The Role of Program-Supported Mentoring Relationships in Promoting Youth Mental Health, Behavioral and Developmental Outcomes

Abstract

This study examined the relationship between youth mentoring status and behavioral, developmental, and emotional outcomes for 859 youths aged 6–17 participating in a national survey of Big Brothers Big Sisters community mentoring relationships (MRs). Youth self-reported behaviors and mental health occurred at the baseline assessment (before being paired to a mentor) and at 18 months follow-up. Youth mentoring status was categorized as follows: (1) continuous MR less than 12 months (n = 131); (2) continuous MR 12 or more months (n = 253); (3) dissolved MR less than 12 months (n = 110); (4) dissolved MR 12 or more months (n = 70); 5) MR with a second mentor (re-matched; n = 83); and (6); never mentored (n = 212). Structural equation model results at 18 months revealed that mentored youths, especially those in MR lasting 12 or more months (continuous or dissolved), reported significantly fewer behavioral problems and fewer symptoms of depression and social anxiety than did non-mentored youths. They also reported stronger coping skills and emotional support from parents. Mentored girls and boys in long-term relationships experienced positive outcomes. Re-matched girls displayed better outcomes than did never-mentored girls while there was some evidence of harmful outcomes for re-matched boys. Threats to internal validity are examined including the possibility of pre-existing baseline differences between mentored and non-mentored youths. Implications for mentoring programs are discussed.

Full text

The Role of Program-Supported Mentoring Relationships
in Promoting Youth Mental Health, Behavioral
and Developmental Outcomes
David J. DeWit
1
&David DuBois
2
&Gizem Erdem
3
&Simon Larose
4
&
Ellen L. Lipman
5
#Society for Prevention Research 2016
Abstract This study examined the relationship between
youth mentoring status and behavioral, developmental, and
emotional outcomes for 859 youths aged 6–17 participating
in a national survey of Big Brothers Big Sisters community
mentoring relationships (MRs). Youth self-reported behaviors
and mental health occurred at the baseline assessment (before
being paired to a mentor) and at 18 months follow-up. Youth
mentoring status was categorized as follows: (1) continuous
MR less than 12 months (n= 131); (2) continuou s M R 1 2 or
more months (n=253); (3) dissolved MR less than 12 months
(n= 110); (4) dissolved MR 12 or more months (n=70); 5)
MR with a second mentor (re-matched; n= 83); and (6); never
mentored (n= 212). Structural equation model results at
18 months revealed that mentored youths, especially those
in MR lasting 12 or more months (continuous or dissolved),
reported significantly fewer behavioral problems and fewer
symptoms of depression and social anxiety than did non-
mentored youths. They also reported stronger coping skills
and emotional support from parents. Mentored girls and boys
in long-term relationships experienced positive outcomes. Re-
matched girls displayed better outcomes than did never-
mentored girls while there was some evidence of harmful
outcomes for re-matched boys. Threats to internal validity
are examined including the possibility of pre-existing baseline
differences between mentored and non-mentored youths.
Implications for mentoring programs are discussed.
Keywords Emotional problems .Behavioral problems .
Developmental outcomes .Youth mentoring programs
Background
Meta-analyses of youth mentoring programs have found that
youths paired to an adult mentor experience significant im-
provements in behavioral and psycho-social outcomes com-
pared to non-mentored youths (DuBois, Portillo, Rhodes,
Silverhorn, and Valentine 2011;Meyerson2013;Tolan,
Henry, Schoeny, Lovegrove, and Nichols 2014). Mentoring
theorists (e.g., Rhodes, Spencer, Keller, Liang, and Noam
2006) suggest that these improved outcomes occur through
several mechanisms of change. By, for example, modeling
effective adult communication and pro-social behavior, men-
tors may help youths express and regulate emotions and
choose adaptive coping strategies for managing environmen-
tal stressors. In addition, close and secure attachments with
mentors may help youths to develop positive internal working
models that improve their perceptions of the value of interper-
sonal relationships (e.g., parents, peers). Mentors may also
enhance youth self-esteem through praise and educational
and recreational activities that allow for the discovery of
unique abilities and improve cognitive functioning by
Electronic supplementary material The online version of this article
(doi:10.1007/s11121-016-0663-2) contains supplementary material,
which is available to authorized users.
*David J. DeWit
ddewit@uwo.ca
1
Social and Epidemiological Research Department, Centre for
Addiction and Mental Health, 100 Collip Circle, Suite 200,
London, Ontario N6G 4X8, Canada
2
University of Illinois, Chicago, IL, USA
3
Koc University Istanbul, Istanbul, Turkey
4
Laval University, Quebec City, Quebec, Canada
5
McMaster University, Hamilton, Ontario, Canada
Prev Sci
DOI 10.1007/s11121-016-0663-2

teaching skills, offering intellectual challenges, and promoting
positive attitudes toward learning.
While mentoring appears to be a promising approach for
improving youth well-being, program effect sizes have been
low to moderate at best with little evidence of sustained im-
pact (DuBois et al. 2011). One possible explanation for the
modest impact of mentoring is that most community evalua-
tions compare mentored versus non-mentored youths.
Unfortunately, these comparisons risk masking potential var-
iation in program impacts possibly due to variability in the
stability and longevity of the mentoring relationship, attributes
considered to be important benchmarks for defining success-
ful mentoring outcomes (Rhodes 2008).
Mentoring stability and longevity are central to Keller’s
(2005) theory of mentoring relationship development.
According to this theory, the mentoring relationship is a dy-
namic and evolving entity marked by phases of development
including a contemplation phase (i.e., a period of anticipation
and preparation before a relationship is formed), an initiation
or getting acquainted phase, followed by a growth and
maintenance phase in which aspects of the relationship
become solidified, and ending with a phase of decline where
mentors and mentees retreat from the relationship and end
contact altogether. Some youths may enter their relationship
only to have it close within a few months while others
continue uninterrupted for months if not years under the
guidance of the same mentor. Still, others who have
experienced an unplanned closure may enter a new
relationship with a second mentor. Finally, some youths may
wait months before receiving a mentor and remain in the
contemplation phase for an indefinite period.
Keller (2005) asserted that the extent to which youths real-
ized health and social benefits would depend on the stage of
their mentoring relationship. He reasoned that youths who had
reached the growth and maintenance phase would generally
benefit the most because they had sufficient time to solidify
fundamental aspects of their relationship (i.e., expectations for
shared activities and regular contact, self-disclosure of person-
al feelings, trust, affection, and emotional support). Rhodes
and colleagues (2006) suggest that many of these mentoring
relationship qualities lie at the core of successful relationships
and therefore must be in place before positive gains in youth
development occur. These ideas underscore the need for a
mentoring status variable capable of predicting with greater
accuracy the degree to which youths at different phases of
their mentoring relationship experience positive developmen-
tal outcomes. Using data from a large national study of Big
Brothers Big Sisters community mentoring relationships,
the current study will seek to quantify the association
between various categories of youth mentoring status
(broadly aligned with Keller’s stages of mentoring rela-
tionship development) and selected aspects of youth health
and social functioning.
Stability and Longevity of Mentoring Relationships
Despite the emphasis placed by Keller on the heterogeneous
nature of youth mentoring relationships (i.e., phases of rela-
tionship development), research that has unpacked the
mentoring relationship into categories reflective of its stability
and longevity and related those categories to youth develop-
mental outcomes has been slow to emerge. In terms of stabil-
ity, previous work has shown that up to half of all program-
supported community mentoring relationships between
youths and adult mentors end before the standard 12-month
period of commitment (Grossman and Rhodes 2002;
Lymburne r 2006). Evidence suggests that these early relation-
ship endings may be psychologically damaging to the youths
(Grossman and Rhodes 2002), possibly because of mentee
feelings of rejection (Grossman and Rhodes 2002), abandon-
ment (Spencer, Basualdo-Delmonico, Walsh, and Drew
2014), or unfulfilled expectations of a deepening of the rela-
tionship (Goldner and Mayseless 2009). Mentoring relation-
ships that end prematurely have been identified as one reason
why some mentoring programs fail to reach their objectives or
even backfire by producing unintended negative effects
(Rodriguez-Planas 2014).
Longer mentoring relationships, especially those lasting 12
or more months, have been associated with increased health
and social benefits for youths (Grossman and Rhodes 2002;
Dolan, Brady, O’Regan, Russell, Canavan, and Forkan 2011).
For example, Grossman and Rhodes’(2002) study of
mentoring duration and youth development in American Big
Brothers Big Sisters (BBBS) community-based programs
found that youths who were mentored early and remained in
their relationship for at least 12 months benefited the most in
terms of improved self-worth, scholastic and social competen-
cies, and reduced truancy and substance use. Importantly, for
some outcomes, youths mentored less than 3 months were
worse off than non-mentored controls.
Other investigations, however, have failed to find an asso-
ciation between mentoring longevity and positive youth out-
comes (e.g., Eddy et al. 2015; Herrera, DuBois, and Grossman
2013). These conflicting results may be due to differences in
the way mentoring longevity is constructed. A rigorous study
of program-supported mentoring in the US northwest (Herrera
et al. 2013) did not find that longer relationships were more
beneficial. In that study, each of the defined mentoring lon-
gevity categories included youths in ongoing and dissolved
relationships. In contrast, in the study conducted by Grossman
and Rhodes (2002), it appears that youths in ongoing and
dissolved relationships lasting 12 or more months were com-
bined, while the shorter duration categories only contained
youths whose relationships had dissolved.
Less is known about the mentoring experiences of re-
matched youths. Research has shown that up to one third of
youths in closed mentoring relationships are re-matched to a
Prev Sci

second mentor within an approximate 1-year period (Herrera
et al. 2013). Whether or not re-matching mitigates harmful
effects associated with early mentoring relationship closures
is unclear. Herrera and colleagues (2013) found no evidence
that re-matching was associated with better youth outcomes
and that re-matched youths were worse off than were youths
in first time relationships because of the greater amount of
time they were mentored. A study of BBBS school-based
mentoring programs found that re-matched youths demon-
strated significantly poorer academic functioning compared
to non-mentored controls despite having been mentored for
an average of 25 weeks (Grossman, Chan, Schwartz, and
Rhodes 2012). Compared to re-matched youths, youths in
ongoing mentoring relationships with the same mentor were
mentored for a shorter period (23 weeks) and achieved the
greatest academic success.
Correlates of Youth Mentoring Status
Characteristics of the mentee and the mentee’s immediate en-
vironment may impact both the formation and maintenance of
program-supported mentoring relationships and youth devel-
opment (Rhodes 2002). Consequently, research aimed at es-
tablishing a link between youth mentoring status and various
outcomes must adjust for these factors. Studies have shown
that the pace with which mentoring relationships develop is
influenced by mentee demographics with boys and racial mi-
nority youths waiting significantly longer to be mentored than
other groups (Furano, Roaf, Styles, and Branch 1993;Moodie
and Fisher 2009). Mentored girls have been found to experi-
ence longer mentoring relationships than boys do (Rhodes,
Lowe, Litchfield, and Walsh-Samp 2008). There is also evi-
dence that relationships involving older youths tend to be
shorter (Grossman and Rhodes 2002; Rhodes, Schwartz,
Willis, and Wu 2014). Mentored youths whose mentoring
relationships end prematurely are also more likely to suffer
from mental health and behavioral difficulties and exposure
to negative environmental factors (e.g., physical or sexual
abuse, unstable family relationships, parent mental illness)
(Grossman and Rhodes 2002; Lymburner 2006; Rhodes
2002) that place them at greater risk of experiencing poor
outcomes. Herrera and colleagues (2013) reported that youths
in closed mentoring relationships were least likely to be re-
matched if they scored high on an index of personal risk (e.g.,
behavioral difficulties) and low on environmental risk (e.g.,
family stressors). However, none of the other factors in their
models (e.g., youth demographics) emerged as significant
predictors.
Gender, Mentoring Status, and Youth Outcomes
Gender-based theories pertaining to youth mentoring suggest
that girls may exhibit a more favorable response to mentoring
than boys might because they are more likely to derive their
sense of identity from social relationships, demand greater
intimacy, and when confronted with stressors, turn to signifi-
cant others for emotional support (Liang, Bogat, and Duffy
2013). Spencer (2007) has however challenged these assump-
tions noting that they have been based mostly on research
from non-mentored samples of youths. Drawing on qualitative
interviews with adult male mentors and adolescent boys
paired in one-to-one BBBS community mentoring relation-
ships, the author concluded that boys are just as likely as girls
to value emotional closeness in their mentoring relationships
and that the traditional societal norms and stereotypes of male
masculinity may not operate where mentors are trained to give
mentees emotional and instrumental support. To date, a sys-
tematic study of youth gender as a potential moderator of the
impact of mentoring on youth developmental outcomes has
not been conducted.
Current Study
This study will examine the association between categories of
youth mentoring status (broadly reflective of Keller’sstagesof
mentoring relationship development) and aspects of youth
health and social functioning. A secondary objective is to
examine the role of youth gender as a possible moderating
factor. These objectives will be met using the responses of
859 youths participating in a national study of BBBS one-to-
one community mentoring programs.
Program Description
BBBS community one-to-one mentoring programs (BBBS of
Canada 2014) are designed to give youths a one-to-one
mentoring relationship with a caring and responsible adult.
The programs assume that upon developing close and nurtur-
ing relationships with their mentees, mentors will cultivate
protective factors that promote positive youth development.
To qualify for the program, parents/guardians must have a
youth applicant between the ages of 6 and 17, permanently
resides in the BBBS agency’s catchment area, and they must
agree to agency rules. Mentors must be at least 18 years old
with no prior criminal record, provide three personal refer-
ences, and complete an assessment interview. For a minimum
of 1 year, they are expected to commit an average of 2–4hours
each week with their mentee engaging in recreational, leisure,
or skill-based activities. Qualified mentors must undergo
training delivered by program staff on child safety, application
and assessment procedures, roles and responsibilities, match
closure, caseworker support and supervision, recognizing
abuse, and optimizing match quality. To determine a match,
BBBS caseworkers conduct interviews with families and
mentors to assess common interests, mentor ability to meet
the needs of the youth, and personal preferences. A meeting is
Prev Sci

held with the parent, youth, mentor, and agency caseworker
allowing families and mentors time to become acquainted. For
a match to occur, all parties must be in agreement. BBBS
caseworkers provide ongoing support to the mentoring rela-
tionship (e.g., information on BBBS events and advice on
handling match difficulties). Mentors, youths, and parents
are contacted at least once a month for the first 6 months
and bi-monthly thereafter until 12 months. After 12 months,
contact is reduced to every 3 months. Unmatched youths are
assigned to a waiting list until a mentor is found. Waiting
times vary from just weeks to several months.
Method
Participants
Sample Selection and Recruitment
Between 2006 and 2008, 21 BBBS agencies across Canada
were recruited by Big Brothers Big Sisters of Canada (the
national umbrella organization) to participate in the study.
Agencies were chosen on the basis of having large annual
caseloads, a long history of operation, and cultural diversity
of clients. All but one chose to participate. The majority
(80 %) served metropolitan centers and accounted for about
half of all matches made on a year-to-year basis. In 2006, there
were over 120 agencies nationwide. Agencies were assigned a
quota of families to recruit based on their normal client case-
load. Over a 30-month period, 1,279 families were
approached by BBBS caseworkers to participate. To qualify
for the study, families had to be new admissions and passed
the agency’s qualifying assessment. Youth applicants had to
be 6–17 years old. In families with more than one eligible
youth, one was randomly selected to participate. Study
parents/guardians had to have primary parenting responsibili-
ty for the study youth. For joint custody or shared living ar-
rangements, the parent/guardian spending the most time with
the youth was enrolled. Among the families approached, 997
(78 %) agreed to participate and completed a baseline ques-
tionnaire. A comparison of participants with a sample of non-
participants (n=105) on selected demographic characteristics
(i.e., parent gender, age, education, marital status) revealed no
significant differences.
The sub-sample for the current study consists of 859 youths
with information on their mentoring status at the 18-month
follow-up. For youths who did not complete an 18-month
follow-up (n= 224/859), data on the duration and stability of
their mentoring relationship was obtained from the 24- and
30-month follow-ups and used to fill in the missing 18-
month values. The 18-month follow-up was chosen for
assessing youth outcomes because most youths had already
been mentored and many had been in a relationship for at least
12 months, the minimum period of mentor commitment re-
quired by BBBS. Non-completion of the 18-month follow-up
occurred because of scheduling difficulties (71 %) or study
drop out (29 %), the latter occurring primarily as a result of
family withdrawal from BBBS (43 %), a family move (22 %),
or lack of parent or youth interest/time (11 %). Excluded from
the sub-sample were 138 youths without an 18-month follow-
up and information on their mentoring status. These youths
were compared to the sub-sample of 859 on baseline demo-
graphics, perceived supports, and behaviors included in the
multivariate models (see below). Excluded youths had a mean
cognitive behavioral coping score (M= 11.61, SD = 4.65) less
than the score for included youths (M=12.54, SD =5.08) (F
(1, 995) = 4.12, p= .043) and were less likely to report a chron-
ic health condition (21.7 % vs. 30.5 %) (χ
2
(1, N= 996) = 4.41,
p=.036).
Sample Description
The sample (n= 859) consisted of slightly more boys than
girls (50.6 vs. 49.4 %). The average age of youths was
9.74 years (SD =2.20). Most (93 %) resided in urban areas.
Seventy percent were living with a single biological parent
(usually the female parent); 10.8 % with both biological par-
ents; and 19.4 % in other arrangements (e.g., step or foster
parents, grandparents). Thirty-three percent did not have any
siblings at home. Nearly one third (31.9 %) were from visible
ethnic/racial minority groups (non-Caucasian and non-White
in color) (e.g., 84 African Canadian, 108 Aboriginal, 64
Asian, 18 Hispanic). Forty-three percent were White
European, 10.9 % identified themselves as having a
Canadian nationality but with no specified ethnicity/race,
and 15.4 % fell in the category of “other unspecified”(com-
bination of multiple ethnic/racial backgrounds). Families
moved an average of 1.66 times in the past 5 years
(SD = 2.86); 33 % had not moved while 24.9 % moved three
or more times. Parents or guardians were mostly female
(93.1 %). The average age of mentee guardians was
40.28 years (SD = 8.75). Most were divorced, separated, or
widowed (46.3 %) or never married (34.8 %). Nearly two-
thirds (65.0 %) indicated at least some post-secondary educa-
tion; 15.9 % had not received a high school diploma. Nearly
40 % (39.3 %) reported an annual gross household income of
less than $20 K; one third were in receipt of government social
assistance, and under one third (28.8 %) were living in a sub-
sidized dwelling.
Procedures
To recruit families, BBBS agency caseworkers followed a
standardized script describing study objectives, types of ques-
tions asked of participants, and expectations concerning par-
ticipant roles and responsibilities. Families were informed that
Prev Sci

their participation was completely voluntary and that a refusal
to take part would not affect the quality of the service received
(i.e., the agency would proceed as usual with matching the
youth to a mentor). Interested parents/guardians signed the
script and recorded contact information authorizing a baseline
assessment. Families received an in-home baseline assess-
ment prior to a match to a mentor consisting of a 40-min
parent self-administered questionnaire and a 2-hour youth
face-to-face interview in a private room conducted by a
trained interviewer. Informed consent to participate was ob-
tained from youths and parents before the baseline assess-
ment. In-home follow-up assessments were conducted every
6 months until the last follow-up at 30 months. Youths report-
ed on their behavioral and psychological functioning. Parents
reported on their children in each of these areas and their own
social relationships and mental health outcomes. Matched
families also reported on the match determination process,
parent and agency match supports, and match characteristics
(e.g., quality and weekly contact). Youths received two movie
passes after each assessment and parents received a US$5 fast
food gift certificate. Study procedures were approved by the
Centre for Addiction and Mental Health research ethics board.
Measures
Youth outcomes were assessed using youth self-reports of
behavioral problems, poor mental health, perceived social
support, self-esteem, and coping strategies at the study base-
line and 18-month follow-up assessments.
Behavioral Problems Behavioral problems were assessed
using the conduct problems (5 items; α=0.50–0.60),
hyperactivity-inattention (5 items; α= 0.65–0.72), and pro-
social behavior (5 items reverse coded; α=0.60–0.64) sub-
scales of the Strengths and Difficulties Questionnaire (SDQ)
(Goodman and Scott 1999). The SDQ was developed as a
brief measure of psychological adjustment in children and
youths. Item response options range from “not true”to “cer-
tainly true.”Good reliability and validity of the SDQ sub-
scales has been reported in ethnically diverse samples of
youths across a broad age span (Achenbach et al. 2008;
Muris, Meesters, Eijkelenboom, and Vincken 2004). In the
current study, items on each sub-scale were summed. The
resulting indices were used as indicators of youth behavioral
problems in the SEM models.
Mental Health Youth mental health was measured using the
Generalized Social Anxiety and Distress sub-scale (SAD-G)
(four items; α= 0.60–0.64) of the Revised Social Anxiety
Scale for Children (SASC-R) (La Greca 1999) and eight items
(α= 0.76–0.83) from the Center for Epidemiology Studies
Depression Scale (CES-DC) (Weissman, Orvaschel, and
Padian 1980). Four response options on the SAD-G items
range from “not at all”to “all the time.”Four response options
on the CES-DC range from “not at all”to “a lot or all the
time.”The SAD-G has been administered to youths aged 7–
15 years and possesses moderate to good internal consistency
and test-retest reliability and good construct and criterion va-
lidity (La Greca 1999; Myers and Winters 2002). The CES-
DC indicates good internal consistency and retest stability in
12–18 year olds and moderate support for concurrent validity
(Myers and Winters 2002).
Coping Behaviors Coping behaviors were measured using
the Cognitive-Behavioral Problem-Solving (eight items;
α= 0.82–0.89) and Assistance Seeking (four items;
α= 0.57–0.67) sub-scales of the Coping Scale for Children
and Youth (CSCY) (Brodzinsky et al. 1992). Youths complet-
ing the CSCY rated how frequently they adopt a certain cop-
ing strategy to handle a problem. Four response options range
from “never”to “very often.”Previous research using the
CSCY has shown good internal consistency and test-retest
reliability (Dry et al. 2015). Evidence of construct validity
was reported in the original study of scale development
(Brodzinsky et al. 1992). Due to a low item-total correlation,
the item “Ikeptmyfeelingstomyself”was dropped from the
Assistance Seeking sub-scale.
Peer self-esteem Peer self-esteem was measured using a six-
item abbreviated version of the peer sub-scale (α=0.62–0.73)
of the HARE Self-Esteem Scale (Shoemaker 1980). Five re-
sponse options range from “strongly agree”to “strongly
disagree”.
Perceived social support Peer and teacher support were mea-
sured using items from the Peer Support (five items; α=0.68–
0.74) and Teacher Support (five items; α= 0.77–0.83) sub-
scales of the Social Support Appraisal Scale (SSAS)
(Dubow and Ullman 1989). The SSAS was developed to
gauge children’s perceptions of family, peer, and teacher sup-
port. Five response options range from “never”to “always.”
Sub-scales of the SSAS have demonstrated good reliability
(α= 0.78–0.83) and construct validity (Dubow and Ullman
1989). Emotional support from parents was measured using
the Perceived Emotional Support sub-scale (six items;
α= 0.83–0.90) of the 15-item Wills Parental Support Scale
(WPSS) (Wills, Vaccaro, and McNamara 1992). Four re-
sponse options range from “not at all”to “very much.”
TheWPSSwascreatedasashortmeasureoffunctional
support for children and adolescents. Youths indicate their
level of agreement with statements pertaining to perceived
support availability from parents. High internal consisten-
cy (α= 0.88) has been reported elsewhere (Wills and
Shinar 2000).
Prev Sci

Independent Variable Youth mentoring status reflected the
stability and longevity of the mentoring relationship in the
interval between the baseline assessment and the 18-month
follow-up. Dummy-coded categories included the following:
(1) continuously mentored less than 12 months; (2) continu-
ously mentored 12 or more months; (3) dissolved mentoring
relationship less than 12 months; (4) dissolved mentoring re-
lationship 12 or more months; and (5) re-matched to a second
mentor. Never-mentored (un-matched) youths served as the
comparison group.
Covariates Study covariates in the multivariate models (see
below) were chosen based on previous theory and research on
youth mentoring (e.g., Rhodes 2002). Youth characteristics
were gender (1 = boys 0 = girls), age (continuous), living ar-
rangements (two dummy-coded categories: living with a sin-
gle biological parent and living in other arrangements vs. a
reference group of both biological parents), ethnic/racial mi-
nority status reported by the parent/guardian (1 = Aboriginal,
African, Asian, and Hispanic Canadian 0= all others), number
of siblings at home (continuous), number of family moves
(past 5 years) (continuous), presence of a chronic health con-
dition (1= yes 0 =no), and sought help from a mental health or
social service professional in the past 12 months (1 = yes
0 = no). Parent/guardian characteristics included age
(continuous) and education (1 = <high school; 0 = other).
Parent-reported gross annual household income (1 = <20 K;
0 = other), receipt of government social assistance (1 = yes;
0 = no), and living in a government subsidized dwelling
(1 = yes; 0 = no) were combined into a count variable to cap-
ture the depth of family poverty. Other covariates reported by
parents included family functioning (α= 0.86), measured
using the 13-item general functioning sub-scale of the
McMaster Family Assessment Device (Byles, Bryne, Boyle,
and Offord 1988); parent depression (α= 0.93), measured
using the 20-item Center for Epidemiological Depression
Scale (Radloff 1977); parent social anxiety (α=0.92), mea-
sured using the 17-item Social Phobia Inventory (α= 0.92)
(Connor et al. 2000); and neighborhood problems (α=0.92),
measured using 6 items from the revised Simcha-Fagan
Neighborhood Questionnaire (McGuire 1997).
Analytic Approach
Structural equation modeling in M-plus (Muthen and Muthen
2011) was used to examine the relationship between
mentoring status and youth mental health, behavioral, and
developmental outcomes. At the 18-month follow-up, each
outcome was treated as a latent endogenous construct repre-
sented by three or more indicators and simultaneously
regressed on (1) youth mentoring status (five dummy-coded
categories); (2) scores for the same outcome measured at the
baseline assessment (latent exogenous construct represented
by three or more indicators); and (3) potential baseline con-
founders or model covariates (e.g., youth demographics, per-
sonal and environmental factors) measured without error and
treated as single observed indicators. Confounders were
allowed to co-vary with the mentoring status variable. To en-
hance statistical power and reduce skewness on the latent con-
struct indicators, items on the outcome measures were
parceled. SEM multiple groups was performed to assess the
plausibility of the model for male and female youths. To
establish measurement invariance, equality constraints were
imposed on the factor loadings and intercepts of the latent
constructs across time and gender groups. Cross group equal-
ity constraints were imposed to assess statistically significant
differences between boys and girls in the magnitude and di-
rection of the pathway linking mentoring status to youth out-
comes. Standard errors were adjusted for nesting of youths
within mentoring agencies. Full Information Maximum
Likelihood (Graham 2009) was used for handling missing
data on outcomes at baseline (2 %) and follow-up (26 %).
Results
Tab le 1(top portion) provides a percentage breakdown of the
youth mentoring status construct for the total sample and then
separately by gender. To validate the construct, the first five
categories were cross-classified with mentoring relationship
characteristics (quality, weekly contact, and number of
mentoring activities) identified as important benchmarks for
defining successful mentoring relationships (Herrera, Sipe,
and McClanahan 2000). The means and standard deviations
associated with these variables are reported in the bottom por-
tion of Table 1. Results showed that youth mentoring status
accurately discriminated between two of the three benchmark
indicators. Statistically significant associations occurred be-
tween relationship quality and youth mentoring status.
Consistent with previous research showing a strong positive
association between mentoring relationship quality and lon-
gevity (Rhodes et al. 2014), post hoc comparisons revealed
significantly higher mean quality values on the long-term
mentoring categories relative to the short-term dissolved and
re-matchedcategories. A statistically significant association in
the expected direction was also found for number of monthly
activities with a higher mean activity value for youths in con-
tinuous long-term versus short-term dissolved relationships.
There was no association between weekly contact and
mentoring status.
Prior to estimating the SEM models, differences across
categories of youth mentoring status for each baseline
outcome were explored. Statistically significant group dif-
ferences in the estimated mean values were found for pro-
social behavior (F(5, 853) = 4.10, p= .001), parent emo-
tional support (F(5, 853) = 4.93, p= .001), and peer
Prev Sci

support (F(5, 853) = 4.41, p= .001). Post hoc comparisons
revealed significant differences between non-mentored
youths and youths in the long-term mentoring categories
(with lower mean pro-social and social support values for
non-mentored youths). These variables were included as
covariates in the SEM models.
Tab le 2presents standardized structural coefficients in each
of the SEM models for the pathways leading from the
dummy-coded mentoring status variable to the youth
behavioral, social support, and mental health outcomes (latent
endogenous constructs). Estimates are compared against a
comparison group of never-mentored youth and adjusted for
the corresponding baseline outcomes along with demographic
and environmental covariates. Details of the measurement and
structural components of the model are provided in an adden-
dum. For the total sample, results showed that youths in long-
term stable or dissolved mentoring relationships experienced
better outcomes than did never-mentored youths with statisti-
cally significant negative relationships occurring for behavior-
al problems and symptoms of depression and social anxiety
and positive relationships occurring for coping strategies and
parent emotional support. Entering a second mentoring rela-
tionship was also beneficial. Specifically, statistically signifi-
cant positive relationships occurred between the re-matched
category and cognitive behavioral coping and perceived
parent support. Membership in short-term dissolved
mentoring relationships was not a statistically significant cor-
relate in any of the models.
Turning to gender, girls in long-term relationships (ongoing
and dissolved) did relatively better than did never-mentored
girls with statistically significant negative relationships occur-
ring for behavioral problems and depressed mood and positive
relationships for self-esteem. For girls and boys, involvement
in long-termdissolved relationships was negatively associated
with symptoms of depressed mood and social anxiety relative
to never-mentored youth. Unlike girls, boys in long-term
mentoring relationships were more likely than never-
mentored boys to report stronger perceptions of emotional
support from peers and parents. Girls re-matched to a second
mentor performed better than never-mentored girls did on
several outcomes with statistically significant negative rela-
tionships occurring for depressed mood and social anxiety
and positive relationships for peer self-esteem, coping, and
parent emotional support. In contrast, re-matched boys either
showed no difference or for some outcomes (e.g., peer self-
esteem) did significantly worse thannever-mentored boys did.
Statistically significant cross-gender comparisons in the mag-
nitude of the estimated parameters occurred for re-matched
youths for peer self-esteem (χ
2
(1) = 13.91, p= <.001) and
depressed mood (χ
2
(1) = 4.85, p=.03).
Tabl e 1 Description and validation of youth mentoring status construct
Con < 12 (Cat A) Con 12 + (Cat B) Dis < 12 (Cat C) Dis 12 + (Cat D) Rematch (Cat E) No match
Tot al
1
15.3 29.5 12.8 8.1 9.7 24.7
Girls
1
8.0 34.4 17.9 10.6 12.5 16.5
Boys
1
22.3 24.6 7.8 5.7 6.9 32.6
Quality MR (PR)
2
13.93
c,e
(1.52)
14.21
c,d,e
(1.34)
11.42
a,b,d
(3.21)
13.34
b,c,e
(2.09)
12.19
a,b,d
(2.65)
–
F(4, 582) = 38.50, p< .001
Quality MR (CR)
2
14.5
c,e
(1.31)
14.61
c,e
(.89)
12.70
a,b,d
(2.56)
14.33
c,e
(1.10)
13.22
a,b,d
(2.27)
–
F(4, 588) = 32.37, p< .001
Contact MR
2
.81
(.40)
.83
(.43)
.76
(.45)
.83
(.36)
.76
(.43)
–
F(4, 587) = 0.87, p=.48
Activities MR
2
7.49
b
(3.74)
8.76
c
(3.90)
6.69
b
(3.92)
7.93
(4.19)
7.68
(3.38)
–
F(4, 532) = 4.42, p= .002
Note: Mentoring relationship quality was assessed on the basis of youth and parent global feelings of relationship closeness, warmth, trust, respect, and
happiness (α=0.86–0.88). Scale items were summed to form separate indices, one for youth and one for parents, each with a minimum value of 5 and a
maximum value of 15. Weekly contact was measured by the number of days in a typical week youth reported doing things with their mentor. Number of
activities reflected the number of different mentoring activities in the past month youth reported sharing with their mentor (e.g., playing a sport).Gender
comparisons. Category A: (χ
2
(1, N= 859) = 33.80, p< .001), Category B: (χ
2
(1, N= 859) = 9.99, p= .002), Category C: (χ
2
(1, N= 859) = 19.65,
p< .001), Category D: (χ
2
(1, N= 859) = 6.79, p= .009), Category E: (χ
2
(1, N= 859) = 7.72, p= .005), Category F: (χ
2
(1, N= 859) = 30.07, p< .001).
Superscript letters means post-hoc comparisons using the Games-Howell test (p<.05).
PR Parent Report, CR Child Report
1
Percentages
2
Means with standard deviations shown in parentheses
Prev Sci

Tabl e 2 Mentoring status and
youth psychosocial outcomes Mentoring status
a
Outcome Con < 12 Con 12 + Dis < 12 Dis 12 + Rematch
Behavioral problems
Total −0.09 (0.05) −0.13 (0.05)* 0.05 (0.07) −0.06 (0.04) 0.02 (0.03)
Girls −0.04 (0.06) −0.22 (0.10)* −0.01 (0.13) −0.15 (0.06)* −0.12 (0.10)
Boys −0.10 (0.07) −0.07 (0.07) 0.09 (0.08) −0.01 (0.06) 0.12 (0.07)
Tota l m o d el:X
2
=192.67;df = 97;p = .000;RMSEA= 0.034;CFI = 0.92;TLI = 0.88
Gender model:X
2
=307.16;df = 194;p=.000;RMSEA = 0.037;CFI = 0.92;TLI = 0.87
Depressed mood
Total −0.10 (0.04)* −0.12 (0.05)** −0.06 (0.05) −0.18 (0.03)*** −0.06 (0.04)
Girls −0.13 (0.07) −0.18 (0.05)*** −0.12 (0.09) −0.19 (0.04)*** −0.15 (0.05)**
Boys −0.07 (0.07) −0.06 (0.08) 0.05 (0.04) −0.20 (0.05)*** 0.03 (0.06)
Tota l m o d el:X
2
=168.86;df = 159;p = .281;RMSEA = 0.008;CFI= 0.99;TLI = 0.99
Gender model:X
2
=326.22;df = 318;p = .32;RMSEA= 0.008;CFI = 0.99;TLI = 0.99
Social anxiety
Total −0.08 (0.04) −0.05 (0.04) −0.08 (0.04) −0.13 (0.02)*** −0.05 (0.04)
Girls −0.02 (0.06) −0.07 (0.07) −0.12 (0.07) −0.14 (0.05)** −0.10 (0.04)*
Boys −0.10 (0.06) −0.07 (0.06) −0.02 (0.05) −0.13 (0.04)*** −0.01 (0.05)
Tota l m o d el:X
2
=222.43;df = 159;p = .001;RMSEA = 0.022;CFI= 0.96;TLI = 0.94
Gender model:X
2
=436.91;df = 318;p=.000;RMSEA = 0.030;CFI = 0.92;TLI = 0.90
Peer self-esteem
Total 0.02 (0.05) 0.05 (0.06) 0.06 (0.08) 0.02 (0.04) 0.04 (0.06)
Girls −0.01 (0.08) 0.19 (0.06)** 0.13 (0.11) 0.09 (0.04)* 0.23 (0.06)***
Boys 0.01 (0.08) −0.02 (0.08) 0.03 (0.07) −0.01 (0.04) −0.12 (0.06)*
Tota l m o d el:X
2
=153.12;df = 101;p = .001;RMSEA = 0.025;CFI= 0.95;TLI = 0.92
Gender model:X
2
=274.01;df = 202;p=.001;RMSEA = 0.029;CFI = 0.93;TLI = 0.90
Coping (cognitive)
Total −0.01 (0.04) 0.11 (0.05)* 0.02 (0.07) 0.04 (0.04) 0.09 (0.04)*
Girls 0.10 (0.07) 0.17 (0.09) 0.12 (0.12) 0.05 (0.07) 0.17 (0.06)**
Boys −0.07 (0.06) 0.09 (0.07) −0.05 (0.06) 0.04 (0.07) 0.05 (0.05)
Tota l m o d el:X
2
=217.21;df = 159;p = .002;RMSEA = 0.021;CFI= 0.98;TLI = 0.97
Gender model:X
2
=398.75;df = 318;p=.001;RMSEA = 0.024;CFI = 0.97;TLI = 0.97
Coping (assistance)
Total 0.01 (0.06) 0.12 (0.05)* 0.05 (0.06) 0.10 (0.04)* 0.06 (0.04)
Girls 0.02 (0.10) 0.12 (0.09) 0.14 (0.11) 0.12 (0.07) 0.14 (0.05)**
Boys −0.03 (0.06) 0.15 (0.06)** −0.06 (0.08) 0.06 (0.06) −0.02 (0.05)
Tota l m o d el:X
2
=151.70;df = 101;p = .001;RMSEA = 0.024;CFI= 0.94;TLI = 0.90
Gender model:X
2
=266.89;df = 202;p=.002;CFI = 0.92;TLI = 0.88
Parent emotional support
Total 0.10 (0.04)* 0.09 (0.04)* 0.06 (0.06) −0.01 (0.04) 0.07 (0.035)
Girls 0.10 (0.05) 0.08 (0.07) 0.11 (0.11) −0.01 (0.05) 0.11 (0.04)**
Boys 0.10 (0.06) 0.12 (0.05)** −0.03 (0.04) −0.01 (0.07) 0.03 (0.05)
Tota l m o d el:X
2
=102.31;df = 97;p = .336;RMSEA= 0.008;CFI = 1.00;TLI = 0.99
Gender model:X
2
=215.87;df = 194;p=.063;RMSEA = 0.016;CFI = 0.99;TLI = 0.98
Peer social support
Total 0.13 (0.05)* 0.12 (.07) 0.06 (0.06) 0.04 (0.06) 0.04 (0.05)
Girls 0.08 (0.05) 0.06 (.09) 0.08 (0.08) 0.01 (0.10) 0.06 (0.09)
Boys 0.15 (0.07)* 0.20 (.09)* −0.04 (0.08) 0.11 (0.05)* 0.03 (0.07)
Tota l m o d el:X
2
=190.26;df = 97;p = .000;RMSEA= 0.033;CFI = 0.91;TLI = 0.86
Gender model:X
2
=350.18;df = 194;p=.000;RMSEA = 0.043;CFI = 0.86;TLI = 0.79
Prev Sci

Discussion
This study moved beyond a simple comparison of mentored
versus non-mentored youths to create a mentoring status var-
iable with a range of different mentoring categories broadly
corresponding to the stages of mentoring described in Keller’s
theory of mentoring relationship development. It was expect-
ed that these mentoring categories would differentiate youths
achieving positive developmental outcomes from those who
did not. A secondary objective was to explore the roleof youth
gender as a moderator of these relationships.
In general, this study provided support for a more refined
mentoring construct. After adjusting for potential confounders
known to influence the development of mentoring relation-
ships and positive youth outcomes (i.e., personal and environ-
mental factors), youths in mentoring relationships lasting 12
or more months experienced health and social benefits com-
pared to never-mentored youths. This result supports existing
models of youth mentoring (e.g., Rhodes et al. 2006) which
posit that the processes of change through which mentoring
influences youth social-emotional, identity, and cognitive de-
velopment (e.g., changes in the way youth view themselves or
approach relationships with others) take time to fully develop
such that positive changes in development are not likely to
surface until well into the growth and maintenance phase of
the mentoring relationship.
Why youths in dissolved long-term relationships faired just
as well as those in ongoing long-term relationships is unclear.
To explore this question, a close examination of the study data
revealed that two thirds of the dissolved relationships ended
exactly at 12 months. This finding could mean that the BBBS
policy of a minimum 12-month commitment may have had
the unintended effect of leading some mentors and mentees
early in the program to anticipate, expect, or even plan for
relationships to close at the end of the 12-month period. If
many of these mentors and mentees parted on good terms, it
might explain why youths in the dissolved long-term category
benefited just as much as youths in the ongoing category.
Unfortunately, it is not possible to test this hypothesis. The
current study did not collect information on mentor or mentee
expectations around relationship continuity or possible pre-
arranged informal agreements for relationships to end to co-
incide with the minimum 12-month period of commitment.
The absence of health and social benefits associated with
youth involvement in short-term dissolved relationships is al-
so an important finding. The null findings may simply reflect
youth disappointment with mentoring relationships gone
awry, as suggested by the lower quality ratings for this group
in the descriptive results, or negative feelings associated with
unplanned endings. Several authors (e.g., Goldner and
Mayseless 2009; Spencer, Basualdo-Delmonico, Walsh, and
Drew 2014) have suggested that mentoring relationships that
dissolve early may be damaging to youths because of feelings
of rejection or abandonment or unfulfilled expectations.
Re-matched youth did not experience health and social
benefits on most outcomes. This finding is somewhat consis-
tent with previous investigations (e.g., Grossman et al. 2012;
Herrera et al. 2013) showing that being paired to a second
mentor had either no effect and in some instances negative
effects on youth outcomes. One explanation for these findings
is that many re-matched youths may have recently left un-
pleasant relationships with their first mentor and were in the
process of getting acquainted with their new partner.
However, due to the small sample, comparing re-matched
youths in short- versus long-term relationships was not
possible.
Finally, youth in ongoing relationships lasting less than
12 months also experienced positive outcomes relative to
never-mentored youth. Youth in this category started their
mentoring relationships later and were still being mentored
at follow-up. Given the shorter duration of their relationships,
it possible that some were still experiencing the excitement
and euphoria often associated with meeting someone new
for the first time (Keller’s get acquainted phase) and therefore
had developed an exaggerated sense of satisfaction reflective
of a “honeymoon”effect that gave rise to their better standing
relative to never-mentored youth. The positive benefits may
have also occurred because many of the youth in the ongoing
short-term mentoring category had been in a mentoring rela-
tionship lasting 10 or 11 months allowing for sufficient time
Tabl e 2 (continued) Mentoring status
a
Tea che r sup po rt
Total 0.10 (0.05) 0.07 (0.06) 0.08 (0.08) 0.01 (0.03) 0.01 (0.05)
Girls 0.12 (0.07) 0.09 (0.10) 0.10 (0.14) 0.02 (0.05) 0.12 (0.07)
Boys 0.07 (0.08) 0.09 (0.07) 0.05 (0.06) 0.02 (0.04) −0.15 (0.08)
Tota l m o d el:X
2
=143.99;df =101;p=.003;RMSEA= 0.022; CFI = 0.97; TLI = 0.95
Gender model:X
2
=240.68;df =202;p=.032;RMSEA= 0.021; CFI = 0.97; TLI = 0.96
Note: results are standardized structural path coefficients with standard errors shown in parentheses
*p<.05;**p<.01;***p< .001
a
Comparison group: never-mentored or unmatched youths
Prev Sci

for bonding to develop between the mentor and mentee. At
least some of these youths would have continued beyond the
12-month mark following the 18-month interview.
Gender differences in the relationship between youth
mentoring status and development varied by the category of
the mentoring status variable and the developmental outcome.
Girls in long-term relationships experienced fewer behavioral
problems and higher self-esteem while boys experienced
greater emotional support from peers and parents. Moreover,
re-matched girls did relatively better than never-mentored
girls did. In contrast, re-matched boys did not experience ben-
efits and for some outcomes did more poorly than never-
mentored boys did. These results support Spencer’s(2007)
position that the traditional gender-based stereotypes applied
to non-mentored samples of youth do not necessarily apply to
mentoring relationships and that boys are just as likely as girls
to value emotional closeness in their mentoring relationships
resulting in positive outcomes. The varied findings could also
mean that boys and girls have different expectations of what
they hope mentoring will achieve or possibly that mentoring
fulfills certain needs related to gender at different times in the
evolution of the relationship (Pryce, Kelly, and Guidone
2013). Investigating these possibilities will require careful
study of the motivational factors and relational processes
influencing the development of the mentoring relationship
tied to youth gender and that might be responsible for the
observed differences in this study.
The finding that re-matched girls experienced positive out-
comes and not re-matched boys could mean that girls expect
more out of their mentoring relationships than boys do and
therefore are less willing to settle for relationships that fall
short of satisfying their needs for intimacy. If true, girls who
leave their first mentoring relationships may be more highly
motivated than boys are to enter a second relationship, choos-
ing a mentor that is a better fit with their interests, and conse-
quently, demonstrating a greater commitment to making the
relationship work. An alternative explanation is that re-
matched girls left their first relationship sooner than boys did
providing more time to become fully acquainted with their
new mentor. These explanations, however, are speculative
and do not explain why re-matched boys demonstrated wors-
ened behavior on some outcomes. Further study is needed to
explain this finding.
The results of this study are consistent with previous stud-
ies (e.g., DuBois et al. 2011) showing statistically significant
weak to moderate-sized positive associations between
mentoring and youth health and social functioning.
However, the magnitude of the associations was less than
expected. One possible reason for the absence of stronger
relationships is that mentees and mentors who reach orsurpass
the12-month period of commitment may make downward ad-
justments to their time commitments to closely conform to
what they perceive to be desirable or realistic. DuBois and
Neville (1997) reported that contact between mentors and
their mentees became less frequent the greater the longevity
of the relationship. The weak relationships could also be due
to reduced program supports. In most Canadian BBBS pro-
grams, ongoing support tomentoring relationships isprovided
by BBBS caseworkers but is reduced after 12 months. Last,
the results could have been due to low scores on the baseline
outcomes that created a floor effect limiting any improvement
associated with having an adult mentor. To explore this pos-
sibility, the baseline means of several outcomes were com-
pared against those reported in general population samples
(e.g., La Greca 1999;Murrisetal.2004;Stopa,Barrett,and
Golingi 2010). For the most part, youth in this study recorded
similar or higher means for problem behavior and mental
health outcomes and lower means for coping behavior ruling
out a floor effect.
Limitations
This study benefited from a good response rate and a large
ethnically diverse sample of youths. However, there were also
some limitations. First, the sample was comprised primarily of
youths from metropolitan centers and thus should not be
viewed as representative of all BBBS community applicants.
Second, the finding that youths in long-term mentoring rela-
tionships experienced positive outcomes may have occurred if
they were well-adjusted to begin with, making it less likely
that sustained mentor support was a contributing factor.
Preliminary analyses found that these youths demonstrated
higher levels of baseline parent and peer support and pro-
social behavior than other groups did. Although these factors
were included as model covariates, it is possible that youths in
this category benefited more because they received at the out-
set encouragement and support from parents that enabled
them to continue their relationships or because they reported
pro-social attitudes and behaviors reflecting a greater readi-
ness for change and that contributed to greater mentor satis-
faction with the relationship. Third, it is possible that youths in
long-term mentoring relationships did better because they ex-
pected to do so or because of similar expectations by parents.
These reporting biases may have impacted the results. Fourth,
while this study found positive benefits associated with youths
in long-term mentoring relationships, it did not examine how
benefits were achieved. Keller (2005) suggested that long-
term mentoring relationships allow for vital aspects of rela-
tionship quality (e.g., affection) to mature and that relation-
ships that possess these qualities are likely to have the greatest
impact on youth development. Future investigations are re-
quired to test these mediational pathways. Finally, this study’s
reliance on youth self-reports may be problematic since
youths may exaggerate behaviors or withhold sharing person-
al feelings (Brener, Billy, and Grady 2003). Future analyses
will address these biases by examining parent perspectives.
Prev Sci

Implications for Programming
The results of this study have important implications for the
delivery of effective mentoring programs. To reduce early
mentoring relationship closures and increase the positive im-
pact of mentoring on youth development,mentoring programs
may need to implement more effective methods of screening
to ensure that prospective mentors are committed to helping
youth succeed over the long-term, improve the process of
pairing mentors to mentees to account for the needs and
expectations of both parties, and offer enhanced training en-
couraging mentors and mentees to meet or surpass the expect-
ed 1-year period of commitment to the mentoring relationship.
Programs may also need to apply greater scrutiny to relation-
ships at the early stages of development through more intense
caseworker contacts aimed at identifying problems before
they escalate. At the same time, programs should raise aware-
ness that relationship closures will happen and seek to mini-
mize harm by ensuring that mentors and mentees notify each
other of their intention to end the relationship and explain their
decision in a sensitive manner. Finally, mentored girls and
boys experienced health and social benefits that varied by
their mentoring status. Further study of what underlies these
differences is required in advance of making gender-specific
recommendations around program content and practices.
Compliance with Ethical Standards
Funding Funding for this research was provided by grants from the
Canadian Institutes of Health Research (MOP 81115 and MOP
130435). The authors are grateful for the kind support of BBBS of
Canada and the participating agencies.
Conflicts of interest The authors declare that they have no conflicts of
interest.
Ethical approval All study procedures were conducted in accordance
with the ethical standards of the CAMH Research Ethics Board and the
1964 Helsinki Declaration and later amendments or comparable ethical
standards.
Informed consent Informed consent was obtained from all individual
study participants.
References
Achenbach, T. M., Becker, A., Dopfner, M., Heiervang, E., Roessner, V.,
Steinhausen, H.-C., & Rothenberger, A. (2008). Multicultural as-
sessment of child and adolescent psychopathology with ASEBA
and SDQ instruments: research findings, applications, and future
directions. Journal of Child Psychology and Psychiatry, 49,251–
275. doi:10.1111/j.1469-7610.2007.01867.x.
Big Brothers Big Sisters of Canada. (2014). Big Brothers Big Sisters of
Canada National Learning Center. http://www.bigbrothersbigsisters.
ca/en/Home/default.aspx.
Brener, N. D., Billy, J. O. G., & Grady, W. R. (2003). Assessment of the
factors affecting the validity of self-reported health risk behavior
among adolescents: evidence from the scientific literature. Journal
of Adolescent Health, 33,436–457.
Brodzinsky, D. M., Elias, M. J., Steiger, C., Simon, J., Gill, M., & Hitt, J.
C. (1992). Coping Scale for children and youth: scale development
and validation. Journal of Applied Developmental Psychology, 13,
195–214.
Byles, J., Bryne, C., Boyle, M. H., & Offord, D. R. (1988). Ontario child
health study: reliability and validity of the general functioning sub-
scale of the McMaster family assessment device. Family Process,
27,97–104.
Connor, K. M., Davidson, J. R. T., Churchill, L. E., Sherwood, A.,
Foa, E., & Weisler, R. H. (2000). Psychometric properties of the
social phobia inventory (SPIN). British Journal of Psychiatry,
176,379–386.
Dolan, P., Brady, B., O’Regan, C., Russell, D., Canavan, J., & Forkan, C.
(2011). Big brothers big sisters of Ireland: evaluation study. Report
1: RCT and implementation report. Galway: Child and Family
Research Centre, National University of Ireland, Galway.
Dry, S. M., Kane, R. T., & Rooney, R. M. (2015). An investigation into
the role of coping in preventing depression associated with perfec-
tionism in pre-adolescent children. Frontiers in Public Health, 3,1–
9. doi:10.3389/pubh.2015.00190.
DuBois, D. L., & Neville, H. A. (1997). Youth mentoring: investigation
of relationship characteristics and perceived benefits. American
Journal of Community Psychology, 25,227–234.
DuBois, D. L., Portillo, N., Rhodes, J. E., Silverhorn, N., & Valentine, J.
C. (2011). How effective are mentoring programs for youth? A
systematic assessment of the evidence. Psychological Science in
the Public Interest, 12,57–91. doi:10.1177/1529100611414806.
Dubow, E. F., & Ullman, D. G. (1989). Assessing social support in ele-
mentary school children: the survey of Children’s social support.
Journal of Clinical Child Psychology, 18,52–64.
Eddy, J.M., Herrera, D., Martinez, C.R., Grossman, J., Cearley, J.J.,
Harachi, T., Kjellstrand, J., Haggerty, K., Wheeler, A., Feldman,
B., Seeley, J., & Gau, J. (2015). Twelve-year professional youth
mentoring program for high risk youth: Continuation of a longitu-
dinal randomized controlled trial. US Department of Justice.
Document No. 248595.
Furano, K., Roaf, P. A., Styles, M. B., & Branch, A. Y. (1993). Big
Brothers/Big Sisters: a study of program practices. Philadelphia,
PA: Public/Private Ventures.
Goldner, L., & Mayseless, O. (2009). The quality of mentoring relation-
ships and mentoring success. Journal of Youth and Adolescence, 38,
1339–1350. doi:10.1007/s10964-008-9345-0.
Goodman, R., & Scott, S. (1999). Comparing the strengths and difficul-
ties questionnaire and the child behavior checklist: is small beauti-
ful? JournalofAbnormalChildPsychology,27,17–24.
Graham, J. W. (2009). Missing data analysis: making it work in the real
world. Annual Review of Psychology, 60,549–576. doi:10.1146/
annurev.psych.58.110405.085530.
Grossman, J. B., & Rhodes, J. E. (2002). The test of time: predictors and
effects of duration in youth mentoring relationships. American
Journal of Community Psychology, 30, 199–219. doi:10.1023/
A:1014680827552.
Grossman, J. B., Chan, C. S., Schwartz, S. E. O., & Rhodes, J. E. (2012).
The test of time in school-based mentoring: the role of relationship
duration and re-matching on academic outcomes. American Journal
of Community Psychology, 49,43–54.
Herrera, C., Sipe, L., & McClanahan, W. S. (2000). Mentoring school-
age children: relationship development in community and school-
based programs. Philadelphia, PA: Public/Private Ventures.
Herrera, C., DuBois, D. L., & Grossman, J. B. (2013). The role of risk:
mentoring experiences and outcomes for youth with varying risk
profiles. New York, NY: Public/Private Ventures.
Prev Sci

Keller, T. E. (2005). The stages and development of mentoring relation-
ships. In D. L. DuBois & M. J. Karcher (Eds.), Handbook of youth
mentoring (pp. 82–99). Thousand Oaks, CA: Sage Publications.
La Greca, A. M. (1999). Social anxiety scales for children and adoles-
cents. Manual and instructions for SASC, SASC-R, SAS-A
(adolescents) and parent versions of the scales. Miami, FL: Author.
Liang, B., Bogat, A., & Duffy, N. (2013). Gender in mentoring relation-
ships. In D. L. DuBois & M. J. Karcher (Eds.), Handbook of youth
mentoring (pp. 159–175). Thousand Oaks, CA: Sage Publications.
Lymburner, J. (2006). Predictors of match duration in a Big Sisters
mentoring program. (Ph.D. dissertation), Simon Fraser University.
Retrieved from http://summit.sfu.ca/item/6640.
McGuire, J. B. (1997). The reliability and validity of a questionnaire
describing neighborhood characteristics relevant to families and
young children living in urban areas. Journal of Community
Psychology, 25,551–566.
Meyerson, D. A. (2013). Mentoring youth with emotional and behavioral
problems: a meta-analytic review. (Doctor of Philosophy (PhD)),
DePaul University. Retrieved from http://via.library.depaul.edu/
csh_etd/56
Moodie, M. L., & Fisher, J. (2009). Are youth mentoring programs good
value for money? An evaluation of the Big Brothers Big Sisters
Melbourne Program. BMC Public Health, 9,41–49. doi:10.1186/
1471-2458-9-41.
Muris, P., Meesters, C., Eijkelenboom, A., & Vincken, M. (2004). The
self-report version of the strengths and difficulties questionnaire: its
psychometric properties in 8–13 year old non-clinical children.
British Journal of Clinical Psychology, 43,437–448.
Muthen, L. K., & Muthen, B. O. (2011). M-plus. Statistical analysis with
latent variables. User’sguide. Los Angeles, CA: Authors. Version
6.11.
Myers, K., & Winters, N. C. (2002). Ten-year review of rating scales II:
scales for internalizing disorders. Journal of the American Academy
of Child and Adolescent Psychiatry, 41,634–659.
Pryce, J., Kelly, M. S., & Guidone, S. R. (2013). Mentor and youth
matching. In D. L. DuBois & M. J. Karcher (Eds.), Handbook of
youth mentoring (2nd ed., pp. 427–439). Thousand Oaks, CA: Sage
Publications.
Radloff, L. S. (1977). The CES-D Scale: a self-report depression scale for
research in the general population. Applied Psychological
Measurement, 1,385–401.
Rhodes, J. (2002). Stand by me: the risks and rewards of mentoring
today’syouth. Cambridge Massachusetts: Harvard University Press.
Rhodes, J. E. (2008). Improving youth mentoring interventions through
research based practice. American Journal of Community
Psychology, 41,35–42.
Rhodes, J., Spencer, R., Keller, T. E., Liang, B., & Noam, G. (2006). A
model for the influence of mentoring relationships on youth devel-
opment. Journal of Community Psychology, 34,691–707. doi:10.
1002/jcop.20124.
Rhodes, J., Lowe, S. R., Litchfield, L., & Walsh-Samp, K. (2008). The
role of gender in youth mentoring relationship formation and dura-
tion. Journal of Vocational Behavior, 72,183–192. doi:10.1016/j.
jvb.2007.09.005.
Rhodes, J.E., Schwartz, S.E.O., Willis, M.M., & Wu, M.B. (2014).
Validating a mentoring relationship quality scale: Does match
strength predict match length? Youth and Society, 1-23, doi: 10.
1177/0044118X14531604.
Rodriguez-Planas, N. (2014). Do youth mentoring programs change the
perspectives and improve the life opportunities of at-risk youth? IZA
WorldofLabor,62.doi:10.15185/izawol.62
Shoemaker, A. L. (1980). Construct validity of area-specific self-esteem:
the Hare Self-Esteem Scale. Educational and Psychological
Measurement, 40,495–501.
Spencer, R. (2007). “I just feel safe with him”: emotional closeness in
male youth mentoring relationships. Psychology of Men and
Masculinity, 8,185–198.
Spencer, R., Basualdo-Delmonico, A., Walsh, J., & Drew, A. L. (2014).
Breaking up is hard to do: a qualitative interview study of how and
why youth mentoring relationships end. Youth and Society.doi:10.
1177/0044118X14535416.
Stopa, J. E., Barrett, P. M., & Golingi, F. (2010). The prevention of
childhood anxiety in socioeconomically disadvantaged communi-
ties: a universal school-based trial. Advances in School Mental
Health Promotion, 3,5–24.
Tolan, P. H., Henry, D. B., Schoeny, M. S., Lovegrove, P., & Nichols, E.
(2014). Mentoring programs to affect delinquency and associated
outcomes of youth at risk: a comprehensive meta-analytic review.
JournalofExperimentalCriminology,10,179–206. doi:10.1007/
s11292-013-9181-4.
Weissman, M. M., Orvaschel, H., & Padian, N. (1980). Children’s symp-
tom and social functioning self-report scales: comparison of
mother’s and children’s reports. Journal of Nervous and Mental
Disease, 168,736–740.
Wills, T. A., & Shinar, O. (2000). Measuring perceived and received
social support. In C. Sheldon, L. G. Underwood, & B. H. Gottlieb
(Eds.), Social support measurement and intervention: a guide for
health and social scientists (pp. 86–135). New York: Oxford
University Press.
Wills,T. A., Vaccaro, D., & McNamara, G. (1992). The role of life events,
family support, and competence in adolescent substance use: a test
of vulnerability and protective factors. American Journal of
Community Psychology, 20,349–374. doi:10.1007/BF00937914.
Prev Sci