnicity) as well as measures of the ways in which pairs spend time together (e.g., time spent on the phone) and other program-related factors (including mentor report of amount of post-match training/contact with staff, community vs. school-based program, and required commitment – short-term intensive, short-term non-intensive, or long-term). Analyses were conducted at the individual mentor, rather than program level and adjustments to standard errors for clustering/non-independence of mentors within programs were not implemented.

Evaluation Outcomes

Mentor-reported Feelings of Closeness Toward Mentee
Mentors’ reports of the amount of pre-match training or orientation they had received were positively and significantly associated with their reported feelings of closeness toward their mentees. Mentors who attended fewer than two hours of pre match orientation or training reported the lowest levels of closeness, whereas mentors with the highest levels of closeness reported attending 6 hours or more of orientation or pre match training.

Number of Hours Spent with Mentees in Social Mentoring Activities
Mentor reports of the amount of pre-match training or orientation received were positively and significantly associated with mentor-reports of time spent in social mentoring activities with their mentees.

Additional Findings
When comparing mentors in school- and community-based programs, mentor reports of the amount of pre-match training/orientation received were found to exhibit similar associations with mentor reports of match relationship quality. In both instances, the highest levels of closeness were reported by mentors attending 6 or more hours of pre-match training or orientation.

Structural equation modeling analyses showed that reported amount of pre-match training was indirectly related to mentor reports of relationship closeness and supportiveness. In particular, mentors who reported receiving more pre-match training also tended to report spending more time with their mentees and to engage I social activities with them; these measures, in turn, were predictive of mentor reports of greater closeness and supportiveness in their match relationships. In interpreting this finding, the report authors noted that pre-match training may establish the importance of spending time with one’s mentee and provide mentors with insight into the value of building a relationship through the activities they engage in their mentees (paraphrased from Herrera et al., 2000, p. 31). In other research, Parra et al. (2002) similarly found support for a pathway linking mentor reports of the quality of pre-match training received in a Big Brothers Big Sisters community-based mentoring program were linked indirectly to mentor reports of greater relationship closeness through measures that included their reported levels of contact with mentees and engagement with them in program-relevant activities.

Evidence Classification

Insufficient Evidence (Inadequate Design Quality for Assessing Effects of the Practice; Limited or Inconsistent Outcome Evidence)

Evaluation Methodology

Herrera and colleagues (2007) examined correlates of pre-match training for mentors as part of a randomized control evaluation of the Big Brothers Big Sisters of America (BBBSA) school-based mentoring program. This evaluation included 1,139 youth in grades 4 to 9 and programs operated by 10 BBBSA affiliates. A total of 554 volunteers, who were assigned to serve as Big Brothers or Big Sisters (“Bigs”) for the youth assigned to the treatment group, completed baseline surveys at the start of the study. Almost three quarters (72 percent) of these Bigs were female. Approximately half of the Bigs (48%) were 18 years old or younger (i.e., high school students) and an additional 17 percent were 19 to 24 years old (predominately college students). The remaining Bigs were adults ages 25 and older. Approximately one fourth of the Bigs (23%) belonged to a racial or ethnic minority group, African-American being the most common of these (8% of all Bigs). Measures were completed by Big and other informants (teachers and youth) on three occasions: Baseline: Beginning of First School Year, First Follow-Up: End of First School Year, and Second Follow-Up: Late Fall of Second School Year. Further details on the study design and sample can be found in Herrera et al.(2007).

On the survey that Bigs completed at the end of the first school year, they responded to the following item: “How much time was spent in individual training before the start of your match?” A question about group training was also asked but did not differentiate between training completed prior to and after the start of the match. Bivariate correlations were computed between reported amount of individual pre-match training received and two outcomes: mentor-reported feelings of closeness toward his or her mentee at the second follow-up and whether or not the mentoring relationship continued into a second school year as reported by program staff. Analyses were conducted at the individual match, rather than program level and adjustments to standard errors for clustering/non-independence of match data within programs were not implemented.

Evaluation Outcomes

Mentor-reported Feelings of Closeness Toward Mentee
Mentor reports of the amount of individual pre-match training received had a significant positive correlation with mentor-reported feelings of closeness toward their mentees at the second follow-up assessment of the study. The size of this association was between levels typically suggested to be indicative of a small- or medium-sized relationship between variables.

Mentoring Relationship Continued into a Second School Year
Mentor reports of the amount of individual pre-match training received also had a significant positive correlation with continuation of the mentoring relationship into the second year as indicated by staff reports. The size of this association was closest to the level typically suggested to be indicative of a small- or medium-sized relationship between variables.

Evidence Classification

Promising

Evaluation Methodology

Herrera and colleagues (2013) examined correlates of pre-match training for mentors as part of an evaluation of seven one-to-one, volunteer-based mentoring programs, five of which were BBBSA community-based mentoring programs. This evaluation included 1,310 youth who ranged in age from 8- to 15-years old and were from a variety of ethnic backgrounds (57% were ethnic minorities). About two-thirds of the youth came from single-parent homes and about two fifths were from low-income households (i.e., annual incomes below $20,000). In addition, nearly three quarters (71%) faced some type of “individual-level” risk—for example, academic struggles, behavior problems or mental health concerns. When divided into groups based on individual and environmental risk levels, approximately one-fourth (26%) were designated as relatively high on both types of risk.

The volunteer mentors paired with youth were predominately White (82%) and had an average age of 32, although nearly one-quarter were college students (owing primarily to one of the programs being university-based with all college student mentors). Nearly half of the mentors (44%) reported having prior experience working with youth with behavioral, social or emotional difficulties and similar numbers reported experience working with youth from diverse cultural backgrounds (45%) and youth living in poverty (35%). Survey measures for the study were completed by youth, mentors, and parents both at baseline and at a 13-month follow-up (parents completed a survey at follow-up in only a portion of the evaluation). Further details on the study design and sample can be found in Herrera et al.(2013).

Whether or not mentors participated in pre-match training during the 13 month study period was determined based on records maintained by program staff as part of the study. At the beginning of the study, all but two of the seven programs reported having required pre-match training for mentors as defined by several criteria (e.g., offered before or within one month of the start of the match, included interactive or experiential component), with the length of the training being from 1 to 3 hours. By the end of the research, all programs required pre-match training of their mentors by the end of the research and the length of the training had increased to a range of between two to seven hours. Overall, approximately 1 in 2 mentors (56%) received pre-match training.

Analyses examined participation in pre-match training as a predictor of several different mentoring relationship outcomes. These included whether or not mentor-mentee meetings were frequent (three or more times per month) and whether or not the mentoring relationship lasted at least 12 months, both of which were assessed via program records; youth-report indices of mentoring relationship quality were also examined (i.e., levels of goal or growth focus and developmental centeredness, respectively, in the relationship and the youth’s feelings of closeness toward his or her mentor). All analyses controlled for the youth’s gender, age, race/ethnicity, and risk status group as well as for program.

Evaluation Outcomes

Mentor-Mentee Meeting Frequency
Herrera et al. (2013) found that mentor participation in pre-match training was a significant predictor of whether or not mentors met frequently with their mentors during the 13 month study period. Mentors who received pre-match training were more likely to meet with their mentee regularly (3 or more times during 70 percent or more of the months the match was active) in comparison to those who did not receive this type of training.

Mentoring Relationship Lasting 12 Months or Longer
Herrera et al. (2013) found that mentor participation in pre-match training was a significant predictor of whether or not the mentoring relationship lasted at least 12 months. Mentors who received pre-match training were more likely to have relationships lasting a year or more in comparison to those who did not receive this type of training.

Mentoring Relationship Quality: Growth/goal Focus
Herrera et al. (2013) found that mentor participation in pre-match training was a significant predictor of mentee-reported level of growth/goal focus in the mentoring relationship. Reported levels of growth/goal focus were higher among mentors who received pre-match training in comparison to those who did not receive this type of training.

Mentoring Relationship Quality: Youth Centeredness
Herrera et al. (2013) found that mentor participation in pre-match training was a significant predictor of mentee-reported level of youth centeredness in the mentoring relationship. Reported levels of youth centeredness in the mentoring relationship were higher in the case of youth whose mentors received pre-match training in comparison to those who did not receive this type of training.

Mentoring Relationship Quality: Closeness
Herrera et al. (2013) found that mentor participation in pre-match training was a significant predictor of mentee-reported feelings of closeness toward their mentors. Reported levels of closeness were higher in the case of youth whose mentors received pre-match training in comparison to those who did not receive this type of training.

Additional Findings
Herrera et al. (2013) also tested whether mentor participation in pre-match training as a predictor of mentoring relationship outcome varied according to youth individual and environmental risk status. No significant differences were found.

Evidence Classification

Insufficient Evidence (Inadequate Design for Assessing Effects of the Practice)

Evaluation Methodology

McClanahan (1998) examined correlates of pre-match training for mentors as part of a study examining the nature and content of the relationships that developed between mentors and mentees involved in the Hospital Youth Mentoring Program (HYMP). This program was designed to match students at risk of low academic achievement with mentors to help them complete school, expose them to a variety of health-related professions, give them career guidance and in some cases, provide them with actual work experience. Data from 13 program sites the was triangulated through three different means – surveys of mentors, youth and program coordinators, focus groups with mentors and youth, and archival documents. Surveys were collected from 269 of 465 mentors across the 13 program sites for a 58% response rate. The surveys asked about demographic information, reason for involvement with the program, type and amount of training received, meeting frequency, location of meetings, length of the match, type of activities the pair engaged in, and the amount of career guidance in the mentoring relationship.

The HYMP sites in the study served a total 515 students who were interested in a health care career, 380 of whom responded to a study questionnaire for a response rate of 74%. A third of the students had been participating for two or more years in the programs. Given the focus on urban youth, the majority of the students were from minority backgrounds (91%). Most of the student participants were Black or African American (53%), Hispanic (19%), or Asian American (13%). The majority of the participating students were girls (66%). Although the students ranged from 14 to 22 years in age, most were 16 to 18 years (73%) as the program targeted recruitment of high school youth. Almost quarter of the students came from families receiving public assistance and almost half of the students reported receiving reduced price or free lunch at school.

All mentors were hospital employees who volunteered to be mentors. About half the mentors were in supervisory positions at hospitals (48%) and a quarter of mentors (25%) had previously mentored in another program. Almost half of the volunteer mentors were White (51%), with the next largest racial/ethnic group being African American or Black (36%). The mentors were predominately female (73%) and most were middle aged adults (61% between 31- and 50-years old).

Whether or not the mentors participated in pre-match training was determined through mentor surveys and program coordinator interviews. Eighty two percent of the mentors reported receiving at least one hour of formal pre-match orientation or training. The number of hours of pre-match training offered to mentors across program sites ranged from 1 to 5. Bivariate correlations were computed between reported amount of pre-match training received and five relationship outcomes – length of the match, mentor support, whether the mentor took a developmental or prescriptive approach to the relationship, engagement in social activities with mentee and the amount of career mentoring taking place in the relationship. Analyses were conducted at the individual mentor, rather than program level and adjustments to standard errors for clustering/non-independence of data within program sites were not implemented.

Evaluation Outcomes

Length of the Match
McClanahan (1998) found that mentor reports of the amount of individual pre-match training received had a significant positive correlation with the duration of the mentoring relationship as indicated by staff reports and program documents. This association was small- to medium-size in magnitude.

Mentor Support
McClanahan (1998) found that there was a non-significant trend for mentor reports of the amount of individual pre-match training received to be associated positively with the level of support that mentees reported receiving from their mentors.

Mentor Developmental vs. Prescriptive Style
McClanahan (1998) found that mentor reports of the amount of individual pre-match training received was not related to mentee reports of whether the mentor took a developmental or prescriptive approach to the relationship.

Engage in Social Activities
McClanahan (1998) found that mentor reports of the amount of individual pre-match training received had a significant positive correlation with mentor reports of the level of social activities engaged in with mentees. This association was small- to medium-size in magnitude.

Engage in Career Mentoring
McClanahan (1998) found that mentor reports of the amount of individual pre-match training received had a significant positive correlation with mentor reports of the amount of career-related mentoring taking place in the relationship. This association was small in magnitude.

Evidence Classification

Insufficient Evidence (Inadequate Design for Assessing Effects of the Practice; Limited or Inconsistent Outcome Evidence)

Evaluation Methodology

Wheeler & DuBois (2009) examined correlates of pre-match training for mentors as part of a survey of the community-based mentoring programs of 131 BBBSA agencies between May and September, 2009, including 80 agencies with available data in BBBSA’s Agency Information Management (AIM) system (referred to here as “AIM agencies”). The response rate for all agencies was 34% and for AIM agencies was 81%. The survey completed by agencies included questions asking about various aspects of pre-match volunteer orientation/training within their community-based mentoring programs, including the materials used, whether it was offered on an individual or group basis, whether it included interactive exercises, and duration.

A dichotomous (yes/no) variable was created to indicate whether the agency reported providing at least one hour of pre-match orientation/training (52% of the agencies reported offering this level of training). A similar variable was created to indicate whether two or more hours of orientation/training was provided (18% of programs reported doing so). These variables were examined for correlations with the following indices of community-based (CB) match quality and duration: 6 and 12 month retention rates, average match length, and 3 month child and volunteer ratings on a measure of mentor-mentee relationship quality. These latter measures were available only for AIM agencies. Analyses included bivariate correlations as well as partial correlations that controlled for any of several assessed agency characteristics (e.g., % of female volunteers, % of children from low-income families) that had a correlation (significant or near-significant) with the match quality or duration measure being considered.

Evaluation Outcomes

6-month Match Retention Rate
Wheeler and DuBois (2009) found that there was a non-significant trend for agencies offering more than two hours of pre-match training to have higher percentages of matches lasting at least six months. This was no longer true, however, in analyses that included control variables.

12-month Match Retention Rate
Wheeler and DuBois (2009) found that there was a non-significant trend for agencies offering more than two hours of pre-match training to have higher percentages of matches lasting at least 12 months. This was no longer true, however, in analyses that included control variables.

Average Match Length
Wheeler and DuBois (2009) found that there was a non-significant trend for agencies offering more than two hours of pre-match training to have higher average match lengths. This was no longer true, however, in analyses that included control variables.

3-month Child-reported Match Relationship Quality
Wheeler and DuBois (2009) did not find evidence of an association between whether or not an agency provided 2 or more hours of pre-match training and mentee-reported match relationship quality at the 3-month point of the match.

3-month Volunteer-reported Match Relationship Quality
Wheeler and DuBois (2009) did not find evidence of an association between whether or not an agency provided 2 or more hours of pre-match training and volunteer-reported match relationship quality at the 3-month point of the match.

Additional Findings
Analyses focused on whether the agency reported offering at least one hour of pre-match training did not reveal evidence of any associations with the above measures of match duration or quality. It was found, however, that higher scores on a 2-item scale assessing whether pre-matching training was offered in a group format and included interactive exercise had a significant positive association with 3-month child-reported match relationship quality. This association remained significant in analyses that included control variables.

Evidence Classification

Insufficient Evidence (Inadequate Design for Assessing Effects of the Practice; Limited or Inconsistent Outcome Evidence)

Evaluation Methodology

Kupersmidt, Stump, and colleagues (2017) examined benchmark program practices and standards in the Elements of Effective Practice for Mentoring (MENTOR, 2009) as predictors of match length. The data were obtained from 45 Big Brothers Big Sisters agencies across the U.S. and included 29,708 matches. Participating agencies ranged in size from fewer than 30 matches to more than 13,000 represented in the available data. Matches were equally distributed between community-based and school- or site-based; the study limited their analyses to community-based matches. Surveys were completed by program staff, and archival data was obtained from the BBBSA Agency Information Management (AIM) system. Mentors were predominantly female (63%), White (66%) and non-Hispanic (91%), with an average age of 31.74. Mentees were also predominantly female (59%) but were more balanced in terms of race and ethnicity: 28% White and 17% Hispanic. Mentees were 10.78 years of age on average at the start of their match.

Match relationship length was calculated based on agency records of the start and end dates of each match; match length for those matches that were still open at the end of the study was calculated as the time between the start of the match and the last day of the study. Matches were further coded dichotomously as open or closed. Matches with a length of less than 12 months (the requested time of commitment from BBBS) were coded as prematurely closed. Matches of 24 months or longer in duration were coded as long-term matches (analyses using this variable excluded open matches of less than 24 months in duration). Program practices were assessed using a 31-item version of the Elements Quality Improvement Process (EQUIP) program self-assessment questionnaire completed by program staff. The items on the survey were combined into 22 items reflecting the 22 benchmark practices included in the EEPM. Items were coded as the individual practice being implemented or not, with a total benchmark score calculated as the sum of the implemented practices. Benchmark practices were further reduced to six standards: recruitment, screening, training, matching, monitoring & support, and closure. Programs were coded as implementing a standard if all of the benchmark practices for the standard were endorsed; a total standard score was calculated as the sum of the standards implemented (out of a possible 6). Further details on the study design and sample can be found in Kupersmidt, Stump et al. (2017).

Evaluation Outcomes

Match Length
Kupersmidt, Stump, et al. (2017) found that, after controlling for match cohort and youth age and sex, the training standard was significantly associated with greater match length.

Evidence Classification

Insufficient Evidence (Inadequate Design Quality for Assessing Effects of Practice; Limited or Inconsistent Outcome Evidence)
 

Evaluation Methodology

Kupersmidt, Stelter and colleagues (2009) examined the effectiveness of a web-based pre-match training program, Preparing for Mentoring. Fifty-one mentoring programs (representing a range of different program models) were assigned to blocks based on the length of their pre-match training. Programs were randomly assigned to the training group (25 programs, 80 mentors) or a wait-list control group (26 programs, 47 mentors). Mentors were 34.6 years of age on average (range 18-16) and were predominantly White (61.9%). Mentors came from a diverse range of educational backgrounds (30.4% high school diploma or GED, 13.0% associate’s degree or trade school certificate, 36.2% bachelor’s degree, and 20.3% postbaccalaureate education). The Training for Mentoring program used multimedia resources (custom animations, videos, and audio) and were programmed to be self-directed. The training was organized into three courses organized around the life cycle of the mentoring relationship. The first course, Building the Foundation, focused on the contemplation and initiation stages of mentoring, and included topics such as defining mentoring, common mentoring goals, and developing realistic expectations. The second course, Ethics and Safety, focused on the promotion of welfare and safety of mentees. The third course, Building and Maintaining the Relationship, focused on strategies to develop a close and positive long-term relationship.

All participating mentors completed surveys at baseline and repeated surveys post-training. Control group mentors completed a second survey equivalent to the Training group post-training survey approximately two weeks after the baseline survey. Training group mentors completed the post-training survey after completing the web-based training, with an additional set of Consumer Satisfaction questions about the training. Mentors completed 3 knowledge tests related to the 3 mentoring training courses, with correct responses summed across the three tests. The Mentor Motivations Questionnaire was used to asses volunteer motivation to mentor. Three of the subscales were used in analyses: civic responsibility, overinvolvement with children, and social reciprocation (the fourth subscale, self-enhancement, was excluded as most participants reported that the items were not reasons for their becoming a mentor). The Mentor Expectations Questionnaire assesses volunteer expectations and yields two subscales: unrealistically positive expectations (reported in findings below) and negative expectations. The Mentoring Roles Questionnaire assessed how frequently mentors expected to play various roles with their mentees and were organized into three groups: appropriate roles for mentors (e.g., role model), inappropriate roles for mentors (e.g., financial resource), and program-specific roles (e.g., tutor, advocate). Mentors also completed a scale assessing their attitudes about youth, with higher scores indicating more positive attitudes (e.g., youth are fun to be around). Mentors also completed a Mentor Preferences Scale developed for the study, in which they rated the perceived difficulty of mentoring children with 24 types of individual and family characteristics (e.g., youth 14-18, who are in trouble with the law). A self-efficacy and readiness scale (adapted from a previous version to include training-specific items) assessed how ready or prepared mentors were to mentor a youth. Program effectiveness for improving mentor knowledge and attitudes was tested in intent-to-treat analyses. Specifically, ANCOVA analyses were used to compare the Training and Control group, controlling for baseline levels of outcomes and mentor demographic characteristics (gender and educational status).

Evaluation Outcomes

Self-Efficacy to Mentor
Kupersmidt, Stelter et al. (2017) found that mentors in the Training group reported significantly higher self-efficacy scores than those in the Control group.

Mentoring Knowledge
Kupersmidt, Stelter et al. (2017) found that mentors in the Training group had significantly higher knowledge scores than those in the Control group.

Mentor Preferences
Kupersmidt, Stelter et al., (2017) found that participation in the online pre-match training was not significantly associated with mentors’ ratings of the perceived difficulty of mentoring children with various individual and family characteristics.

Positive Attitudes about Youth
Kupersmidt, Stelter et al., (2017) found that participation in the online pre-match training was not significantly associated with mentors’ reported attitudes about youth.

Mentor Motivation: Civic Responsibility
Kupersmidt, Stelter et al. (2017) found a group difference in the opposite direction to what was predicted: Control group mentors reported higher scores on civic responsibility motivation than Training group mentors. Although not statistically significant, the difference had a small-to-moderate effect size.

Mentor Motivation: Overinvolvement with Children
Kupersmidt, Stelter et al. (2017) found a group difference in the opposite direction to what was predicted: Training group mentors reported higher scores on overinvolvement with children motivation than Control group mentors. This difference was not statistically significant and had a small effect size.

Mentor Motivation: Social Reciprocation
Kupersmidt, Stelter et al., (2017) found that participation in the online pre-match training was not significantly associated with mentors’ social reciprocation motivation for mentoring.

Mentor Expectations: Unrealistically Positive
Kupersmidt, Stelter et al. (2017) found that mentors in the Training group reported significantly lower unrealistically positive expectations than those in the Control group.

Mentoring Roles: Inappropriate for Mentors
Kupersmidt, Stelter et al. (2017) found that mentors in the Training group reported significantly lower scores on the roles inappropriate for mentors scale than those in the Control group.

Mentoring Roles: Program Specific
Kupersmidt, Stelter et al. (2017) found a group difference in the opposite direction to what was predicted: Control group mentors reported higher scores on program-specific mentoring roles than Training group mentors. This difference was not statistically significant and had a small-to-moderate effect size.

Additional Findings
Mentors in the Training group reported significantly higher scores on the roles appropriate for mentors subscale than the Control group. Training group mentors also reported lower unrealistically negative expectation scores, but this difference was not significant and had a small effect size.

Evidence Classification

Insufficient Evidence (Limited or Inconsistent Outcome Evidence; Limited Fidelity of Implementation)

Evaluation Methodology

Stump and colleagues (2018) conducted secondary data analyses using archival data from 45 BBBS agencies (representing 25,252 mentoring relationships or “matches”) to examine the impact of mentoring and mentoring program practices on children of incarcerated parents (COIP). Participating agencies ranged in size from fewer than 15 matches in their archival data to more than 5,000 (mean program size of 561.16 matches). Overall, 19.36% of matches included a youth with a parent who had been incarcerated. All matches were one-to-one mentoring relationships, with slightly more (55%) community-based matches compared with school- or site-based settings (45%). Enhancements to mentoring programs serving COIP and non-COIP youth were assessed via telephone interview with BBBS program staff. Enhanced program practices included having specialized mentor training for mentoring of COIP, as well as having specific goals related to serving COIP and having funding dedicated to providing services for COIP. COIP and non-COIP matches were compared on a range of different demographic characteristics (characteristics of mentees, the mentees’ living situation, mentor demographic characteristics, and match characteristics – community-based match versus site- or school-based) and differed on all characteristics except for youth age. The 20 covariates included in analyses included youth and mentor gender, race, ethnicity and age; youth living situation (e.g., foster care, single parent home), family income and subsidized lunch status; mentor education, marital status and profession (helping profession versus other profession); and match type (community- versus site-based).

Archival data was obtained from the BBBSA Agency Information Management (AIM) system. Measures came from mentee report, mentor report, and staff report AIM data. Youth social acceptance was measured on a 6-item scale, rated by mentees, with higher scores representing higher perceived social acceptance). Youth parental trust was rated with 3 items rated on a 4-point scale, with higher scores indicating a greater sense of trust. School attendance was assessed with youth ratings on 2 items (number of absences and number of lates in the past 30 days) with higher scores representing higher attendance. Youth scholastic competence was measured using a 6-item scale with items relating to perceived academic achievement (e.g., “I am very good at my school work”), scored so that higher scores represent greater scholastic competence. Youth educational expectations were measured using a 3-item scale asking about future educational plans, with higher scores indicating greater educational expectations. Youth grades were based aggregated ratings across 4 items (e.g., “how are you doing in mathematics?”) with higher scores indicating higher grades. Match relationship length was drawn from agency archival records, based on the start date and end date of each match, with all analyses conducted on matches that had already closed. The Mentor Strength of Relationship survey is 14-item scale measuring mentor’s feelings about their mentoring relationship, scored so that higher scores indicate stronger relationships. Inverse Propensity Weighting (IPW) was employed in order to calculate estimators of average treatment effect on the treated (ATT) for use in analyses, which allowed the data set to be adjusted for the relatively small proportion (19%) of youth with an incarcerated parent, and to balance the COIP versus non-COIP populations on the set of covariates to increase the likelihood that observed differences are due to the child’s COIP status rather than other background factors. Logistic regression analyses on youth and match outcomes controlled for the 20 covariates noted above, baseline value on the outcome of interest (if assessed), with the effect of focus being the mentoring enhancement (specifically the enhanced training) x COIP status interaction. Further details on the study design and sample can be found in Stump et al., (2018).
 

Evaluation Outcomes

Match Relationship Length
Stump et al., (2018) found that COIP who came from programs that had specialized mentor training had longer matches than COIP who came from programs that did not have specialized training.

Mentor Strength of Relationship
Stump et al., (2018) found that mentors of COIP who came from programs that had specialized mentor training reported stronger relationships than mentors of COIP who came from programs that did not have specialized training.

Social Acceptance (mentee report)
Stump et al., (2018) found that specialized mentor training for mentors serving COIPs was not significantly associated with mentee-reported social acceptance.

Parental Trust (mentee report)
Stump et al., (2018) found that specialized mentor training for mentors serving COIPs was not significantly associated with mentee-reported parental trust.

School Attendance (mentee report)
Stump et al., (2018) found that specialized mentor training for mentors serving COIPs was not significantly associated with mentee-reported school attendance.

Scholastic Competence (mentee report)
Stump et al., (2018) found that specialized mentor training for mentors serving COIPs was not significantly associated with mentee-reported scholastic competence.

Educational Expectations (mentee report)
Stump et al., (2018) found that specialized mentor training for mentors serving COIPs was not significantly associated with mentee-reported educational expectations.

Grades (mentee report)
Stump et al., (2018) found that specialized mentor training for mentors serving COIPs was not significantly associated with mentee-reported grades.