Relationships Among Supports and Barriers and Career and Educational Outcomes

Abstract

Interest in the roles of environmental supports and barriers in career and educational development has increased steadily over the past few decades, enough to warrant a meta-analysis of this vast and still growing literature. The current study presents the results of a meta-analytic investigation, employing 276 samples drawn from 249 published articles (N = 104,440), on the relationships of supports and barriers to nine different career and educational outcomes. Employing a random effects meta-analytic model and sampling and measurement error-corrected effect size estimates, the study found that supports tended to account for more variance (M = 10%) across all outcomes than did barriers (M = 3%). Several moderators were also found, suggesting that (a) men’s self-efficacy beliefs and outcome expectations are more strongly related to supports than women’s, (b) Latino(a) students’ levels of school engagement and performance are more weakly related to supports than White students, and (c) supports (and perhaps barriers) seem to be more highly related to elementary school students’ levels school engagement than high school students. The implications of the results are discussed as well as potential avenues for future investigation suggested by some gaps in the literature uncovered in this meta-analysis.

Keywords

Meta-analysis supports barriers educational development career development

Interest in understanding the roles of environmental barriers and supports in the career and educational development process has mushroomed over the past several decades, with interest in the possible career limiting aspects of barriers predating concerted attention (at least in the vocational psychology literature) on the potential facilitating influences of supports. The barriers construct entered vocational psychology discourse primarily through writings on women’s career development (e.g., Farmer, 1976; Fitzgerald & Crites, 1980; O’Leary, 1974) as a potential explanatory mechanism for women’s limited career aspirations and achievements.

Research on the barriers construct grew rapidly, facilitated by the development of barriers measures (e.g., McWhirter, 1997; McWhirter, Rasheed, & Crothers, 2000; Swanson & Tokar, 1991b), to encompass other marginalized groups (see Brown & Lent, 2016). Today, most scholars agree that the experience or perception of barriers is negatively related to a variety of different educational (e.g., persistence, performance, and school engagement) and career (aspirations, expectations, self-efficacy beliefs, and outcome expectations) outcomes. However, the magnitudes of these effects are still unknown. Also unknown at this time is the comparative magnitudes of barriers’ relationships to different types of career and educational outcomes. For example, there seems to be consensus that barriers may limit career aspirations and negatively affect academic persistence, but we do not know whether barriers may similarly relate to aspirations and academic persistence or whether they might have a more potent influence on one rather than the other.

Finally, there are several potential moderators of the negative relationships between barriers and educational and career outcomes that have yet to be addressed in a systematic way, most notably, gender, race/ethnicity, and year in school. For example, both women and persons of color have been found consistently (e.g., Luzzo, 1993, 1995; McWhirter, 1997) to report more barriers to career and educational development than do men and Caucasians. However, unknown is whether gender or race/ethnicity has a differential impact on how barriers relate to educational and career outcomes—are there gender and race/ethnicity differences in the negative associations between barriers and career and educational outcomes?

Another potential moderator is individuals’ developmental levels or year in school. There is evidence, for example, that educational and career aspirations developed by 8th grade are quite stable and are among the strongest predictors of 12th grade and young adult educational and career aspirations (e.g., Lee & Rojewski, 2009; Mello, 2008). There is also evidence that students’ experience in school before the eighth grade have a significant influence on eighth-grade aspirations (Rojewski & Yang, 1997). Thus, might the influence of barriers on students educational and career outcomes be more pronounced for younger (i.e., elementary and junior high) than older (high school and college) students? Thus, one goal of this study was meta-analytically to estimate the magnitudes of the negative relationships between barriers and eight different career and educational outcomes and to test whether gender, race/ethnicity, and year in school moderate these relationships.

Although the origin of interest in the support construct in the vocational psychology literature is harder to pinpoint, it has had a long history in the developmental psychology and mental health literatures and much is known about its relationship to developmental and mental health outcomes (Chu, Saucier, & Hafner, 2010; Cohen & Wills, 1985; Rueger, Malecki, Pyun, Aycock, & Cole, 2016). Recent meta-analyses have suggested that support is positively related to well-being (Chu et al., 2010) and negatively related to depression (Rueger et al., 2016), with the magnitude of these relationships being quite similar (.20 to .26 in absolute value). These meta-analyses have also reported that the relationships between supports and well-being and mental health outcomes vary by sources of support (e.g., teacher vs. family vs. peers), gender, and developmental level.

There appears to be consensus among vocational psychology scholars (e.g., Brown & Lent, 2016; Lent, Brown, & Hackett, 2000) that support is related positively to educational and vocational outcomes, but (as is the case with barriers) the magnitude of this positive relationship is unknown. Also, unknown is whether these positive relationships might be moderated (as they are in the mental health literature) on the bases of gender and developmental level (year in school). The support literature has revealed, for example, that girls and women tend to report consistently higher average levels of support than do boys and men (e.g., Colarossi & Eccles, 2003; Malecki & Demaray, 2003). However, it is not clear whether these gender differences translate into gender differences in the associations between support and educational and vocational outcomes. And given the importance of early educational performance and educational and career aspirations to later aspirations and attainment as discussed earlier, might the positive influences of support (like the negative influences of barriers) be more pronounced among younger children (elementary and junior high aged) than adolescents and young adults (high school and college aged)? Research on racial/ethnic differences in support has revealed that different racial/ethnic groups (e.g., African American and non-Hispanic Whites) have different configurations of informal support networks (e.g., congregation-based sources of support in African American vs. White communities), but the degree to which these sources of support relate to career and educational outcomes has been infrequently studied (Taylor, Chatters, Woodward, & Brown, 2013). Thus, the second goal of this study was to estimate the magnitude of the relationships between support and eight career and educational outcomes and explore the potential moderating influences of gender, race/ethnicity, and year in school.

We also had two other goals. For one, we wanted to estimate the comparative magnitudes of the relationship of supports and barriers to the same set of career and educational outcomes. We expected, given recent literature on the importance of attending to support versus barriers in career interventions (see Brown & Ryan Krane, 2000; Liu, Huang, & Wang, 2014), that support would account collectively for more variance in these outcomes than would barriers.

Our other goal was to explore, when possible, the relationships among specific types of supports and barriers to career and educational outcomes. The literatures on supports and barriers have clearly revealed that both constructs are multidimensional. In the case of barriers, Swanson and Tokar (1991a, 1991b) identified several categories of environmental barriers, including (a) discouragement from pursuing career goals, (b) sex discrimination, (c) racial discrimination, (d) discrimination due to other factors (e.g., disability, age), (e) job-market constraints, (f) multiple role conflict, and (g) work–family conflict. Research on environmental supports has revealed that supports can be categorized by both type (emotional, instrumental, informational, and appraisal) and source (e.g., family, parent, peer, sibling, and teacher; Malecki & Demaray, 2003). Meta-analytic evidence has also suggested that support from family members and peers is more strongly inversely related to depression than support from teachers and close friends (Rueger et al., 2016).

Given these findings, we also wanted to explore the relationships of types of barriers and supports and sources of supports to career and educational outcomes. However, there were often insufficient data to explore all possible relationships. Thus, we present meta-analytic results on these relationships only when there was sufficient data (k = 5 or more) to yield somewhat stable effect size estimates.

We, therefore, undertook a meta-analysis of the relations between barriers and supports to a variety of different educational and career outcomes. Although recent theories (Duffy, Blustein, Diemer, & Autin, 2016; Lent, Brown, & Hackett, 1994, 2000) have explicitly incorporated the supports and barriers constructs offering specific predictions about their role in career and educational development, our meta-analysis, being the first one to appear in the literature, is largely an exploratory investigation of the relationships between supports and barriers and career and educational outcomes addressing several questions: (a) how much total variance in career and educational outcomes is accounted for by barriers; (b) how much total variance in career and educational outcomes is accounted for by supports; (c) do supports account for more outcome variance than do barriers; and (d) do gender, race/ethnicity, and year in school moderate the relationships between barriers and supports and career and educational outcomes?

Method

Barriers and Supports—Definitions and Classifications

Although early definitions of barriers (e.g., Swanson & Tokar, 1991a, 1991b) included both environmental (e.g., discrimination) and intrapersonal (e.g., lack of confidence) factors that might hinder career and educational development, subsequent writing (e.g., Lent et al., 2000) has suggested that such a nondifferentiated view of barriers may obscure different paths through which intrapersonal and environmental factors may impede or facilitate educational and career development. For example, including both discrimination and low self-efficacy (confidence) as barriers may obscure the paths through which these two variable impede career and educational development (e.g., that the influence of discrimination on career and educational outcomes may be mediated by self-efficacy beliefs). Thus, we adopted Lent, Brown, and Hackett’s (2000) definitions of barriers as “negative contextual influences” that make career and educational progress difficult (p. 39) and used Swanson and Tokar’s (1991a, 1991b) classification of environmental barriers initially to code types of barriers. Our classification system went through several iterations as coding progressed resulting in a final coding scheme that included six potential types of barriers to educational and career progress—(a) discrimination, (b) discouragement from educational and career pursuits, (c) family conflict, (d) teacher–child conflict, (e) negative or inhospitable environment, and (f) perceived barriers other than discrimination and discouragement. The latter (perceived barriers other than discrimination and discouragement) tended to ask about a broad range of potential barriers to different types of educational and career outcomes (e.g., barriers to occupational mobility, college going, and career attainment). Some measures coded into this category included individual items related to discrimination and discouragement, but these were coded in the “other than” category if they did not include separate scales measuring discrimination or discouragement. We also categorized all forms of discrimination together to have a sufficient number of samples to investigate this type of barrier’s relationships to all career and educational outcomes and to engage in planned moderator analyses.

We also used Lent et al.’s definition of environmental supports as “environmental variables that can facilitate” the process of career and educational development (p. 42) to conceptualize environmental supports and initially coded supports by type and source. As with barriers, our classification system went through several iterations as coding progressed and resulted in eight support categories. The first five categories cross sources of support (parent, sibling, peer, teacher, and romantic partner) with two types of support (instrumental and emotional). Instrumental support was defined as outcome-related support (e.g., support for career goals) and combined instrumental, informational, and appraisal supports. Emotional support was defined, consistent with the support literature (see Malecki & Demaray, 2003), as the provision of caring, love, empathy, and trust that was not outcome focused. Other categories included support whose source was not specified, financial support, and positive or hospitable environments. We included positive or hospitable environments as a source of support as well as its opposite (negative or inhospitable environment) as a barrier because prior writing on the null environment (Betz, 1989) has suggested that the absence of hostility or in-hospitability does not mean that a work or school environment is positive and supporting.

We initially used social cognitive career theory (SCCT; Lent et al., 1994; Lent & Brown, 2006, 2008) to identify potential educational and career outcomes to include in the meta-analysis. Initial outcomes suggested by SCCT included academic performance, academic persistence, educational and career goals, educational and career aspirations, educational and career self-efficacy beliefs, educational and career outcome expectations, educational and career interests, and educational and career satisfaction. We identified several other outcomes as our coding progressed. These included career decidedness and school engagement as well as a number of other outcomes that were studied too infrequently (e.g., career search activities, school belonging, vocational identity, and career maturity) to be included in our meta-analysis.

Although there were several outcomes that could be coded as indices of academic performance (e.g., self-reported grade point average [GPA], recorded GPA, and test performance), we chose to use recorded GPA as the sole index of academic performance (hereafter called GPA). Measures of both intended and actual persistence were used to operationalize persistence. Consistent with Lent, Brown, and Hackett (1994) and research on the aspiration-expectation gap (see Rojewski, 2005), we separated expected (termed goals in SCCT) from aspired to outcomes. Thus, measures were coded as aspirations only if they asked about participants’ ideal career or level of education, while they were recorded as goals if they asked about expected educational and career choices as well as choice commitment and certainty.

We coded variety of different indices of self-efficacy and outcome expectations as long as they were career or educationally focused (e.g., measures of generalized self-efficacy beliefs were excluded). Our operationalization of self-efficacy beliefs included education/academic (i.e., confidence in successfully completing academic tasks, courses, or subjects), occupational/vocational (i.e., confidence in succeeding in different occupations), career planning, career decision-making, and career search self-efficacy. Measures coded as outcome expectations included anticipated outcomes for educational attainment, occupational choice, college major choice, and career decision-making. Interests included both occupational and educational interests, while satisfaction was operationalized by measures of school and college major satisfaction. School engagement was operationalized by measures that focused on the degree to which participants engaged in school work and school activities. We also coded school belonging in a separate category to reflect participants’ attachment to their schools, but this category yielded too few studies to be included in the meta-analysis.

Literature Search

We employed three strategies to identify all available published studies that could provide relevant effect size estimates. First, we searched PsycINFO and Web of Knowledge databases using the following support and barriers terms initially to direct the search: social support, perceived support, parent support, family support, peer support, friend support, teacher support, school support, workplace support, environmental barriers, perceived barriers, discrimination, gender discrimination, ethnic discrimination, age discrimination, academic discouragement, family conflict, and teacher conflict. Each of these terms was paired, using the Boolean search “and” term, with the following outcomes: academic performance, GPA, academic persistence, tenure, intentions to leave, academic expectations, academic aspirations, career expectations, career aspirations, self-efficacy, outcome expectations, academic interests, career interests, and interests. Later searches added school engagement, school belonging, vocational identity, career decidedness, and career indecision.

Second, we manually searched recent (2 years) tables of contents and online first entries of 68 journals that yielded at least one potential study from the database. Third, we consulted the reference sections of all articles generated via the first two procedures to identify other relevant studies. On the basis of abstract descriptions, 394 articles were identified as potentially relevant to the current meta-analysis.

Inclusion and Exclusion Criteria

To be included in the meta-analysis, primary studies had to meet the following criteria. First, we only included quantitative studies of the relationship between some form of support or barrier and a career or educational outcome (i.e., literature reviews and qualitative studies were eliminated). Second, we then retained only studies that contained a correlation or other statistic (e.g., t, univariate F, M, and SD) that could be converted into bivariate correlations. For example, if a two-group, between-group study reported a t or univariate F, these test statistics were converted to bivariate correlations via standard conversion formulae (Cohen, 1992). If a between group study failed to provide an exact test statistic value and degrees of freedom (or an F with more than one degree of freedom in the numerator), but included group means and standard deviations, means and standard deviations were used to calculate a Cohen’s d that was subsequently converted into a bivariate correlation. Studies were eliminated if they failed to provide a bivariate correlation or sufficient information to calculate one (e.g., structural equation modeling studies that failed to provide an input correlation or covariance matrix).

After eliminating studies that were not quantitative, did not included data on relevant dependent and independent variables, or did not provide useable data (k = 145), the final sample for this meta-analysis included 249 studies that yielded 276 independent samples (e.g., some articles contained more than one study or contained independent samples within a single study). References to the 249 included studies are available from the first author upon request. As can be seen in Table 1, the 276 independent samples yielded a total of 104,440 participants (39% men, 46% women, and 16% missing data on gender). The mean age of the sample was 17.47 (SD = 5.36, range = 5.47–38.10) and most were in high school or college (elementary = 22.18%, junior high = 16.62%, high school = 30.94%, college = 23.56%, noncollege adults = 2.94%, and other or not reported = 3.68%). The racial/ethnic composition of the sample was 31.50% African American, 4.03% Asian American, 31.50% Caucasian, 21.35% Latino/a, 1.93% mixed race, 1.79% Native American, 6.50% other, and 11.77% not reported. Forty (10.64%) samples totaling 21,756 participants were from outside of the United States.

Table 1.

Study (j = 248) and Sample (k = 267) Characteristics.

Study Characteristics
Variable	n/k	Percentage	Median (Mean)	SD	Range
Year of publication			2006 (2004)	6.89	1974–2014
Journals (J)	67
Discipline of journal (K)^a
Vocational psych.	62	25
Child psych/develop	48	19
Ed. psych./education	28	12
Couns. psych./couns.	22	9
Minority psych./ed.	22	9
School psych.	17	7
Community psych.	15	6
College student dev.	9	4
Family	6	2
General psych.	6	2
Personality/social psych.	6	2
Psychiatry	3	1
Other^b	4	2
Sample characteristics
Variable	n	Percentage	Mean	SD	Range
Age			17.15	5.36	5.47–38.10
Gender
Male	40,071	38.37
Female	47,843	45.81
Missing	16,526	15.82
Race/ethnicity
African American	18,143	17.37
Asian American	3,460	3.31
Caucasian	27,042	25.89
Latino(a)	18,329	17.55
Native American	1,654	1.58
Mixed race	1,534	1.47
Other	5,579	5.34
Not reported	10,101	9.67
International^c	18,598	17.81
Year in school
Elementary	23,162	22.18
Junior high	17,355	16.62
High school	32,315	30.94
Technical school	102	0.10
College	24,601	23.56
Adult, noncollege	3,073	2.94
Not reported	3,832	3.67

^aNumbers and percentages are for number and percentage of samples published in each discipline. ^bJournals coded as “other” were Southern Communication Journal, Journal of School Health, International Journal of Stress Management, and Journal of Homosexuality. ^cInternational samples included Western European (k = 17, n = 12,622), Asian/Pacific Islander (k = 3; n = 2,257), Caribbean (k = 2; n = 649), African (k = 2; n = 853), and multinational (k = 16; n = 5,375).

Coding Procedures

A code book was developed and revised several times to capture (a) study (year of publication, journal, country of origin, whether theory was tested and the title of the theory) and participant characteristics; (b) barriers, supports, and career and educational outcome measures; (c) a priori planned moderator variables (gender, race/ethnicity, age/year in school); and (d) statistics for effect size. Coding was completed by teams of two to three graduate student research assistants and presented in a group meeting attended by the first author and all coders. Discrepancies were resolved by team consensus.

Coded participant characteristics included (a) total sample size, (b) gender (number and percentage of men and women), race/ethnicity (number and percentage of African American/Black, American Indian/Alaska Natives, Asian American/Pacific Islander, Hispanic American/Latino(a), White/Caucasian, mixed race, Other), and (c) year in school (number elementary, junior high, high school, community/junior college, technical/vocational school, 4-year college, graduate/professional school, adult not in school, other). Although most studies did not provide estimates of the socioeconomic status (SES) of participants, we coded indicators (percentage of free or reduced lunch) of SES when they were available. Sample means, standard deviations, and score ranges were recorded for each measure as well as reported reliability estimates.

Meta-Analytic Procedures

We used Hunter and Schmidt’s random-effects meta-analysis method to synthesize correlation coefficients across independent samples (Hunter & Schmidt, 2004; Schmidt & Le, 2004). Correlations were corrected for measurement error in both the independent (supports and barriers) and dependent (career and educational outcomes) variables using internal consistency estimates reported in the primary studies. Reliability for GPA was estimated to be 1.00. For studies that did not report internal consistency estimates, several methods were used to impute them. First, we searched the literature for reliability estimates on the scale obtained from samples similar to the primary study (e.g., we searched for reliability estimates reported for high school students when the primary study sample was composed of high school students). Second, when the first method failed to yield sample-appropriate reliability estimates, we contacted the corresponding author to obtain reliability estimates from the primary study. Third, when the first and second methods failed to yield sample-appropriate reliability estimates, we estimated reliability via sample size weighted meta-analyses using all reliability estimates available for a particular scale score. The latter method was employed to estimate reliability in 14 (5%) of the 276 samples. We then used the corrected correlations to calculate the overall sample-size weighted true score correlation (ρ).

If a study yielded several samples addressing the meta-analytic or moderator question, the arithmetic mean was used to provide a single effect size estimate for that study. For example, if a study reported correlations separately by race, correlations were averaged in the overall meta-analyses. However, in the analyses of race as a moderator, the samples were considered as independent samples and their correlations were, therefore, disaggregated (i.e., not averaged).

We examined the standard error of the mean true score correlation by computing 95% confidence intervals (CIs) to estimate whether the estimated true score correlation differed from 0. We took several steps to gauge the likely presence of moderators and to test moderator effects. First, we computed 80% credibility intervals (CRs) for the mean true score correlation and percentage of variance due to artifacts for each overall correlation. Wide CRs and a small percentage of artifact variance suggest the presence of moderators of the effect size estimate.

For moderator analyses, we computed true score correlations and 95% CIs in the different moderator categories (e.g., male and female). Overlap between the CIs suggests that the difference between the true score correlations is artifactual and that there is no meaningful moderating effect. Nonoverlapping CIs, however, suggest that the difference between the true score correlations is not artifactual and that there may be a meaningful moderating effect.

For the moderator analyses involving sex and race, there were insufficient samples including exclusively male or White participants to conduct the analyses. We, therefore, assigned samples to gender and race/ethnicity categories on the basis of a 75% rule—if 75% or more of a sample was of a specific gender or race/ethnicity, it was assigned to its corresponding category (e.g., samples coded as male had at least 75% male participants).

Results

Table 2 summarizes the meta-analytic results for the relationships between supports (upper panel) and barriers (lower panel) and nine educational and career outcomes. We include only outcomes for which five or more samples yielded effect size estimates. Eight of the nine outcomes appear in both panels (GPA, persistence, goals, self-efficacy, outcome expectations, interests, school engagement, and educational and career aspirations), with school satisfaction appearing only for supports and indecision appearing only for barriers.

Table 2.

Results of Meta-Analysis of Supports and Barriers With Career and Educational Outcomes.

Variable	k	N	r	ρ	SDρ	95% CI [LL, UL]	80% CR [LL, UL]	% Artifact
Supports
GPA	42	14,995	.11	.13	.10	[.09, .16]	[.00, .26]	25.40
Persistence	28	16,780	.20	.23	.10	[.10, .37]	[−.13, .77]	5.74
Goals	35	10,712	.25	.30	.16	[.25, .36]	[.10, .51]	13.94
Self-efficacy	88	36,004	.27	.33	.13	[.30, .36]	[.16, .50]	14.32
Outcome expectations	41	14,664	.35	.42	.12	[.37, .46]	[.26, .57]	16.78
Interests	26	10,526	.26	.31	.12	[.26, .36]	[.16, .46]	18.02
Satisfaction	13	5,892	.38	.47	.16	[.38, .56]	[.27, .68]	8.82
Engagement	35	17,658	.27	.33	.18	[.27, .39]	[.10, .56]	7.56
Aspirations	21	6,321	.18	.23	.20	[.13, .32]	[−.03, .48]	11.49
Barriers
GPA	30	8,956	−.14	−.15	.11	[−.20, −.11]	[−.30, −.01]	23.11
Persistence	18	7,240	−.21	−.26	.25	[−.38, −.14]	[−.58, 06]	5.16
Goals	27	8,954	−.07	−.09	.14	[−.15, −.03]	[−.26, .08]	19.54
Self-efficacy	63	22,703	−.12	−.13	.21	[−.19, −.08]	[−.40, .14]	7.65
Outcome expectations	32	9,881	−.15	−.18	.14	[−.23, −.13]	[−.36, .00]	19.05
Interests	16	6,574	−.08	−.10	.11	[−.16, −.04]	[−.24, .04]	22.24
Indecision	15	3,693	.21	.24	.18	[.15, .34]	[.02, .47]	14.58
Engagement	16	6,700	−.15	−.19	.20	[−.29, −.09]	[−.44, −.06]	8.05
Aspirations	13	3,894	−.08	−.10	.14	[−.19, −.02]	[−.28, .07]	21.43

Note. k = number of samples; N = total sample size; r = sample size weighted correlation, not corrected for measurement error, ρ = sample size weighted correlation corrected for measurement error, SDρ = standard deviation of ρ; CI = 95% confidence interval; CRI = 80% credibility interval; % artifact = percentage of variance attributable to statistical artifact; LL = lower limit; UL = upper limit.

As can be seen in Table 2, the true score correlations between supports and barriers and the nine career and educational outcome were all significant (95% CIs did not include 0). True score correlations ranged from .13 (GPA) to .47 (satisfaction) for supports (M =.31) and from −.09 (goals) to −.26 (persistence) for barriers (M = −.16). Overall supports appeared to account for more variance (10%) in these career and educational outcome than did barriers (3%). Except for the relationships with GPA and persistence, the true score correlations involving support were often substantially larger than the true score correlations involving barriers. (e.g., .30 vs. −.09 for goals).

Despite the fact that supports and barrier both related significantly to all outcomes, the 80% CRs and percentages of variance due to artifact displayed in Table 2 suggest substantial heterogeneity of true score correlations within each cell. In three cases (support persistence, support aspirations, and barriers aspirations), the 80% CRs spanned 0, suggesting that the populations estimates may not all be positive (in the case of the support-persistence and aspirations relationships) or negative (in the case of barriers–aspiration relationships). In the other cases, the CRs were wide but did not span 0, suggesting that the population estimates are somewhat heterogeneous but positive (for supports) and negative (for barriers).

The substantial heterogeneity displayed within each cell, suggested that an exploration of potential moderators may yield useful information. Tables 3 and 4 summarize the results of our a priori hypothesized moderators. Again, we explored moderators only when there were at least five samples available for the analyses. As a result, gender as a moderator was tested only for the relationship between supports (Table 3) and barriers (Table 4) with self-efficacy beliefs and outcome expectations. The results of these analyses suggested that gender moderated the support-self-efficacy and outcome expectations relationships, with the true score correlations in both cases being significantly higher (p = .05) for men than for women. Gender, however, did not appear to moderate the relationships between barriers and these two outcomes.

Table 3.

Results of Moderator Analyses for Supports.

Variable/Demographic	k	N	ρ	95% CI [LL, UL]	% Artifact
Gender
Self-efficacy
Male	8	2,992	.41	[.34, .48]	29.61
Female	15	3,152	.21	[.14, .28]	31.34
Outcome expectations
Male	6	2,650	.59	[.50, .69]	100
Female	5	1,473	.24	[.20, .27]	10.88
Race/ethnicity
Self-efficacy
White	23	7,161	.33	[.28, .39]	20.98
African American	12	2,538	.24	[.19, .30]	67.63
Latino (a)	9	2,360	.24	[.13, .35]	18.42
Outcome expectations
White	11	3,794	.35	[.29, .40]	35.11
African American	7	1,483	.37	[.30, .43]	80.90
Latino (a)	5	1,585	.34	[.25, .44]	33.47
Engagement
White	9	6,310	.42	[.36, .49]	16.48
Latino(a)	7	2,108	.23	[.15, .31]	40.25
GPA
White	9	3,046	.22	[.17, .26]	66.30
African American	8	3,029	.08	[.02, .15]	39.45
Latino (a)	5	708	.27	[.16, .38]	63.16
Year in school
GPA
Elem	9	2,115	.18	[.12, .23]	78.26
Junior high	6	4,193	.13	[.06, .20]	23.91
High school	16	4,620	.14	[.06, .21]	18.57
College	9	3,789	.08	[.02, .14]	35.23
Self-efficacy
Elem	6	3,428	.33	[.21, .46]	10.68
Junior high	14	4,018	.37	[.25, .49]	8.79
High school	22	6,301	.30	[.24, .36]	21.45
College	33	12,603	.37	[.34, .41]	23.25
Outcome expectation
Junior high	7	1,700	.36	[.28, .44]	48.43
High school	15	4,561	.36	[.33, .39]	98.89
College	15	7,584	.46	[.37, .53]	7.93
Engagement
Elem	11	4,312	.46	[.37, .45]	9.89
High school	11	4,249	.27	[.21, .32]	46.63

Note. k = number of samples; N = total sample size; r = sample size weighted correlation, not corrected for measurement error; ρ = sample size weighted correlation corrected for measurement error; SDρ = standard deviation of ρ; CI = 95% confidence interval; % artifact = percentage of variance attributable to statistical artifact; LL = lower limit; UL = upper limit.

Table 4.

Results of Moderator Analyses for Barriers.

Variable/Demographic	k	N	ρ	95% CI [LL, UL]	% Artifact
Gender
Self-efficacy
Male	6	2,767	−.18	[−.26, −.10]	26.94
Female	16	4,058	−.09	[−.17, −.02]	23.91
Outcome expectations
Male	6	2,783	−.19	[−.28, −.10]	38.17
Female	6	1,453	−.13	[−.24, −.02]	21.43
Race/ethnicity
Self-efficacy
White	13	3,811	−.24	[−.32, −.16]	20.87
African American	9	1,627	−.11	[−.27, −.02]	21.83
Outcome expectations
White	11	3,794	−.17	[−.31, −.03]	12.53
African American	7	1,438	−.26	[−.42, −.10]	17.20
GPA
White	5	1,279	−.15	[−.29, −.01]	19.17
African American	8	2,138	−.12	[−.20, −.05]	37.01
Year in school
GPA
Elem	5	965	−.20	[−.30, −.09]	40.22
High school	10	4,033	−.12	[−.18, −.06]	28.94
College	10	2,085	−.10	[−.19, −.00]	25.23
Self-efficacy
Junior high	7	1,818	−.31	[−.39, −.23]	42.2
High school	17	4,323	−.21	[−.31, −.11]	11.25
College	26	9,189	−.05	[−.15, .04]	5.64
Outcome expectations
High school	15	4,175	−.26	[−.33, −.20]	27.12
College	11	4,830	−.16	[−.24, −.06]	12.77
Engagement
High school	6	2,637	−.13	[−.23, −.02]	20.94

The moderating effects of race/ethnicity on the relationships between support and GPA, self-efficacy, outcome expectations, and school engagement are also displayed in Table 3. Results revealed race/ethnicity to be a significant (p = .05) moderator of the relationships of supports and school engagement and GPA, with Latino(a) students obtaining significantly lower true score correlations with school engagement than Caucasians. Although the number of samples employing African American students did not meet our k of at least five criterion in the school engagement analyses (k = 3), the true score correlation (ρ = .12) obtained from these three samples was also significantly lower (CI [.00, .24]) than that obtained from White students. In the case of GPA, African American samples obtained significantly lower true score correlations than did the White and Latino(a) samples. Race/ethnicity, however, did not moderate the true score correlations between barriers and GPA, self-efficacy beliefs, and outcome expectations (the moderating effect of race/ethnicity on barriers–school engagement relationships could not be tested).

Year in school was tested as a potential moderator of the relationships between both supports (Table 3) and barriers (Table 4) and GPA, self-efficacy beliefs, outcome expectations, and school engagement. Results revealed that year in school moderated the relationships between supports and school engagement—the true score correlation was significantly (p = .05) more positive for elementary school students than for high school students. Although there were only four samples available for an analysis of year in school as a moderator of barrier-school engagement relationships a similar moderator effect was present—the true score correlation was significantly more negative for elementary school students than for high school students (95% CI [−.50, −.24] vs. [−.23, −.02] for high school students).

There were insufficient samples to cross fully types (instrumental and emotional) and sources (e.g., parents, teachers) of support. Thus, we collapsed over types of support and present in Table 5 the results of analyses for sources of support, again only including effect sizes obtained from five or more samples. In general, it appears that support from parents, peers, and teachers related positively and significantly to GPA, persistence, aspirations, goals, self-efficacy, outcome expectations, interests, and engagement, with effect sizes ranging from .14 (peers with GPA and self-efficacy) to .35 (teachers with goals). There was only one significant difference found for sources of support; namely, that the correlation between peer support and self-efficacy (ρ = .14) was lower than the relationships of parent and teacher support with self-efficacy (ρ = .32 and .34, respectively). Hospitable environments also related significantly to school engagement (ρ = .37). Finally, it appears that both parent and teacher support are significantly more strongly related to school engagement (ρ = .26 and .37, respectively) than with GPA (ρ = .19 and .14, respectively).

Table 5.

Results for Analyses of Specific Support.

Support Type	K	N	r	ρ	SDρ	95%CI [LL, UL]	% Artifact
GPA
Parent	22	5,906	.16	.19	.06	[.16, .22]	78.56
Peer	14	4,256	.12	.14	.08	[.10, .19]	54.42
Teacher	22	7,897	.12	.15	.14	[.08, .22]	14.30
Persistence
Parent	16	3,586	.27	.32	.10	[.23, .39]	33.54
Peer	15	4,045	.23	.30	.21	[.17, .43]	8.11
Teacher	12	3,248	.18	.23	.14	[.13, .33]	18.46
Aspirations
Parent	20	7,079	.18	.23	.18	[.14, .33]	8.56
Peer	5	643	.18	.23	.04	[.19, .28]	100
Teacher	6	1,669	.19	.25	.07	[.18, .32]	75.31
Goals
Parent	20	4,386	.15	.18	.13	[.14, .25]	28.75
Peer	9	1,721	.17	.20	.10	[.12, .27]	49.63
Teacher	10	2,046	.30	.35	.20	[.21, .49]	10.99
Self-efficacy
Parent	58	23,850	.26	.32	.14	[.27, .36]	10.36
Peer	20	12,733	.10	.14	.13	[.07, .21]	10.26
Teacher	25	7,572	.28	.34	.11	[.29, .39]	25.62
Pos. Env.	7	8,365	.23	.31	.03	[.29, .34]	100
Outcome expectations
Parent	18	5,330	.27	.33	.13	[.26, .40]	18.36
Peer	9	2,720	.21	.25	.07	[.18, .32]	47.32
Teacher	10	3,105	.28	.34	.09	[.29, .40]	45.45
Interests
Parent	10	3,006	.17	.22	.12	[.13, .30]	25.78
Engagement
Parent	15	5,466	.21	.26	.12	[.19, .34]	18.55
Peer	10	2,402	.10	.17	.16	[.09, .32]	16.99
Teacher	16	6,421	.30	.37	.21	[.25, .49]	4.94
Pos. Env.	8	7,527	.30	.37	.08	[.29, .45]	10.26

Note. Pos. Env. = positive or hospitable environment; k = number of samples; N = total sample size; r = sample size weighted correlation, not corrected for measurement error; ρ = sample size weighted correlation corrected for measurement error; SDρ = standard deviation of ρ; CI = 95% confidence interval; % artifact = percentage of variance attributable to statistical artifact; LL = lower limit; UL = upper limit.

Analyses of the relations between specific types of barriers and career and educational outcomes are displayed in Table 6, when five or more studies were available for the analyses. Several results stand out. First, perceptions of discrimination are related significantly (p = .05) to academic persistence (ρ = −.31) but not to GPA (ρ = −.08), goals (−.02), self-efficacy (−.06), outcome expectations (ρ = −.11), or interests (ρ = −.09). On the other hand, barriers other than discrimination are significantly negatively related to all outcomes (ρ’s range from −.30 for persistence to −.10 for self-efficacy). The relationships of inhospitable environments and family conflict with GPA were significant (p = .05), with effect sizes of −.19 and −.23, respectively. Inhospitable environments also correlated significantly with school engagement, but the positive relationship between supportive environments and engagement (ρ = .37) was significantly higher (p = .05) than the negative relationship between inhospitable environments and engagement (ρ = −.10).

Table 6.

Results for Analyses of Specific Barriers.

Barrier Type	k	N	R	ρ	SD _ρ	95%CI [LL, UL]	% Artifact
GPA
Discrimination	11	2,910	−.07	−.08	.12	[−.16, .00]	25.40
Inhospitable env.	12	4,558	−.16	−.19	.11	[−.26, −.11]	20.49
Family conflict	8	1,337	−.20	−.23	.11	[−.31, −.15]	54.34
Persistence
Discrimination	7	2,820	−.24	−.31	.25	[−.55, −.07]	3.56
Perceived barriers	6	2,454	−.25	−.30	.17	[−.46, −.13]	7.28
Inhospitable env.	5	1,966	−.07	−.10	.19	[−.29, .10]	7.44
Aspirations
Perceived barriers	8	1,986	−.10	−.13	.11	[−.22, −.04]	37.12
Goals
Discrimination	7	1,650	.02	.02	.09	[−.06, .11]	55.49
Perceived barriers	18	6,854	−.09	−.12	.13	[−.19, −.04]	17.18
Self-efficacy
Discrimination	12	3,181	−.05	−.06	.12	[−.14, .02]	25.36
Perceived barriers	41	13,817	−.10	−.10	.21	[−.18, −.03]	6.47
Inhospitable env.	9	4,268	−.20	−.27	.08	[−.35, −.20]	25.79
Outcome expectations
Discrimination	11	2,775	−.09	−.11	.14	[−.21, .00]	19.56
Perceived barriers	20	6,226	−.19	−.22	.12	[−.28, −.16]	21.55
Interests
Discrimination	5	1,585	−.02	.09	−.02	[−.12, .07]	40.13
Perceived barriers	9	4,001	−.13	−.15	.10	[−.23, −.08]	24.09
Engagement
Inhospitable env.	8	3,225	−.08	−.10	.12	[−.20, .00]	18.17
Indecision
Perceived barriers	10	2,776	.22	.26	.16	[.14, .38]	12.68

Note. Inhospitable env. =inhospitable environment; k = number of samples; N = total sample size; r = sample size weighted correlation, not corrected for measurement error, ρ = sample size weighted correlation corrected for measurement error, SDρ = standard deviation of ρ; CI = 95% confidence interval; % artifact = percentage of variance attributable to statistical artifact; LL = lower limit; UL = upper limit.

Discussion

The present study explored meta-analytically the relations between environmental supports and barriers and nine different educational and career outcomes. The results supported expectations that supports would relate positively and barriers negatively to all outcomes. However, the results were equally clear that supports appear to be more facilitative of these outcomes than barriers are hindering. On the average, supports accounted for approximately 10% of outcome variance versus approximately 3% for barriers. In terms of individual outcomes, the true score correlations were significantly larger for supports than barriers in relation to educational and career goal setting, aspirations, self-efficacy beliefs, outcome expectations, interests, and school engagement. No differences were found for school performance (GPA) or academic persistence. In both cases, the relationships of barriers and supports were stronger for persistence than performance, suggesting that positive and negative environmental influences may influence academic persistence rates and intentions more than they influence school performance.

The results of these analyses are consistent with prior findings that support building efforts account for more variance in career-choice and job-finding intervention outcomes than barrier reduction methods (Brown & Ryan Krane, 2000; Liu et al., 2014). They also provide empirical support for findings of prior qualitative studies that have highlighted the importance of supports to the career success of African American (Pearson & Bieschke, 2001; Richie et al., 1997), White (Richie et al., 1997), and Latina (Gomez et al., 2001) women and the career development of Native Americans (Juntunen et al., 2001), despite all samples reporting experiencing considerable barriers to their success and development. The results also suggest that empirical efforts to develop effective support-building interventions may be particularly important means of promoting positive educational and career development across the life span, especially in promoting academic persistence and school engagement, academic and career aspirations and goals, and robust educational and career decision-making self-efficacy beliefs and outcome expectations. The data also suggest that support building interventions to promote greater academic persistence may need to be complemented by concerted efforts to reduce the negative environmental conditions that seem to be equally (but negatively) associated with persistence efforts and intentions.

Our moderator analyses also have important implications, especially for the timing of environmental interventions to promote greater school engagement. Consistent with prior research on the importance of early school performance on later academic aspirations and expectations (e.g., Rojewski & Yang, 1997), we found that supports (especially from teachers and parents and from a supportive school environment) and barriers (especially experiencing a negative or inhospitable school environment) are more strongly related to the school engagement of elementary students than of junior high and high school students. Past research has also consistently demonstrated a positive relationship between school engagement and performance (Perry, Liu, & Pabian, 2010). Together these data and our results suggest that efforts to promote greater school engagement need to start during the elementary school years—trying to engage disengaged high school students may be too late. Our data also suggest that targeting teacher and parent support and creating supportive school environments (while concomitantly reducing environmental in-hospitability) during the elementary school years may be particularly called for. The stronger true score correlations between teacher and parent support with school engagement than with GPA suggest that influence of teacher and parent support on school performance (GPA) may be mediated by school engagement as found in at least one prior study Perry et al. (2010).

Our other moderator analyses suggested that men’s self-efficacy beliefs and outcome expectations may be more responsive to support than women’s self-efficacy beliefs and outcome expectations, despite the fact that women tend to report more support than do men (e.g., Colarossi & Eccles, 2003). We also found that Latino(a) and African American students’ school engagement and performance may be less responsive to supportive experiences than White students’ engagement and performance. Although the moderating effects of race require further study since they were based on a small number of samples, the implications are troubling given the importance of school engagement and performance to later aspirations and expectations. Thus, further research is needed both to confirm these results and to understand better why the school engagement of African American and Hispanic students seems less responsive to support than the performance and engagement of White students. The applied implications of such research could be substantial.

We were somewhat surprised by the results obtained for discrimination as a barrier—only the true score correlation between persistence and discrimination emerged as statistically significant. Discrimination accounted for approximately 9% of the variance in academic persistence versus 1% or less of the variance for the other career and educational outcomes. Nonetheless, the results are clear that the experience or perception of discrimination in the school environment is both a statistically and practically significant impediment to academic persistence and supports continued efforts to reduce discrimination and promote equality in the school environment. However, it is equally clear that future research needs to attend to other types of educational- and career-limiting barriers to begin to disentangle how these different types of nondiscriminatory variables may affect career and educational development, both alone and in interaction with experiences of discrimination. Further research also needs to disentangle discrimination by studying whether different forms of discrimination (e.g., gender, race, and age) may have differential relations to different career and educational outcomes (i.e., we collapsed all forms of discrimination together to facilitate our analyses).

There are several other directions for future research suggested primarily by the gaps in the literature revealed by the present meta-analysis. For one, future research is needed on gender and race/ethnicity as potential moderator variables. For example, we could only test for gender-moderating effects on the relations of supports and barriers to self-efficacy beliefs and outcome expectations. We found a significant moderator effect for the relationship of supports to both outcomes (true score correlations were higher for men than for women). Future research should address gender as a potential moderator with other career and educational outcomes. Similarly, we were only able to test for the moderating effects of race/ethnicity on a limited number of career and educational outcomes (i.e., GPA, self-efficacy, outcome expectations, and school engagement). Race/ethnicity emerged as a potentially important moderator in these analyses and deserves future study as a moderator of other support/barrier-outcome relationships. The moderator tests involving race and ethnicity also revealed that Asian Americans and Native Americans have been understudied in the supports and barriers literature.

It would be particularly useful in the study of race/ethnicity as a moderator to explore more specifically the relations of different types and sources of supports to educational and career outcomes. For example, there is literature to suggest that there are race/ethnicity differences in sources of supports. Taylor, Chatters, Woodward, and Brown (2013), using a large national sample, found no differences in family support received by African Americans and non-Hispanic Whites but several important differences in friend, fictive kin, and congregation support. Non-Hispanic Whites reported more friend support than did African Americans, but African Americans reported more fictive kin and congregation support than did Whites. The former (fictive kin) was defined as individuals who are unrelated by either blood or marriage but regard one another in kinship terms, while the latter (congregation) involved church congregation members. It is obvious from the present meta-analyses that the latter two sources of support have not yet found their way into the vocational psychology literature but need to be included along with teacher, friend, family, and other sources of support in future research on the relation of supports to career and educational outcomes, especially with race/ethnicity as a moderator.

The present meta-analysis is not without limitations primarily due to the fact that we had to eliminate a number of useful published research because of incomplete reporting practices in primary studies. The result was that several analyses that we planned could not be conducted and many that we did conduct were based on small numbers of samples. The degree to which these “missing” studies might have influenced our results is unknown. Some older studies had to be excluded because of substantial data reporting limitations (e.g., failure to report means and standard deviations or exact values of test statistics for nonsignificant results). However, the majority of more recent studies that had to be eliminated were path analytic and structural equation modeling studies that failed to provide input correlation or covariance matrices. We, therefore, encourage researchers not only to continue to report relevant descriptive statistics and exact test statistic values but also to follow recommended practice in reporting the results obtained from structural equation modeling and other multivariate statistical methodologies (McDonald & Ho, 2002; Weston & Gore, 2006). Input correlation or covariance matrices need always to be reproduced as a table in the article. If such a table is too large to be included in the body of the article, it could be included as an appendix or as an electronic supplemental source. Finally, better reporting of psychometric data obtained on instruments employed in a study would facilitate future meta-analyses as well as allow researchers to explore construct-level relationships obtained by correlations between fallible measures (see Brown, 2015).

Limitations notwithstanding, the present meta-analysis suggests that the external barriers and supports represent important constructs for continued investigation and important targets for preventive interventions to promoter optimum educational and career development for all. The results of the present meta-analysis provide evidence that barriers and supports do relate, as expected, to career and educational outcomes. Thus, future research should move beyond this basic question and begin addressing more theoretically and practically useful questions such as those outlined in this discussion, including using existing theoretical frameworks (Duffy et al., 2016; Lent & Brown, 2006, 2008, 2013; Lent et al., 1994) to direct such efforts.

Footnotes

Authors’ Note

Requests for the references used in this meta-analysis should also be directed to the first author.

Acknowledgments

We thank Kyle Bugh, Andrea Carr, Phil Cooke, Rebeca Daniels, Gwendolyn Foehringer, Jason Hacker, Michelle Johnson, Ethan Rucker, Tom Sak, and Anne Siena for their assistance with this research and Rebeca Daniels and Katherine Matthews for reading and commenting on an earlier draft of this article.

Declaration of Conflicting Interests

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Meghan Roche, Matthew Abrams, Kristen Lamp, Kyle Telander, and Alexander Tatum were supported with graduate assistantships from The Graduate School and School of Education of Loyola University Chicago.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Betz

N. E.

(1989). Implications of the null environment hypothesis for women’s career development and for counseling psychology. The Counseling Psychologist, 17, 136–144.

Brown

S. D.

(2015). On statistical wizardry, construct proliferation, and other challenges for our science. The Counseling Psychologist, 43, 614–628.

Brown

S. D.

Lent

R. W.

(2016). Vocational psychology: Agency, equity, and well-being. Annual Review of Psychology, 67, 541–565.

Brown

S. D.

Ryan Krane

N. E.

(2000). Four (or five) sessions and a cloud of dust: Old assumptions and new observations about career counseling. In Brown

S. D.

Lent

R. W.

(Eds.), Handbook of counseling psychology (3rd ed., pp. 740–766). New York, NY: Wiley.

Chu

Saucier

D. A.

Hafner

(2010). Meta-analysis of the relationships between social support and well-being in children and adolescents. Journal of Social and Clinical Psychology, 29, 634–645.

Cohen

(1992). Statistical power analysis. Current Directions in Psychological Science, 1, 98–101.

Cohen

Wills

T. A.

(1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98, 310–357.

Colarossi

L. G.

Eccles

J. S.

(2003). Differential effects of support providers on adolescents’ mental health. Social Work Research, 27, 19–30.

Duffy

R. D.

Blustein

D. L.

Diemer

M. A.

Autin

K. L.

(2016). The psychology of working theory. Journal of Counseling Psychology, 63, 127–148.

10.

Farmer

H. S.

(1976). What inhibits achievement and career motivation in women? The Counseling Psychologist, 6, 12–14.

11.

Fitzgerald

L. F.

Crites

J. O.

(1980). Toward a career psychology of women: What do we know? What do we need to know? Journal of Counseling Psychology , 27, 47–62.

12.

Gomez

M. J.

Fassinger

R. E.

Posser

Cooke

Mejia

Luna

(2001). Voces Abriendo caminos (voices forging paths): A qualitative study of the career development of notable Latinas. Journal of Counseling Psychology, 48, 286–300.

13.

Hunter

J. E.

Schmidt

F. L.

(2004). Methods of meta-analysis: Correcting error and bias in research findings. Thousand Oaks, CA: Sage.

14.

Juntunen

C. L.

Barraclough

D. J.

Bronek

C. L.

Seibel

G. A.

Winrow

S. A.

Morin

P. M.

(2001). American Indian perspectives on the career journey. Journal of Counseling Psychology, 48, 274–285.

15.

Lee

I. H.

Rojewski

J. W.

(2009). Development of occupational aspirations: A piecewise latent growth model of selected influences. Journal of Vocational Behavior, 75, 82–90.

16.

Lent

R. W.

Brown

S. D.

(2006). Integrating person and situation perspectives on work satisfaction: A social-cognitive view. Journal of Vocational Behavior, 69, 236–247.

17.

Lent

R. W.

Brown

S. D.

(2008). Social cognitive career theory and subjective well-being in the context of work. Journal of Career Assessment, 16, 6–21.

18.

Lent

R. W.

Brown

S. D.

(2013). Social cognitive model of career self-management: Toward a unifying view of adaptive career behavior across the life-span. Journal of Counseling Psychology, 60, 557–568.

19.

Lent

R. W.

Brown

S. D.

Hackett

(1994). Toward a unifying social cognitive theory of career and academic interest, choice, and performance [Monograph]. Journal of Vocational Behavior, 45, 79–122.

20.

Lent

R. W.

Brown

S. D.

Hackett

(2000). Contextual supports and barriers to career choice: A social cognitive analysis. Journal of Counseling Psychology, 47, 36–49.

21.

Liu

Huang

J. L.

Wang

(2014). Effectiveness of job search interventions: A meta-analytic review. Psychological Bulletin, 140, 1009–1041.

22.

Luzzo

D. A.

(1993). Ethnic differences in college students’ perceptions of barriers to career development. Journal of Multicultural Counseling and Development, 21, 227–236.

23.

Luzzo

D. A.

(1995). Gender differences in college students’ career maturity and perceived barriers in career development. Journal of Counseling and Development, 73, 319–322.

24.

Malecki

C. K.

Demaray

M. K.

(2003). What type of support do they need? Investigating student adjustment as related to emotional, informational, appraisal, and instrumental support. School Psychology Quarterly, 18, 231–252.

25.

McDonald

R. P.

M. R.

(2002). Principles and practice in reporting structural equation analyses. Psychological Methods, 7, 64–82.

26.

McWhirter

E. H.

(1997). Perceived barriers to education and career: Ethnic and gender differences. Journal of Vocational Behavior, 50, 124–140.

27.

McWhirter

E. H.

Rasheed

Crothers

(2000). The effects of high school career education on social-cognitive variables. Journal of Counseling Psychology, 47, 300–341.

28.

Mello

Z. R.

(2008). Gender variation in developmental trajectories of educational and occupational expectations and attainment from adolescence to adulthood. Developmental Psychology, 44, 1069–1080.

29.

O’Leary

V. E.

(1974). Some attitudinal barriers to occupational aspirations in women. Psychological Bulletin, 81, 809–826.

30.

Pearson

S. M.

Bieschke

K. J.

(2001). Succeeding against the odds: An examination of familial influences on the career development of professional African American women. Journal of Counseling Psychology, 48, 301–309.

31.

Perry

J. C.

Liu

Pabian

(2010). School engagement as a mediator of academic performance among urban youth: The role of career preparation, parental career support, and teacher support. The Counseling Psychologist, 38, 269–295.

32.

Richie

B. S.

Fassinger

R. E.

Linn-Geschmay

Johnson

Prosser

Robinson

(1997). Persistence, connection, and passion: A qualitative study of the career development of highly achieving African American-Black and white women. Journal of Counseling Psychology, 44, 133–148.

33.

Rojewski

J. W.

(2005). Occupational aspirations: Constructs, meanings, and application. In Brown

S. D.

Lent

R. W.

(Eds.), Career development and counseling: Putting theory and research to work (pp. 131–154). Hoboken, NJ: Wiley.

34.

Rojewski

J. W.

Yang

(1997). Longitudinal analysis of select influences on adolescents’ occupational aspirations. Journal of Vocational Behavior, 51, 375–410.

35.

Rueger

S. Y.

Malecki

C. K.

Pyun

Aycock

Coyle

(2016). A meta-analytic review of the association between perceived social support and depression in childhood and adolescence. Psychological Bulletin, 142, 1017–1067.

36.

Schmidt

F. L.

(2004). Software for Hunter-Schmidt meta-analysis method. Iowa City, IA: University of Iowa, Department of Management and Organizations.

37.

Swanson

J. L.

Tokar

D. M.

(1991a). College students’ perception of barriers to career development. Journal of Vocational Behavior, 38, 92–106.

38.

Swanson

J. L.

Tokar

D. M.

(1991b). Development and initial validation of the career barriers inventory. Journal of Vocational Behavior, 39, 344–361.

39.

Taylor

R. J.

Chatters

L. M.

Woodward

A. T.

Brown

(2013). Racial and ethnic differences in extended family, friendship, fictive kin, and congregational informal support networks. Family Relations, 62, 609–624.

40.

Weston

Gore

P. A.

(2006). A brief guide to structural equation modeling. The Counseling Psychologist, 34, 719–751.