Mentor-Mentee Activity Guidance
Evidence Rating for this Practice:
Promising (Promising: 4 Studies; Null Effect: 2 Studies; Insufficient Evidence: 2 Studies)
In four of the eight studies reviewed, the practice of mentor-mentee activity guidance was associated with better outcomes. In these four studies, the outcome evidence and the methodology used for assessing effects of the practice both satisfied criteria for a designation of Promising. In two of the remaining 4 studies, the methodology used for assessing the effects of the practice met criteria for rigor but the outcome evidence did not reach the threshold for a designation of Promising; these studies were thus classified as Null Effect. The final 2 studies were designated as Insufficient Evidence because the methodology used for assessing effects of the practice did not meet relevant criteria for rigor.
The pattern of evidence designations across the reviewed studies meets criteria for an overall evidence rating of Promising for the practice. This rating is based on currently available research. Findings and conclusions may change as new research becomes available.
Description of Practice:
Mentor-mentee activity guidance involves supporting matches with engaging in particular types of activities or discussions. The activities that are the focus of the guidance can vary widely in association with considerations such as the characteristics, backgrounds, and needs of mentees, mentor backgrounds, skills, and interests, and program goals. Possibilities include relatively specific types of outings or project-based activities as well as more broadly defined kinds of activities or discussions that align with a particular goal (e.g., mentee spark development). The timing, number, frequency, and length of the target activities can also vary; in some instances activity guidance may take the form of relatively limited “icebreaker” activities at the outset of relationships, whereas in others it may involve a much more extensive collection of activities that is intended to encompass the full duration of relationships. The level of guidance provided similarly can range from simple lists of activities to step-by-step instructions. Activity guidance may be delivered in a variety of formats, such as mentor training sessions, group workshops for mentors and mentees, match support contacts, and printed or on-line resources. Furthermore, the activities involved may be presented as either optional or required.
This practice is distinguished from pre-match or post-match training for mentors as well as match support for mentors, the aims of which are typically broader and more multi-faceted than activity guidance. It is also distinct from one-time events (e.g., holiday parties) or other similar opportunities (e.g., attending a sporting event) that may be sponsored by programs as these are not typically oriented toward facilitating particular types of mentor-mentee interactions or activities. Finally, when goal setting is a focus, matches may be helped to pursue activities that support progress toward selected goals. However, because such guidance is likely to be only one of several components involved in supporting matches to work toward goals, goal setting/pursuit is considered to be a distinct practice from mentor-mentee activity guidance.
The primary goal of the practice is to facilitate mentees and mentors engaging in interactions and activities that will strengthen the mentoring relationship and promote positive outcomes for the mentee.
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 programs.
Theory and Evidence-Informed Principles:
Mentor-mentee activity guidance is not guided by a particular theoretical perspective. However, this practice is consistent with frameworks that emphasize the need for mentors and mentees to engage in specific types of activities or experiences in order to fully realize goals of prevention (Cavell & Elledge, 2014) or promotion of positive development (Lerner et al., 2014). Lerner and colleagues (2014), for example, discussed ways in which mentors could be coached to engage in different types of activities with their mentees to facilitate each of the “Cs” of positive youth development (i.e., competence, confidence, connection, character, caring, contribution). Mentor-mentee activity guidance is likely to often be directed toward strengthening the knowledge, skills, and efficacy beliefs of mentors for engaging in particular types of interactions with mentees. These factors are emphasized as being influential in numerous established theories of behavior change, including the Integrated Behavior Model (Montano & Kasprzyk, 2008).
Corresponding Elements of Effective Practice:
This practice is most relevant to the area of Monitoring and Support within the Elements of Effective Practice.
The successful implementation of this practice is likely to require staff to have skill and experience in the delivery format(s) that are being used. For example, where a training or workshop format is utilized, staff should have a mastery of the substantive content and be experienced with effective methods of group facilitation and instruction. Of key importance in all instances is likely to be an awareness on the part of staff of the need to provide guidance to mentors and youth in ways that are responsive to their interests and desires for self-direction. Staff ideally will be experienced in the use of strategies that are well-aligned with this principle (e.g., motivational interviewing).
Kaufman (2010) tested the effects of an intervention for female mentors (“Big Sisters”) in a Big Brothers Big Sisters of America (BBBSA) community-based mentoring program that was intended to increase their levels of communication with their mentees (“Little Sisters”) about sexual health issues. Big Sisters were recruited from a state chapter of BBBSA in New England. To be eligible for the study, Big Sisters had to be matched with a Little Sister for a minimum of three months and the Little Sister had to be at least 10 years old. Any Big was excluded if her Little’s parent/guardian had specified sexual topics as a conversation that should not be pursued by the Big when spending time with the Little.
Of 265 Big Sisters eligible for the study, 66 completed the intervention as well as both baseline and follow-up surveys (140 were unable to be contacted or declined participation and 59 were no-shows for the baseline assessment). The participating Big Sisters were randomly assigned to either the sexual health communication intervention being investigated (n = 32) or a control group (n = 34). The average age of the participating Big Sisters was 41 years; two-thirds (67%) had completed at least a bachelor’s degree. A majority of the Big Sisters were women of color (African American: 37.8%; Latina: 28.8%; Caucasian: 25.8%; multiracial: 4.5%). Most had been matched with their Little Sister for three years or less (less than 1 year: 25.8%; 1–2 years: 21.2%; 2–3 years: 21.2%). On average, at baseline Big Sisters reported spending 6.25 hours with their Little Sisters at their last meeting and a majority reported seeing their Little Sisters once every two to three weeks (n = 36, 54.5%). The intervention and control groups did not differ significantly on any of the preceding variables. The Little Sisters in this study were, on average, 13 years old with a racial/ethnic distribution similar to their Big Sisters (African-American: 37.9%; Latina: 28.8%; Caucasian: 25.8%; multiracial: 4.5%).
The intervention was delivered by the researcher in a single session, in a small group format (2-6 Big Sisters per group), lasting approximately three hours, and using various modalities, including discussions, videos, and role-playing (see Kaufman, 2010, for a more detailed description). The intervention was designed to increase the self-efficacy beliefs of Big Sisters for talking with their Little Sisters about issues related to sexual health, such as avoiding pregnancy, STIs, HIV, using condoms, and delaying sexual debut as well as their perceptions of such conversations being likely to have positive outcomes. In addition, Big Sisters were encouraged to talk with their Little Sisters about the positive aspects of sexuality, such as being comfortable with oneself and seeing sex as a positive experience between two people when it is consensual and safe. Big Sisters in the control group also participated in a similarly structured session, but with a focus on promoting discussions about bullying and peer pressure with Little Sisters. Audiotapes of both types of sessions were content coded by research assistants to ensure that all components were covered consistently across sessions and that content discussed in the sexual health training was not discussed in the control group training.
Study outcomes were measured on four different occasions using surveys of Big Sisters: pre-intervention (baseline) and at weeks 2, 4, and 6. The outcome measures included scales assessing perceptions of self-efficacy and outcome expectations for discussing sex-related topics (having sex, getting pregnant, STIs/HIV, changes due to puberty, and romantic relationships with boys) with the Little Sister. Additional measures assessed whether the Big Sister talked to her Little about sex-related topics in the preceding two weeks as well as the frequency and length of these conversations (separate scales were used to assess discussions relating to each topic). Analyses revealed no significant differences between intervention and control groups at baseline on these latter measures. Rates of attrition (non-survey completion) varied across follow-up, with 89% of participants completing at least two follow-up surveys. However, attrition rates did not vary significantly between the intervention and control groups at any of the three follow-ups; attrition was also found to be unrelated to baseline scores on outcome measures or participant demographic characteristics for the sample overall as well as in interaction with group status (intervention vs. control). Primary analyses used analysis of variance (ANOVA) and logistic regression to test for differences between the intervention and control groups on the outcome measures at each follow-up assessment, controlling for baseline scores on the outcome.
Mentors' Self Efficacy for Talking About Sexual Health Issues
Kaufman (2010) found no significant difference, at the 6 week follow-up, between Big Sisters who received the intervention and those in the control group with regard to their reported levels of self-efficacy for discussing sexual health issues with their Little Sisters. Results did show that Big Sisters increased their scores on this measure over the 4 times of assessment; however, this was similarly true of Big Sisters in each group.
Mentors' Outcome Expectancies for Talking About Sexual Health Issues
Kaufman (2010) found no significant difference, at the 6 week follow-up, between Big Sisters who received the intervention and those in the control group with regard to their reported outcome expectations for discussing sexual health issues with their Little Sisters.
Total-time “Having Sex” was Discussed
This study found no significant difference, at the 6 week follow-up, between Big Sisters who received the intervention and those in the control group with regard to whether they reported having discussed the topic of having sex with their Little Sisters during the preceding 2 weeks.
Total-time "STIs/HIV" was Discussed
This study found no significant difference, at the 6 week follow-up, between Big Sisters who received the intervention and those in the control group with regard to whether they reported having discussed the topic of STIs/HIV with their Little Sisters during the preceding 2 weeks.
Total-time "Romantic Relationships with Boys" was DiscussedAdditional Findings
This study found no significant difference, at the 6 week follow-up, between Big Sisters who received the intervention and those in the control group with regard to whether they reported having discussed the topic of romantic relationships with boys with their Little Sisters during the preceding 2 weeks.
Kaufman (2010) also tested for evidence of effects of the intervention under study at the 2 and 4 week follow-up assessments, using alternative formulations of the study outcome measures (e.g., total time a topic was discussed), and only among Big Sisters who reported low levels of discussion of sex-related topics with their Little Sisters at the baseline assessment. In general, as with the findings reported above, these analyses failed to indicate significant differences between the group receiving intervention designed to promote discussion of sexual health issues and the control group that participated in the alternative session focused on promoting discussion of bullying/peer pressure.
At study baseline, the majority of Big Sisters (56%) reported that they never or rarely discussed having sex with their Little Sisters. The corresponding percentages of never or rarely discussing other topics were similar or higher and as follows: getting pregnant (58%), STIs/HIV (77%), changes due to puberty (67%), and romantic relationships with boys (65%).
Studies 2 (Practice: Suggested Mentor-Mentee Activities) & 3 (Practice: Required Mentor-Mentee Activities)
Study 2 – Promising; Study 3 – Insufficient Evidence
Wheeler & DuBois (2009) examined correlates of practices relating to mentor-mentee activity guidance as part of a survey of 113 BBBSA affiliate agencies regarding their community-based mentoring programs; the sample included 80 agencies with data available in BBBSA’s Agency Information Management (AIM) system. Data were collected via an online survey conducted between May and September of 2009. The response rate was 34 percent for all agencies (i.e., 131 of 385 total) and 81 percent for AIM agencies. The survey included questions asking whether the agency provided matches with a list of ideas for activities (reported by 92% of programs) and whether matches were asked to complete specific types or categories of activities (reported by only 4% of programs).
Analyses examined the associations of each reported type of practice with average scores on measures of mentoring relationship quality (Strength of Relationship [SOR] measure) taken 3 months into the relationship as reported by both the volunteer (Big Brother or Big Sister) and child as well as on measures of relationship duration (6 and 12 month retention rates, average match length). These measures were derived from the AIM system and thus the analyses were limited to the subset of 80 agencies with AIM data. Analyses included bivariate correlations as well as partial correlations that controlled for any of several assessed agency characteristics (e.g., % of women volunteers, % of children with 2 or more risk factors) that had a significant or near-significant (p<.10) correlation with the match quality or duration measure being considered.
3-month Child-reported Strength of Relationship (SOR) Score
Wheeler and DuBois (2009) found no significant correlation between agency reports of either providing matches with a suggested list of activities or required types or categories of activities and 3-month child-reported SOR scores.
3-month Volunteer-reported Strength of Relationship (SOR) Score
Wheeler and DuBois (2009) found a significant and positive correlation between agency report of providing matches with a suggested list of activities and 3-month volunteer-reported SOR scores. However, the corresponding correlation of the 3-month volunteer SOR score with report of having required types or categories of match activities was not significant.
6-month Match Retention Rate
Wheeler and DuBois (2009) found a significant and positive correlation between a report of having a suggested list of mentor-mentee activities and the 6 month match retention rate (i.e., the % that lasted at least six months). The correlation between having required mentor-mentee activities and 6 month match retention was in a negative direction and, although not statistically significant, was large enough to indicate an association of substantive magnitude.
12-month Match Retention Rate
Wheeler and DuBois (2009) found a significant and positive correlation between agency reports of having a suggested list of mentor-mentee activities and the 12 month match retention rate (i.e. the % of matches that lasted at least 12 months). The correlation between having required mentor-mentee activities and 12 month match retention was in a negative direction and, although not statistically significant, was large enough to indicate an association of substantive magnitude.
Average Match Length
Wheeler and DuBois (2009) found a significant and positive correlation between agency report of providing matches with a suggested list of mentor-mentee activities and average match length. However, the corresponding correlation of average match length with report of having required types or categories of match activities was not significant.
Wheeler and DuBois (2009) also examined whether the practices of providing mentors with a resource book that included instructions and resource materials for different activities and providing mentors with suggested activities for specific age groups were correlated with the above described measures of match quality and duration. Thirty percent of programs reported providing resource books to matches and 73% reported providing suggested activities for specific age groups. Analyzes revealed no significant correlations between provision of activity instructions and resource materials and any of the outcomes. However, a negative correlation between this practice and 3 month volunteer-reported SOR scores was large enough to indicate an association of substantive magnitude. Providing suggested activities for specific age groups was correlated significantly and positively with 6 and 12 month retention rates and average match length, but was not correlated with 3 month child or volunteer SOR scores.
Studies 4 (Practice: Incremental Condition) and 5 (Practice: Attribution Condition)
Study 4 – Promising; Study 5 - Promising
Good and colleagues (2003) tested the effects of an intervention designed to reduce vulnerability to stereotype threat and promote academic achievement within the context of school-based mentoring provided to junior high students. The students attended a rural school district in Texas that served a largely low-income and Hispanic population. The sample for research consisted of 138 seventh graders who were randomly enrolled, by school administration, into a computer skills class as part of their junior high school curriculum in which they learned a variety of computer skills, including emailing and web page design. Study participants were predominantly Hispanic (67% Hispanic, 13% Black, and 20% White) and 45% were female.
Students were randomly assigned to one of three intervention conditions (incremental, attribution, and combination) or a control group. In the incremental condition, participants learned about the expandable nature of intelligence and how the brain is able to form new connections throughout one’s lifetime. In the attribution condition, participants learned that all students face academic difficulty during the transition into high school but that most students overcome these difficulties and reach high levels of achievement. Students in the combination condition received both types of intervention. All students (including those in the control condition) were randomly assigned a mentor shortly after the school year began (mid-October), with the interventions beginning once students were proficient in basic computer use and emailing (mid-November). Intervention messages were conveyed to students by mentors during two in-person meetings (in mid-November and end of January) that lasted 90 minutes each and via weekly email correspondence throughout the school year. The ideas conveyed in these discussions with mentors were reinforced by requiring student to create web pages that advocated, in their words and pictures, the messages they received. Mentors were available to help students with their web pages or with any school related issues, including adjusting to the new school environment and useful study strategies. Students also had access to a restricted web page unique to each intervention group and accessible to only those in the group, where they could get ideas for their web pages, helping students to further internalize the message in their condition. Control group students were mentored about the danger of drugs and created anti-drug web pages.
Mentors were 25 college students from the University of Texas. Mentors received three hours of training designed by the school district at the beginning of the school year (early September), which also included methods of conveying each of the experimental messages. Mentors were informed that all intervention conditions were intended to benefit the students academically but they were blinded from the study hypotheses. To avoid individual differences between mentors potentially biasing tests of effects of the study conditions, each mentor was randomly assigned mentoring responsibilities for one to two students in three of the four conditions, resulting in each mentor working with approximately six students.
Student outcomes in math and reading achievement were measured at the end of the school year using the Texas Assessment of Academic Skills test, a statewide standardized achievement test. Outlier analysis eliminated five participants’ math scores from the 138 participants who took the math test and six participants’ reading scores from the 135 participants who took the reading test. The standardized test performances of students in the intervention groups were compared with that of the control group using analysis of variance (ANOVA) tests. Information was not reported on the equivalence of students in the intervention and control conditions on demographic characteristics, pre-existing levels of academic achievement, or other potential influences on the outcomes examined.
Math Test Score
Good et al. (2003) found that students in the incremental condition as well as those in the attribution condition had significantly higher math test scores than students in the control group at post-test.
Reading Test Score
Good et al. (2003) found that students in the incremental condition as well as those in the attribution condition had significantly higher reading test scores than students in the control group at post-test.
Good et al. (2003) also tested for potential differences in effects of the intervention conditions between male and female students. These analyses revealed a significant gender by condition interaction. More specifically, although both male and female students in the incremental, attribution, and combination conditions achieved significantly higher math scores than their counterparts in the control condition, these differences were greater among female students. The net result was that a performance gap that existed among male and female students in the control group (male students performed significantly better than female students) was not evident within any of the intervention conditions. A significant gender by condition interaction for reading scores was not reported.
DuBois et al. (2014) assessed the effect of a mentor-mentee activity curriculum that was designed to both help middle school students develop a growth mindset about their academic abilities and to explore their Sparks (i.e., activities of particular interest) and connect them to their education. The curriculum was implemented on a pilot basis within existing youth serving organizations/programs within six school districts in South King County and Seattle (south Seattle only), Washington.
An environmental scan conducted by Washington State Mentors (WSM) identified six organizations that provided a structured youth mentoring program for middle school youth in a one-to-one format, had program-pilot compatibility, and had the staff capacity necessary to conduct required study activities. Youth in each program who were in the intended age range were identified and assigned to either a group receiving the curriculum from their mentors (the treatment group) or not (the comparison group). Random assignment of youth into treatment and comparison group was conducted differently at each site because of program differences in how mentors and youth were matched and interacted (for example, some programs assigned one mentor to several mentees). The final analysis included data from four of the six programs (one program was dropped due to insufficient mentor attendance for training; another program had a small number of participants and thus all were assigned to receive the curriculum). All four programs were based in a school or other site (e.g. teen center). Further details on the study design and sample can be found in DuBois et al. (2014).
A total of 121 youth participated in the study with 62 assigned to the treatment group and 59 assigned to the comparison group. Participating youth were in the 5th to 9th grades (with 61% being in 7th and 8th grades), 56% were female, 72% were from low-income families (as indicated by student report of receiving free or reduced price school lunch), and 87% reported being from homes in which neither parent had completed a 4-year college degree. The youth in the treatment and comparison groups did not differ significantly on any of these background characteristics. Across all programs, 38 different mentors mentored treatment group youth and 28 different mentors mentored comparison group youth. Mentors of youth in the treatment group received an initial 2-hour training on the mentor-mentee activities conducted in a group-format. This training was intended to provide mentors with the background knowledge (Sparks and mindset concepts and related research) as well as practical strategies needed to support their mentees’ adoption and use of a growth mindset as well as their identification and exploration of Sparks. The curriculum consisted of 5 modules (each of which was designed to be completed within one or two meetings) and 4 associated activity challenges that the youth and mentor were tasked with completing separately after each module. The curriculum addressed the following content areas: Spark concept definition, identification, and selection, Sparks development and connection to school success and academic interests, basic brain neurology, assessing and strategies for changing mindsets, and connecting mindset to school success. Each module was completed during one 45 minute session or two 20-25 minutes sessions.
All participating youth and mentors were surveyed at the beginning of the study (prior to any implementation of the curriculum) and at the end of the study, allowing as much time as possible for implementation of the curriculum within the constraints imposed by the end of the school year. The time period allowed for curriculum implementation varied across programs as well as among mentoring relationships within programs, but typically was approximately 3 months. Pre-test surveys were collected from 97 of the participating youth (for an 80% response rate), including 55 in the treatment group (89%) and 42 in the comparison group (67%). Of these youth, 80 (82%) completed the post-test survey, 40 of whom were in the treatment group (73%) and 40 of whom were in the comparison group (95%). Although the response rates for the pre- and post-test surveys did not differ substantially across programs, they did differ by treatment and comparison group within programs. Pre-test surveys were completed by 95% of mentors in the treatment group and 86% of mentors in the comparison group. Additionally, 83% and 89% of these groups of mentors, respectively, also completed the post-test survey.
Surveys provided information on youth and mentor demographic and other background characteristics as well as study outcomes. Outcomes assessed included 1) youth and mentor reports of mentor support for adopting a growth mindset, Sparks exploration and development, and connecting Sparks to education, mentor role modeling of growth mindset and sparks exploration and development, 2) youth and mentor reports of youth’s adoption of growth mindset about intelligence, Sparks exploration, development, and connection to education, 3) youth self-reports of attitudes and behaviors facilitating learning and academic success (learning goals, positive efforts beliefs, long-term educational expectations, sense of belonging at school, perceived value of school/education, aspirations for educational attainment, effort on school work, perseverance and mastery strategies in school work, school attendance and behavior, time spent on homework), and 4) youth self-reports of overall school performance as well as school performance in core subject areas. Post-test surveys of mentors also collected information on their participation in mentor training and use of the curriculum, the acceptability of the curriculum, and the adequacy of support received from program staff for using the curriculum. Mentors with more than one mentee reported whether they completed each module of the curriculum with any of their mentees, rather than with each mentee. A large majority of mentors reported completing the first two modules of the curriculum and the associated activity challenges that focused on Sparks content; substantially fewer mentors reported completing the modules and activities focused on growth mindset.
Regression analysis was used to assess if there was differential change from pre- to post-test on outcomes for youth assigned to treatment group relative to those assigned to the comparison group (i.e., post-test scores were predicted with control for pre-test scores on the outcome measured). Analyses controlled for a number of variables that showed evidence of differing across treatment and comparison groups at pre-test (mentor reports of strength spotting and growth mindset, and youth and mentor reports of feelings of closeness to one another, support from mentors for Sparks and Sparks exploration, connecting Sparks to education and educational expectations).
Mentor Support for Adopting Growth Mindset (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across treatment (mentoring with curriculum) and comparison (mentoring without curriculum) groups in youth reports of support from mentor for adopting a growth mindset.
Mentor Role Modeling of Growth Mindset (youth survey)
DuBois et al. (2014) found that youth in the treatment and comparison groups differed significantly in change from pre- to post-test in youth reports of mentor role modeling of growth mindset, with those in the treatment group showing more positive change.
Mentor Support for Sparks Exploration (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of support from mentor for Spark exploration.
Mentor Support for Connecting Sparks to Education (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of support from mentor for connecting Sparks to education.
Mentor Role Modeling of Sparks Exploration/Development (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of mentor role modeling of Sparks exploration/development.
Adoption of Growth Mindset (youth survey)
DuBois et al. (2014) found that youth in the treatment and comparison groups differed significantly in change from pre- to post-test in youth reports of growth mindset, with those in the treatment group showing more positive change.
Spark Identification (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of Sparks identification.
Spark Exploration (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of Sparks exploration.
Connecting Sparks to Education (youth survey)
DuBois et al. (2014) found that youth in the treatment and comparison groups differed significantly in change from pre- to post-test in youth reports of connecting Sparks to education, with those in the treatment group showing more positive change.
Learning Goals (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of learning goals.
Positive Efforts Belief (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of efforts belief.
Expectations for Educational Attainment (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of educational expectations.
Sense of Belonging at School (youth survey)
DuBois et al. (2014) found that youth in treatment and comparison groups differed significantly in change from pre- to post-test in youth reports of sense of belonging at school, with those in the treatment group showing more positive change.
Perceived Value of School/Education (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of perceived value of school/education.
Aspirations for Educational Attainment (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of educational aspirations.
Efforts on School Work (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of efforts on school work.
Perseverance (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of perseverance.
School Attendance (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of school attendance.
School Behavior (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of school behavior.
Time Spent on Homework (youth survey)
DuBois et al. (2014) found that there was no significant differential change from pre- to post-test across the treatment and comparison groups in youth reports of time spent on homework.
DuBois et al. (2004) also investigated potential effects of the curriculum on all of the mentor-reported outcomes referenced above as well as all additional youth-reported outcomes, such as school performance. Analyses indicated that youth in the treatment and comparison groups differed significantly in change from pre- to post-test in mentor-reported support for mentee connecting Sparks to education (with those in the treatment group showing more positive change) but not in change for any of the other mentor-reported outcomes. Additionally, a group difference for change in overall school performance that favored the treatment group was large enough to be judged ‘substantively important’, although it did not reach statistical significance.
Hairu-Joshu et al. (2010) assessed the impact of a curriculum-based intervention designed to improve diet and activity behaviors as an element of an existing mentoring program for high needs children. The study recruited 782 children and their parents from tutoring/mentoring programs at 119 OASIS Intergenerational Reading Program (OASIS) and Big Brother Big Sister (BBBS) sites, which were randomly assigned to intervention (n= 74 sites, 418 children) and control conditions (45 sites, 364 children) in a group randomized study design. Intervention children (and their parents) received the standard program plus PARADE (Partners of All Ages Reading About Diet and Exercise). The PARADE curriculum contained modules intended to enhance knowledge of dietary and activity guidelines, identify common and accessible activities, and low cost and accessible fruits and vegetables (FV). Each module of the curriculum contained an individual visit lesson plan, which was used by mentors as a guide to interaction with their mentees during meetings, as well as a storybook and a parent action newsletter. More detail about each curriculum component is available at Haire-Joshu et al., (2010). Children in control sites received the standard tutoring or mentoring program (OASIS or BBBS), which consisted of routine 1 hour visits of the tutor or mentor with the child in various community settings, including libraries, community centers, and after-school areas.
Outcome data were collected before and after the intervention and included 1) child reports of dietary and activity behavior (nutrition and physical activity knowledge, percent challenging themselves to eat five FV every day, and percent challenging themselves to be active at least 1 hour a day) facilitated by the mentor as part of lesson plans, 2) parent reports of the child’s dietary intake and activity level (child’s daily caloric intake, percent caloric consumed from fat, daily serving of FV, and percent time spent in physical activity), and 3) parent reports of their own dietary intake, physical activity level, height, and weight, and 4) measurements of child height and weight, which were used to calculate body mass index (BMI). Mentors/tutors documented attendance and delivery of the intervention curriculum using session content checklists and also rated the student’s participation in discussions and response to curriculum components.
The characteristics of the children at the intervention and control sites were similar, except for age – children in the intervention condition were slightly younger than those in the control condition. About half (48 percent) of the children in the intervention group were female, 54.3 percent were White, 59.7 percent were of normal weight; 56.1 percent of control children were female, 62.4 percent were White, and 56.6 percent were of normal weight. Parents of children in the intervention and control conditions also had similar socio-demographic characteristics (i.e., relation to child, race, employment, marital status, and weight status), except for education and income – parents of children in the intervention condition were more highly educated and reported a higher income than parents of children in the control condition.
Mixed effect models were used to assess changes in the outcomes from pre- to post-intervention, adjusting for intervention site, child age, gender, baseline BMI z-score, and parent education. Analyses of BMI z-score outcomes, however, only adjusted for parent education. The analysis of child report data was based on 451 children (296 at PARADE and 155 at control sites) who completed the pre and post-intervention surveys and the analysis of parent reported outcomes was based 279 children (124 at PARADE and 72 at control sites) with pre and post-intervention parent survey data. The average time between pre-test and post-test was 5.7 months and ranged from 2.1 months to 16.2 months, during which time 56 percent of children in the PARADE condition had received all eight curriculum modules and 82 percent had received at least six.
Knowledge Score (child survey)
Hairu-Joshu et al. (2010) found that child reported knowledge of healthy eating and activity was significantly higher at post-test for children at intervention sites compared to children at control sites.
Percent Challenging Self to Eat 5 FV Every Day (child survey)
Hairu-Joshu et al. (2010) found that the percent of children who reported challenging themselves to eat healthy was significantly higher at post-test for the intervention condition compared to the control condition. The percentage of children who reported challenging themselves to eat healthy increased by an average of 23.3 percentage points from pre- to post-test among children at the intervention sites and by 3.59 percentage points among children at the control sites.
Percent Challenging Self to be Active 1 Hour Every Day (child survey)
Hairu-Joshu et al. (2010) found that the percent of children who reported challenging themselves to be more active was significantly higher at post-test for the intervention condition compared to the control condition. The percentage of children who reported challenging themselves to be more active increased by an average of 21.0 percentage points from pre- to post-test among children at the intervention sites and by 0.04 percentage points among children at the control sites.
Daily Caloric Intake (parent survey)
This research found that parent reported daily caloric intake did not differ significantly between children in the intervention and control sites at post-test.
Percent of Calories from Fat (parent survey)
Hairu-Joshu et al. (2010) found that parent reported percent of calories from fat did not differ significantly between children in the intervention and control sites at post-test.
Daily Servings of Fruits & Vegetables (parent survey)
This research found that parent reported percent time child spent in physical activity did not differ significantly between children in the intervention and control sites at post-test.
Percent Time Spent in Physical Activity (parent survey)
Hairu-Joshu et al. (2010) found that parent reported percent time child spent in physical activity did not differ significantly between children in the intervention and control sites at post-test.
Hairu-Joshu et al. (2010) found that BMI¬-z scores did not differ significantly between children in the intervention and control sites at post-test.
Hairu-Joshu et al. (2010) also conducted stratified analysis to examine any difference in intervention effects for normal weight and overweight children. Analyses found that, for normal weight children, the proportions who reported challenging themselves to be active and to eat five FV every day were significantly higher at post-test for the intervention condition compared to the control condition. For overweight children, these proportions were not significantly different at post-test between intervention and control sites. On the other hand, for overweight children, parent reports of percent of calories consumed from fat were significantly lower at post-test for children in the intervention condition than those in the control condition. For normal weight children, however, parents reports of percent of calories consumed from fat were not significantly different at post-test between children at the intervention and control sites. Additionally, both normal weight and overweight children in the intervention condition achieved significantly higher knowledge scores than their counterparts in the control condition.
Davidson et al. (1987) examined the effects of delinquency treatments that used nonprofessionals within the context of a community-based diversion program on youth recidivism rates. The treatment conditions (of which there were five versions, all of which included mentoring from a college student volunteer) were based on behavioral contracting and child advocacy and involved participants, parents, and significant others. In the primary version of the intervention, the action condition (AC), participating youth were matched with a volunteer college student mentor; the mentor received eight weeks of training using procedures delineated in an eight-unit program manual. The manual was intended to assist with assessing and initiating desired behavioral changes and included both role plays and homework assignments. Participating youth had one-on-one interaction with these mentors for 6-8 hours per week, for 18 weeks, in the youth’s home, at recreational locations, or in any other mutually agreed upon community setting. Mentors were supervised by a pair of psychology graduate students, for two hours per week during the 18 intervention weeks. The action condition-family focus (ACFF) version of the intervention varied from AC in that only youth and their parents and siblings were involved. The action condition-court setting (ACCS) version varied from AC in that the mentors were supervised in the court setting by a caseworker who received 2 hours per week of supervision from the researcher. The relationship condition (RC) differed from AC in that it involved only the youth participant and mentor training focused on the development of empathy, unconditional positive regard, communication skills, and genuineness. The attention placebo condition (APC) varied from AC in that it relied on the natural skills of mentors and provided minimal training (three 2 hour lectures covering general topics in delinquency) and supervision (2 hours a month) and consisted primarily of recreational activities. In the control condition, participants were returned to the court referee for court processing.
Two hundred and twenty-eight youth who had been charged with a wide variety of crimes, including larceny and breaking and entering, were referred to the project by a court intake referee after a preliminary hearing. Youth offenders who would otherwise have been released, were charged with a serious crime, or were already on probation were excluded. 83 percent of participating youth were male, 26 percent were minority, and the mean age was 14.2 years. Participating juveniles were randomly assigned to one of the five intervention conditions or the control condition, stratified by gender, race, court referee, and order of referral. The different study conditions were implemented in different years, except for the AC and control conditions, which were implemented consistently across all project years, which resulted in unequal numbers of participants in the various conditions. The final sample consisted of 213 participants; 76 in AC, 24 in ACFF, 12 in ACCS, 12 in RC, 29 in APC, and 60 in the control condition. Eight youth declined participation after referral and seven additional youth dropped out after randomization (no more than two dropped out from any one condition). Pre-intervention outcome measures (including sex, age, race, prior arrests, prior court petitions, GPA, and self-reported delinquency) were not significantly different between participants in the five intervention conditions and the control condition. The mentors were 153 volunteer college students, randomly selected from a pool of 264 students who had applied for a three-term community projects course. The average age of mentors at the time they began the course was 20.43 years, and 84 percent were White, 90 percent were psychology or criminal justice majors, and 86 percent were juniors or seniors. More information about the study design is available at Davidson et al. (1987).
Official delinquency data was collected, for 1 year prior to referral and 2 years after the intervention, from the police records of 14 local law enforcement agencies and the Law Enforcement Information Network. These data included information on alleged offenses (police contacts), seriousness of alleged offenses, and disposition of alleged offenses. Court records also provided data on court petitions (dates and seriousness of the offenses involved) as well as incarceration length. Analysis, however, revealed that outcome measures were highly correlated with each other and displayed significant interaction between time and condition. Analysis was, therefore, limited to the outcome of number of court petitions at 2 years post intervention and used univariate chi-squared analysis and ANOVA. Because of interest in assessing evidence of effects of activity guidance, the review focused on the comparisons that were made between outcomes for youth in AC, ACFF, ACCS, and RC conditions combined (all of which included activity guidance from a mentor figure) and those for youth in the APC condition (which involved mentoring but without activity guidance).
Davidson et al. (1987) found that, at 2 years post-intervention, the number of court petitions did not differ significantly between those in the AC, ACFF, ACCS, and RC conditions combined and those in the attention placebo condition (APC).
A 29-item self-report delinquency measure assessing the frequency of delinquent behaviors in the 6-week period preceding the interview was also administered to participants at intake and 6, 12, and 18 weeks later, as well as 1 and 2 years after the intervention. A repeated-measures ANOVA comparing the total scale score for the six study conditions (i.e., the 5 treatment conditions and the control condition) by time (intake and 6, 12, and 18 weeks) and ANOVAs comparing scores across conditions at follow-up (1 and 2 years after the intervention) found no significant condition effects or interactions.
External Validity Evidence:
Variations in the Practice
The studies reviewed examined a variety of mentor-mentee activity guidance practices. Most often, mentor-mentee activity guidance was provided in a relatively structured (e.g., curricular) format along with reinforcing activities that mentor-mentee matches worked through. One study evaluated possible effects of whether mentors were simply provided with a list of suggested activities to engage in with their mentees. One study (Wheeler & DuBois, 2009) reported findings suggestive of possible negative effects of requiring mentors to engage in particular types of activities with youth; however, favorable results have been obtained in other studies in which mentors have been tasked with engaging in particular activities with youth (e.g., Hairu-Joshu et al., 2010). The implications of varying levels and types of requirements or flexibility that are incorporated into mentor-mentee activity guidance as a practice have not been studied systematically within particular studies and thus remain in need of clarification.
Studies have also varied in the content and purposes of the activity guidance provided. In some studies the focus of the mentor-mentee activity guidance was academic, in others the guidance had a health focus, and in one study it focused on delinquency prevention. The latter study (Davidson et al., 1987) examined several different forms of guidance in which the specific activities encouraged varied along dimensions such as family involvement, child advocacy, and degree of emphasis of interpersonal behaviors within the mentoring relationship itself. Additional studies that incorporate these types of contrasts in purposes and types of activities will be needed to begin to clarify what implications (if any) such differences have for the effects of practices in this area.Youth
Youth served by the programs in the reviewed studies represented a wide variety of backgrounds and included youth from low income families, as well as youth in rural school districts. Youth also varied within and across samples in types and levels of individual risk (for example, overweight status and prior delinquency) and in most instances the majority belonged to ethnic minority groups. All programs in the studies reviewed targeted youth who were elementary or middle school-aged with varying proportions of male and female students (one study included only female youth served in a Big Brothers Big Sisters program). Two studies investigated possible differences in effects of the practice of interest across subgroups of youth; Haire-Joshu et al. (2010) did so with respect to normal weight and overweight program participants and Good et al. (2003) did so with respect to male and female youth. The latter study found evidence that the type of mentor-mentee activity guidance provided in that research was more beneficial to the subgroup of youth (i.e., girls) that was most vulnerable to negative outcomes in the area examined (i.e., math achievement). However, findings of the other study were mixed in this regard. Overall, little attention has been devoted to testing for subgroup differences and because certain subgroups have been largely absent from study samples (e.g., high school students). Consequently, it is not possible to offer conclusions regarding the extent to which effects of mentor-mentee activity guidance tend to generalize or be differentially strong (or weak) across youth with differing characteristics and backgrounds.
Most of the mentors in the studies reviewed were volunteers, with one study including mentors who were paid staff members. In some studies, mentors were college students, whereas in others they were adult mentors participating in mentoring organizations. Only one study (Kaufman, 2010) provided more detailed demographic information on participating mentors. Studies reviewed also did not test for differences in effect of mentor-mentee activity guidance in relation to mentor characteristics. For these reasons, the applicability of findings to different subgroups of mentors is largely unknown.
The practices of interest in the studies reviewed in most instances were incorporated into existing mentoring programs. All programs used a one-on-one mentoring format, although in some programs each mentor worked with several mentees. As such, the overall conclusion of this review that mentor-mentee activity guidance has a promising evidence base would appear to be most likely to generalize to instances in which this practice is utilized within the context of an already operational program in which mentoring occurs on a 1-to1 format. Among the studies receiving a Promising designation, mentoring was variously provided in the school setting or in the community at large. Findings are thus tentatively supportive of the value of activity guidance in these two types of programs; its effectiveness within programs in which mentors and mentees interact in other types of settings (e.g., workplace) is unstudied.
Studies included in this review assessed effects of the practices of interest on a range of different youth outcomes, including academic attitudes and behaviors, health-related attitudes and behaviors, and delinquency. Findings were mixed within and across outcome areas, making it difficult to assess similarities or differences in effects of mentor-mentee activity guidance as a practice based on the domain of youth outcome being targeted or assessed. Several studies assessed effects of the practice of interest on mentoring relationship outcomes, such as mentoring relationship duration and quality. However, such outcomes were not included consistently. This is a noteworthy omission given that mentoring relationship outcomes may be those most directly and consistently influenced by activity guidance as a practice.
Resources Available to Support Implementation:
Resources to support implementation of mentor-mentee activity guidance can be found under the Resources section of this website. These resources include:
College Positive Mentoring Toolkit – This online toolkit for mentors includes ready-to-use activities, checklists, and background information that can support mentees of all ages as they think about, and plan for, postsecondary education.
Discovering the Possibilities: "C"ing Your Future – This 12-module curriculum and activity guide is designed to assist mentors in working with middle school youth to explore postsecondary education and possible careers.
Finding Mentors, Finding Success – This guidebook for mentees teaches youth how to seek out appropriate mentors who can help them with their next steps in life once leaving a mentoring program or other system of support.
K-12 Journey Map – This tool is intended to help youth and mentors track important milestones as youth make the journey from school to post-secondary education and career planning.
Peer Mentoring Handbook – This resource offers advice, strategies, and other information to older youth who will be serving as peer mentors to younger children in school and community settings.
Young Men's Initiative (YMI) Activities Catalog – This activity guide offers a number of games and fun activities that are designed to get male mentors and mentees interacting and learning together, both individually and in groups.
Davidson, W. S., Redner, R., Blakely, C. H., Mitchell, C. M., & Emshoff, J. G. (1987) Diversion of juvenile offenders: An experimental comparison. Journal of Consulting and Clinical Psychology, 55, 1, 68-75. http://dx.doi.org/10.1037/0022-006X.55.1.68
DuBois, D. L., Felner, J., Heubach, J., & Mayer, N. (2014) Mentoring for academic success: A pilot study. Final report submitted to Raikes Foundation, Seattle, WA.
Good, C., Aronson J., & Inzlicht, M. (2003). Improving adolescents’ standardized test performance: An intervention to reduce the effects of stereotype threat. Applied Development Psychology, 24, 645-662. http://dx.doi.org/10.1016/j.appdev.2003.09.002
Hairu-Joshu, D., Nanney M. S., Elliott, M., Davey, C., Caito, N., Loman, D., . . . Kreuter, M. (2010) The use of mentoring programs to improve energy balance behaviors in high-risk children. Obesity, 18, S1, S75-S83. http://dx.doi.org/10.1038/oby.2009.435
Kaufman M. R. (2010). Testing of the Healthy “Little” Lives project: A training program for Big Sister mentors. American Journal of Sexuality Education, 5, 305-327. http://dx.doi.org/10.1080/15546128.2010.527234
Wheeler, M., & DuBois, D. L. (2009). Analysis of responses to agency practices survey for Big Brothers Big Sisters of America’s Community-Based Mentoring Program. Unpublished report prepared for Big Brothers Big Sisters of America.
Cavell, T.A. & Elledge, L. C. (2014). Mentoring and prevention science. In D. L. DuBois & M. J. Karcher (Eds.). Handbook of Youth Mentoring, 2nd Edition (pp. 29-41). Thousand Oaks, CA: Sage Publications.
Lerner, R. M., Napolitano, C. M., Boyd, M. J., Mueller, M. K., & Callina, K. S. (2014). Mentoring and positive youth development. In D. L. DuBois & M. J. Karcher (Eds.). Handbook of Youth Mentoring, 2nd Edition (pp. 17-28). Thousand Oaks, CA: SAGE.
Montano, D. E., & Kasprzyk D. (2008) Theory of Reasoned Action, Theory of Planned Behavior, and the Integrated Behavior Model. In K. Glanz, B. K. Rimer, & K. Viswanath (Eds.), Health behavior and health education (4th ed., pp. 67-92). San Francisco: Jossey-Bass.