Academic Self-Efficacy as a Predictor

of College Outcomes: Two Incremental

Validity Studies
Paul A. Gore Jr.
ACT

A growing body of literature supports the relationship between students’ self-effi-

cacy beliefs for academic tasks and milestones and their academic performance.
Not surprisingly, some researchers have investigated the role that academic self-
efficacy beliefs play in predicting college success. Two incremental validity studies
were conducted to determine the extent to which academic self-efficacy beliefs
could account for variance in college outcomes beyond that accounted for by
standardized test scores. Results suggest that academic self-efficacy beliefs predict
college outcomes but that this relationship is dependent on when efficacy beliefs
are measured, the types of efficacy beliefs measured, and the nature of the criteria
used.

Keywords: academic self-efficacy, college students, academic achievement,

college success, social cognition

Since the introduction of Bandura’s social learning theory (Bandura, 1977),

the construct of self-efficacy has occupied a central role in psychologists’ attempt
to understand and predict human behavior. Central to this theory is the hypoth-
esis that self-efficacy beliefs help to determine what activities individuals will
pursue, the effort they expend in pursing those activities, and how long they will
persist in the face of obstacles. A review of the literature (e.g., Bandura, 1986)
reveals how widely the construct of self-efficacy has been adopted by different
disciplines within psychology. The discipline of vocational psychology and career
development is no exception. This special issue is a testament to the application
of self-efficacy to career development and a tribute to the individuals responsible
for its introduction (Betz & Hackett, 1981; Hackett & Betz, 1981). Even today,
almost 25 years after these two seminal works, self-efficacy enjoys widespread
attention from vocational and counseling psychologists. For example, more than
11% of all articles published in the Journal of Counseling Psychology, Journal of

Correspondence concerning this article should be addressed to Paul A. Gore Jr., ACT, 500 ACT Drive, Box
168, Iowa City, IA 52243-0168; e-mail: [email protected].

JOURNAL OF CAREER ASSESSMENT, Vol. 14 No. 1, February 2006   92–115

DOI: 10.1177/1069072705281367
© 2006 Sage Publications

92
 Gore / ACADEMIC SELF-EFFICACY AND COLLEGE OUTCOMES    93

Vocational Behavior, and Journal of Career Assessment in the past 5 years include
a reference to self-efficacy in their titles or abstracts.
Given the popularity of this construct, it is not surprising that self-efficacy
beliefs have been explored as possible predictors of students’ academic success
and persistence. For example, several early studies focused on the role of aca-
demic self-efficacy beliefs in predicting the performance and persistence in sci-
ence and engineering fields (Brown, Lent, & Larkin, 1989; Lent, Brown, &
Larkin, 1986, 1987; Siegel, Galassi, & Ware, 1985). Findings from these studies
were generally consistent and supportive of relationships between academic self-
efficacy beliefs and (a) college performance, (b) college persistence, and (c) the
range of perceived career options. Furthermore, each of these studies demon-
strated how self-efficacy beliefs could account for variance in college outcomes
(performance and persistence) beyond that accounted for by more traditional
predictors (e.g., standardized achievement/aptitude measures). Multon, Brown,
and Lent (1991) conducted an early meta-analysis of the relationships between
students’ self-efficacy beliefs for academic tasks and their performance and persis-
tence in school. Their findings suggested that between 11% and 14% of the vari-
ance in academic performance and persistence could be accounted for by an
individual’s academic self-efficacy beliefs. Later studies (e.g., Kahn & Nauta,
2001; Lopez, Lent, Brown, & Gore, 1997; Pajares & Miller, 1995) served to fur-
ther establish the strength of these relationships.
Academic self-efficacy can be defined as individuals’ confidence in their abil-
ity to successfully perform academic tasks at a designated level (Schunk, 1991).
Researchers studying academic self-efficacy have developed instruments that
measure individuals’ confidence in their ability to perform a wide range of tasks.
At the most specific level of measurement, academic self-efficacy items are gener-
ally tied to a specific course or course content. For example, some authors have
operationalized academic self-efficacy as a student’s confidence in his or her abil-
ity to respond correctly to items assessing course content knowledge. Examples of
these measures include the mathematics and verbal self-efficacy scales used by
Zimmerman and Martinz-Pons (1990), the geometry or advanced algebra self-
efficacy scales developed by Lopez and his colleagues (Lopez et al., 1997), or the
various content specific measures developed by Bong (1997). On inspection,
these measures resemble achievement tests. Instead of being prompted for a cor-
rect answer, however, participants are asked to rate how confident they are in their
ability to answer the question correctly.
Another group of measures is also defined by its connection with a specific
content domain or class. In contrast to measures described above, however, stu-
dents completing instruments at this level of specificity are often asked to indicate
their confidence in their ability to attain a specific grade in a class. For example,
Mone and his colleagues (Mone, 1994; Mone, Baker, & Jeffries, 1995) created an
academic self-efficacy instrument for use in an undergraduate management class.
Alternatively, Lent and his colleagues (Brown et al., 1989; Lent et al., 1986, 1987)
developed two different measures of academic self-efficacy for use in studying the
94    JOURNAL OF CAREER ASSESSMENT / February 2006

role of self-efficacy in engineering and science majors. In these studies, Lent et al.
used a measure of educational requirements self-efficacy that assesses students’
confidence in their abilities to successfully complete the educational require-
ments and job duties performed in science and engineering. Additionally, these
authors developed an academic milestone self-efficacy measure that assesses stu-
dents’ confidence in their ability to perform specific accomplishments necessary
for academic success in science and engineering majors. Not surprisingly, this
strategy has been adapted to meet the needs of educators in other academic dis-
ciplines. For example, Harvey and McMurray (1994) developed a nursing aca-
demic self-efficacy instrument based on the curriculum requirements in an
attempt to identify students who might be at risk for attrition. Elias and Loomis
(2002) provided an example of how this strategy of measuring academic self-efficacy
can be adapted to evaluate students’ confidence in their ability to successfully
complete general undergraduate courses.
At the next level of specificity, researchers have developed instruments that
assess students’ self-efficacy beliefs for more generalized academic behaviors. For
example, Pintrich and De Groot (1990) included a self-efficacy scale in their
Motivational Strategies for Learning Questionnaire (MSLQ). This scale includes
items that assess students’ confidence in their ability to master course material,
perform well on course tasks, and receive a high grade. Unlike the instruments
described above, instruments of this type can be used with students in any aca-
demic course (e.g., Bong, 2004). A growing number of measures exist that are
representative of this level of academic self-efficacy (Chemers, Hu, & Garcia,
2001; Juang & Vondracek, 2001; Leach, Queirolo, DeVoe, & Chemers, 2003;
Roeser, Midgley, & Urdan, 1996; Solberg, O’Brien, Villareal, Kennel, & Davis,
1993; Wood & Locke, 1987).
Solberg and his colleagues (Solberg et al., 1993) developed a measure that
assesses students’ confidence in their ability to successfully complete college-
related tasks. The factor structure of this instrument clearly identifies a scale that
operationally defines academic self-efficacy at a more general level (Gore,
Leuwerke, & Turley, in press; Solberg et al., 1993). The course self-efficacy scale
includes items such as “research a term paper” and “write a course paper.”
Interestingly, the College Self-Efficacy Inventory (CSEI) includes another scale
(social self-efficacy) that, we believe, extends the definition of academic self-efficacy
to include pro-academic social behaviors. Items on this factor include “ask a profes-
sor a question outside of class” and “talk with academic advising staff.” Results from
recent research suggest that students who score high on this scale also expect to
participate in campus clubs and organizations, interact with college faculty, and
use campus facilities more often than students with lower scores on this scale
(Gore et al., in press). This broader conceptualization of academic self-efficacy
captures constructs such as social and academic integration that are espoused by
leading theorists in college-student development (Astin, 1999; Tinto, 1993).
Finally, academic self-efficacy may be operationalized as one’s confidence in
one’s ability to successfully perform pro-academic self-regulatory behaviors. Self-
 Gore / ACADEMIC SELF-EFFICACY AND COLLEGE OUTCOMES    95

regulation refers to the degree to which students metacognitively, motivationally,

and behaviorally regulate their learning process (Zimmerman, 1986, 1990). For
example, Zimmerman, Bandura, and Martinez-Pons (1992) found that students’
efficacy beliefs for using self-regulatory behaviors (e.g., finishing homework, orga-
nizing schoolwork, and taking class notes) were related to their efficacy beliefs for
academic achievement and their stated course grade goals.
The most recent measure of academic self-efficacy was developed and vali-
dated by Robbins and his colleagues (Le, Casillas, Robbins, & Langley, 2005;
Robbins, Allen, Casillas, & Peterson, 2005; Robbins et al., 2004). In an effort to
develop a multidimensional instrument for use in predicting college student suc-
cess, these authors conducted a meta-analysis of the relationships between psy-
chosocial and study skills factors and college outcomes. Of the 109 studies in their
meta-analysis, 18 (representing more than 9,000 participants) included a measure
of academic self-efficacy. Results of this study suggest that academic self-efficacy
may account for up to 14% of the variance in college students’ grade point aver-
age (GPA; mean observed correlation r = .38). Furthermore, the authors of this
study observed a significant correlation between academic self-efficacy and col-
lege persistence (mean observed correlation r = .26). Finally, Robbins and his
colleagues concluded that academic self-efficacy beliefs account for variance in
both retention and college GPA beyond that accounted for by more traditional
academic predictors (high school performance and standardized test scores).
As a result of their findings, Le et al. (2005) developed and validated a measure
of student readiness (Student Readiness Inventory [SRI]) that includes an aca-
demic self-efficacy scale (called Academic Self-Confidence [ASC]). They opera-
tionalized their construct as the degree to which a student feels he or she can
perform well in school. In contrast to the CSEI (Solberg et al., 1993), the ASC
scale assesses both students’ perceptions of their specific skills (e.g., “I find it hard
to pick out main ideas in texts”) and their beliefs about their general academic
capabilities (e.g., “I have difficulty keeping up academically with my classmates”).
The present studies were conducted to evaluate the utility of using measures
of academic self-efficacy to predict postsecondary academic success and persis-
tence. The first study describes results from an ongoing longitudinal study of 1st-
year college students’ career and academic development. All students in this
study completed the CSEI, and a small subset of students also completed the
ASC measure. Academic performance (GPA) and institutional persistence (reten-
tion) data were obtained across the first 2 years of college. Data from the second
study come from a large ongoing national study of the SRI. For this study, we
obtained students’ ASC scores and their college academic performance and per-
sistence during the first 2 years of college. The present studies extend existing
literature by using multiple measures of academic self-efficacy to predict multiple
college outcomes. The construct validity of these two measures of academic self-
efficacy would be strengthened by the presence of predictive validity estimates
that are similar across both predictors and criteria. Furthermore, because a small
subset of students completed both instruments (Study 1), we can further evaluate the
96    JOURNAL OF CAREER ASSESSMENT / February 2006

construct validity of the two measures by inspecting the relationships between the
two scales. Finally, the present studies extend previous research by addressing the
incremental validity (Sechrest, 1963) of using academic self-efficacy to predict col-
lege outcomes. There is a large volume of evidence supporting the relationship
between standardized test performance and college success (ACT, 1997; Boldt,
1986; Mathiasen, 1984). To have practical utility in predicting college outcomes,
measures of academic self-efficacy must account for variance in college outcomes
beyond that which can be accounted for by these more traditional predictors.

STUDY 1

Method

Participants and Procedure

Participants for this study were 629 1st-year college students (335 males, 294
females) enrolled in a 3-credit-hour freshman orientation/transition class at a
large public midwestern university. The mean age of participants was 18.1 years
(ages ranged from 17 to 22). Students self-identified their ethnic/racial back-
grounds as follows: White (78%), African American (13%), Latino (3%), and
Asian American (2%). Students’ average ACT composite score was 20.7 (SD =
3.5, scores ranged from 11 to 32). Students self-reported their high school GPAs
as follows: <1.50 (2%), 1.50 to 2.50 (17%), 2.5 to 3.5 (64%), >3.5 (17%). Study
participants were entering freshmen between the fall of 2000 and 2003. Students
completed the instruments described below during the first 2 weeks of the fall
semester. Students completed the College Self-Efficacy Inventory again during
the last 2 weeks of the fall semester.

Measures

Achievement. Students’ ACT composite scores were obtained by consent from

institutional records. The content of the ACT is based on knowledge and skills
that are taught in typical college-preparatory curricula in high schools and that
are considered by college instructors to be essential for academic success in the
1st year of college (ACT, 1997). ACT composite scores can range from 0 to 36
with a recent national average of 20.9.

College self-efficacy. The CSEI (Solberg et al., 1993) consists of 20 items measur-
ing participants’ beliefs in their abilities to successfully complete college-related
tasks. Example items include “talk with your professors,” “make new friends at
college,” and “write a course paper.” Participants responded by indicating how
 Gore / ACADEMIC SELF-EFFICACY AND COLLEGE OUTCOMES    97

confident they were in their ability to successfully complete the tasks using a 10-
point scale from 0 (not at all confident) to 9 (extremely confident). Internal con-
sistency reliability estimates (Cronbach alphas) for scale scores in a sample of
1st-year college students range from .83 to .88 (Gore et al., in press). CSEI scores
negatively correlate with measures of physical and psychological distress and
positively correlate with adjustment, academic persistence, and social integration
(Solberg et al., 1993, 1998; Solberg & Villarreal, 1997). CSEI total scores and
scale scores are computed by averaging item responses.

Academic self-confidence. The ASC scale was developed as a measure of aca-

demic self-efficacy and is one of 10 scales in the SRI (Le et al., 2005). The SRI
was developed to measure motivation, academic-skills, and social engagement
content domains that have been identified through meta-analysis to be valid pre-
dictors of college outcomes (Robbins et al., 2004). The motivation domain mea-
sures personal characteristics that help students focus and maintain goal-directed
activity. This domain contains the ASC scale. In addition, the SRI includes an
academic skill domain, which measures the cognitive, behavioral, and affective
tools students need to complete academic-related tasks, and the social engage-
ment domain, which measures interpersonal factors that influence students’ suc-
cessful integration and adaptation into a postsecondary school environment.
The ASC scale contains 12 items. Participants are asked to indicate the degree
to which they agree or disagree with each item statement on a scale from 1 =
strongly disagree to 6 = strongly agree. Items include “I am confident of my aca-
demic abilities” and “I am a fast learner.” Le et al. (2005) reported an internal
consistency reliability estimate (Cronbach alpha) of .76 for this scale in a sample
of almost 6,000 high school seniors and 1st-year college students. Scores on the
ASC scale are calculated by summing students’ responses to each item. Possible
scores range from 12 to 72. The ASC scale was administered to only the most
recent cohort of students (2002 academic year, N = 139).

College outcomes. Students’ semester (noncumulative) GPAs and enrollment

status were obtained from institutional records. GPAs from the first three con-
secutive college semesters (excluding summer semesters) were used in this study.
Enrollment status was recorded (enrolled, not enrolled) for the second and third
consecutive semesters. Students who enrolled but withdrew from the university
were coded as enrolled. Students in the 2002 cohort who completed both the
CSEI and the ASC scale have GPA data for only their first two college semesters.

Analysis

Hierarchical linear regression was used to evaluate the degree to which ACT
composite, CSEI, and ASC scores predict college GPA. Students’ ACT compos-
ite scores were entered into the analysis in the first step. In Step 2, CSEI subscale
scores (Academic, Social, and Roommate) were entered as a block. Separate
98    JOURNAL OF CAREER ASSESSMENT / February 2006

regressions were conducted using GPAs from the first three college semesters.
Additional analyses were conducted on the small subsample of students who
completed both the CSEI and the ASC scale. In these analyses, CSEI and ASC
scores were entered as a block in Step 2 of the regression.
Hierarchical logistic regression analyses were conducted to evaluate the degree
to which ACT composite and CSEI scores predict college retention. Separate
analyses were conducted using first- to second-semester and 1st- to 2nd-year reten-
tion as dependent variables. Given the small sample of students completing the
ASC and the inherently low base rate of attrition, we decided not to include ASC
scale scores in these logistic regression analyses.
Finally, because CSEIs were completed by our students at both the beginning
and the end of the first semester, and because self-efficacy beliefs are informed by
personal experience, another set of analyses were conducted to determine if end-
of-semester CSEI scores predict colleges students’ performance or persistence.

Results

Descriptive statistics and correlations among the measures used in this study
are presented in Table 1. Small to moderate positive relationships were observed
between scores on the ASC scale and scores on scales of the CSEI, with the high-
est correlation being between ASC and the Course Self-Efficacy scale of the
CSEI. ASC and CSEI were significant, albeit weak, predictors of college GPA.
Of some note, however, are the larger correlations between end-of-semester CSEI
scores and GPA. A subsample of students completed both the ASC and the CSEI.
Correlations between the two measures in this sample ranged from .07 to .46.
Results from the hierarchical linear regression analyses are presented in Table
2. ACT composite score was a significant predictor of GPA across the first three
semesters of college, accounting for between 6% and 7% of the variance. In gen-
eral, CSEI scores obtained at the beginning of students’ first semester in college
failed to account for additional variance in GPA. In contrast, when CSEI scores
obtained from students at the end of the first semester of college were used, col-
lege self-efficacy was a significant predictor of GPA in all three analyses. CSEI
scores accounted for an additional 10% of the variance in first- and second-semester
GPAs and an additional 4% of the variance in third-semester GPA. Of the three
subscales, course self-efficacy was the most consistent predictor of college GPA.
A second set of analyses were conducted using a subsample (n = 137) of stu-
dents who completed both the CSEI and the ASC scale. Results from these
analyses are presented in Table 3. Because of the small sample size used in this
cohort analysis, we will focus on comparing effect sizes observed here to those
reported in the larger analysis. In general, the magnitude of the relationship
between ACT composite scores and GPA was similar to that observed in Table 2.
When taken together, academic self-confidence and college self-efficacy mea-
sured at the beginning of the first college semester accounted for between 2% and
(text continues on p. 104)
 Table 1
Descriptive Statistics and Correlations Among Variables in Study 1

Variables 1 2 3 4 5 6 7 8 9 10 11 12 13
  1. ASC 1.00
  2. CSEI(1)–Total score .37 1.00
  3. CSEI(1)–Course self-efficacy .46 .82 1.00
  4. CSEI(1)–Social self-efficacy .22 .90 .54 1.00
  5. CSEI(1)–Roommate self-efficacy .22 .76 .49 .61 1.00
  6. CSEI(2)–Total score .31 .63 .52 .59 .45 1.00
  7. CSEI(2)–Course self-efficacy .37 .52 .56 .39 .33 .88 1.00
  8. CSEI(2)–Social self-efficacy .22 .62 .42 .66 .40 .93 .67 1.00
  9. CSEI(2)–Roommate self-efficacy .07 .49 .34 .43 .51 .77 .54 .65 1.00
10. Semester 1 GPA .14 .10 .11 .09 .04 .23 .34 .16 .06 1.00
11. Semester 2 GPA .19 .13 .13 .11 .07 .27 .35 .20 .10 .67 1.00
a
12. Semester 3 GPA .05 .05 .06 .00 .19 .21 .19 .06 .55 .59 1.00
13. ACT composite .38 –.03 .04 –.05 –.07 .07 .15 –.04 .03 .24 .24 .26 1.00
Mean 49.95 6.85 6.45 7.73 6.78 7.01 6.76 7.56 6.95 2.77 2.50 2.46 20.73
Standard deviation 8.86 1.08 1.30 1.30 1.17 1.17 1.39 1.32 1.25 0.85 1.01 0.95 3.52

Note. ASC = Academic Self-Confidence scale from the Student Readiness Inventory; CSEI = College Self-Efficacy Inventory; GPA = grade point average.
Semester GPAs are noncumulative. CSEI scores obtained at the beginning of the first semester are designated by (1) and CSEI scores obtained at the end
of the first semester are designated by (2). Correlations >.079 are statistically significant (p < .05).
a. Third-semester GPA not available for students in the subsample who completed the ASC scale.

99
   100
Table 2
Hierarchical Linear Regression of College GPA on ACT Composite Score and Academic Self-Efficacy in Study 1

Dependent Variable Step Predictor Variable B SE B b R R2d F

a
First semester GPA 1 ACT composite 0.057 0.010 .235 .237 .056 34.892a
2 CSEI(1) Course 0.039 0.032 .059 .262 .012 2.606
CSEI(1) Roommate –0.016 0.038 –.022
CSEI(1) Social 0.053 0.035 .081
Second semester GPA 1 ACT composite 0.069 0.012 .240a .240 .058 34.091a
2 CSEI(1) Course 0.055 0.039 .070 .274 .018 3.54
CSEI(1) Roommate 0.003 0.047 .003
CSEI(1) Social 0.063 0.044 .079
Third semester GPA 1 ACT composite 0.071 0.014 .267a .265 .070 25.613a
2 CSEI(1) Course –0.009 0.046 –.014 .276 .006 0.692
CSEI(1) Roommate –0.020 0.055 –.025
CSEI(1) Social 0.069 0.052 .094
 First semester GPA 1 ACT composite 0.042 0.009 .171a .239 .057 37.001a
2 CSEI(2) Course 0.248 0.032 .400a .403 .105 25.469a
CSEI(2) Roommate –0.094 0.034 –.137a
CSEI(2) Social –0.016 0.037 –.026
Second semester GPA 1 ACT composite 0.050 0.011 .173a .237 .056 34.551a
2 CSEI(2) Course 0.283 0.040 .385a .405 .108 24.743a
CSEI(2) Roommate –0.076 0.042 –.093
CSEI(2) Social –0.012 0.047 –.015
Third semester GPA 1 ACT composite 0.063 –0.014 .233a .263 .069 26.288a
2 CSEI(2) Course 0.095 0.060 .125 .331 .040 5.297a
CSEI(2) Roommate –0.087 0.053 –.119
CSEI(2) Social 0.117 0.062 .162

Note. GPA = grade point average; CSEI = College Self-Efficacy Inventory. Standardized and nonstandardized beta estimates reflect values for the final regression
model. Reported R, R2d, and F values correspond to the regression steps. Step 2 of the regression contains ACT composite scores + CSEI subscale scores
in all analyses. CSEI scores obtained at the beginning of the first semester are designated by (1) and CSEI scores obtained at the end of the first semester
are designated by (2).
a. Significant values at p < .05.

101
   102
Table 3
Hierarchical Linear Regression of College GPA on ACT, Academic Self-Confidence,
and Academic Self-Efficacy in a Subsample of Study 1 Participants

Dependent Variable Step Predictor Variable B SE B b R R2d F

a
First semester GPA 1 ACT composite 0.044 0.023 .176 .199 .040 5.572
2 ASC 0.069 0.117 .063 .241 .019 0.645
CSEI(1) Course 0.056 0.081 .070
CSEI(1) Roommate –0.087 0.081 –.112
CSEI(1) Social 0.035 0.072 .053
Second semester GPA 1 ACT composite 0.057 0.031 .177 .205 .042 5.633
2 ASC 0.132 0.155 .092 .264 .028 0.941
CSEI(1) Course 0.057 0.106 .056
CSEI(1) Roommate –0.101 0.107 –.102
CSEI(1) Social 0.096 0.095 .114
First semester GPA 1 ACT composite 0.031 0.023 .122 .199 .040 5.572a
2 ASC 0.012 0.104 .011 .334 .073 2.703a
CSEI(2) Course 0.169 0.056 .310a
CSEI(2) Roommate 0.010 0.070 .015
CSEI(2) Social –0.093 0.072 –.136
 Second semester GPA 1 ACT Composite 0.044 0.030 .136 .205 .042 5.633a
2 ASC 0.040 0.137 .028 .362 .089 3.210a
CSEI(2) Course 0.206 0.072 .296a
CSEI(2) Roommate –0.018 0.093 –.020
CSEI(2) Social 0.011 0.097 .012

Note. GPA = grade point average; ASC = Academic Self-Confidence Scale; CSEI = College Self-Efficacy Inventory. Standardized and nonstandardized beta
estimates reflect values for the final regression model. Reported R, R2Δ, and F values correspond to the regression steps. Step 2 of the regression contains
ACT composite scores + ASC scores + CSEI subscale scores in all analyses. CSEI scores obtained at the beginning of the first semester are designated by
(1) and CSEI scores obtained at the end of the first semester are designated by (2). At the time this study was written, the participants in the subsample used
for these analyses had completed a maximum of two semesters of college.
a. Significant values at p < .05.

103
104    JOURNAL OF CAREER ASSESSMENT / February 2006

3% of the variance in GPA. In contrast, when end-of-semester college self-efficacy

scores were used, ASC and CSEI accounted for between 7% and 9% of the vari-
ance in GPA. Regression coefficients in these analyses suggest that course-related
college self-efficacy beliefs are the strongest predictors of college GPA.
Results from the hierarchical logistic regression of retention on CSEI scores
are presented in Table 4. The high base rate of persistence in this sample (97%
persistence S1-S2 and 83% persistence S2-S3) makes the use of classification
accuracy problematic. At all steps of the analyses the model predicted 100%
retention. We can, however, evaluate the incremental model fit afforded by the
addition of ACT composite and CSEI scores. To do this we used the Wilks’s χ2
(referred to as G2 or 2×log-likelihood ratio). Incremental model fit was evaluated
by comparing G2 changes at each step of the analysis. Changes in G2 are distrib-
uted as a χ2 with degrees of freedom equal to the difference in degrees of freedom
between the two steps. In general, our results suggest that the models containing
ACT composite scores or ACT composite scores plus CSEI scale scores are not
significantly better models than the null intercept model. The sole exception to
this finding was in the model testing the prediction of 1st- to 2nd-year retention
using end-of-semester CSEI scores. In this model, the addition of CSEI scores
resulted in a model that improved on the use of ACT scores alone.

STUDY 2

Method

Sample and Procedures

Institutions. A stratified sample of 4-year degree-granting postsecondary institu-

tions was created to ensure representation across levels of admission selectivity
and geographic location. A majority (N = 25) of the institutions in this sample
agreed to participate in the study. Participating institutions agreed to make a 2-
year commitment to the study, administer the SRI to a minimum of 200 1st-year
students before or within the first 6 weeks of the fall 2003 semester, and provide
individual participant’s semester GPAs and retention information at the end of
the fall 2003, spring 2004, and fall 2004 semesters.

Participants. A total of 7,956 incoming 1st-year students were available in our

sample. Students were randomly selected to participate in the study at the insti-
tutional level. Participants were solicited during summer orientation programs,
during fall orientation programs, and/or within the first 6 weeks of the fall 2003
semester. The majority of participants were female (54.9%) and Caucasian
(67.7%), whereas 20% were African American, 6.8% were Hispanic/Latino, 2%
were Asian, 0.8% were American Indian/Alaskan Native, 0.2% were Native
 Table 4
Hierarchical Logistic Regression of College Persistence on ACT and College Self-Efficacy in Study 1

Dependent Variable Step Variable B SE Exp(B) G df Gdiff dfdiff

S1-S2 persistence 0 Intercept 0.772 1.879 2.164 189.01 0
1 ACT composite 0.061 0.064 1.062 188.35 1 0.66 1
2 CSEI(1) Course –0.178 0.213 0.837 185.42 4 2.93 3
CSEI(1) Roommate 0.069 0.229 1.072
CSEI(1) Social 0.285 0.204 1.330
S2-S3 persistence 0 Intercept 0.806 1.016 2.239 558.39 0
1 ACT composite 0.029 0.031 1.029 557.67 1 0.72 1
2 CSEI(1) Course –0.060 0.103 0.942 554.37 4 3.30 3
CSEI(1) Roommate –0.102 0.121 0.903
CSEI(1) Social 0.199 0.109 1.220
S1-S2 persistence 0 Intercept –1.080 1.474 0.340 210.26 0
1 ACT composite 0.102 0.059 1.107 26.91 1 3.35 1
2 CSEI(2) Course –0.085 0.202 0.918 201.89 4 5.02 3
CSEI(2) Roommate 0.122 0.196 1.130
CSEI(2) Social 0.281 0.213 1.324

(continued)

105
106
Table 4 (continued)

Dependent Variable Step Variable B SE Exp(B) G df Gdiff dfdiff

S2-S3 persistence 0 Intercept 0.067 0.896 1.069 586.23 0
1 ACT composite 0.025 0.031 1.025 584.49 1 1.74 1
2 CSEI(2) Course 0.197 0.107 1.218 574.51 4 9.98a 3
CSEI(2) Roommate –0.173 0.112 0.841
CSEI(2) Social 0.140 0.118 1.150

Note. CSEI = College Self-Efficacy Inventory. S1-S2 denotes first- to second-semester persistence, S2-S3 denotes second- to third-semester persistence (spring
to fall, not including summer). B is the nonstandardized logistic regression coefficient for the final model, and SE is the standard error of that coefficient.
Exp(B) is the odds ratio for each predictor in the final model. G refers to the log likelihood ratio. Gdiff and dfdiff refer to the differences between (a) the intercept
model and the intercept + ACT composite model, and (b) the ACT composite model and the full model with ACT composite + subscales from the CSEI.
CSEI scores obtained at the beginning of the first semester are designated by (1) and those obtained at the end of the first semester are designated by (2).
a. Significant values at p < .05.
 Gore / ACADEMIC SELF-EFFICACY AND COLLEGE OUTCOMES    107

Hawaiian/Other Pacific Islander, and 2.3% were of another race or ethnicity or

preferred not to respond. Participants varied in age from 16 to 58 years old (M =
18.3, SD = 1.5).

Measures

Achievement. Student’s ACT composite scores were obtained by matching

students who participated in the study to the ACT research database. Students’
ACT composite scores ranged from 9 to 35 (M = 21.0, SD = 4.3). ACT compos-
ite scores can range from 0 to 36 with a recent national average of 20.9.

Psychosocial factors. The SRI was administrated to all participants as described

above in the procedures. This instrument was described in detail in the Methods
section of Study 1. Only results from the ASC scale are reported in this study.

Analysis

Because the purpose of this study was to determine the relationship between
academic self-confidence and college outcomes after controlling for the effects of
past achievement (e.g., ACT composite score), hierarchical linear and hierarchi-
cal logistic regression analyses were conducted. Specifically, college outcomes
(GPA or retention) were regressed first on ACT composite scores and then on
ASC scale scores. To account for variation across the 25 institutions in this sam-
ple, random intercept regression models were used.

Results

Correlations and descriptive statistics for the predictor variables and semester
GPAs are presented in Table 5. The results of hierarchical regression analyses are
shown in Table 6. ACT composite scores were significant predictors of both first-
and second-semester GPA, accounting for approximately 20% of the variance in
students’ college performance. The addition of ASC to the equation resulted in a
significant, albeit small, increase in the prediction of college performance. On
average, ACT composite scores were seven to nine times more powerful in pre-
dicting college GPA compared with ASC scores.
Results from the hierarchical linear regression analyses are presented in Table
7. Although there was a slightly higher rate of attrition in this sample compared
with the sample used in Study 1 (90% persistence S1-S2 and 74% persistence S2-
S3), still none of the models more accurately predicted attrition in this sample
compared with the baseline model. A strategy identical to that adopted in Study
1 was used to evaluate the incremental model fit afforded by the addition of ACT
composite scores and ASC scores. Incremental model fit was evaluated by com-
paring G2 changes at each step of the analysis. The use of ACT composite scores
108    JOURNAL OF CAREER ASSESSMENT / February 2006

Table 5
Descriptive Statistics and Correlations Among Variables in Study 2

Variable 1 2 3 4
1. ACT composite 1.00
2. ASC .43 1.00
3. Semester 1 GPA .36 .20 1.00
4. Semester 2 GPA .36 .19 .65 1.00
Mean 21.01 52.61 2.67 2.63
Standard deviation 4.25 9.76 0.98 1.03

Note. ASC = Academic Self-Confidence; GPA, grade point average. Semester GPAs are
noncumulative. All correlations are statistically significant (p < .05).

Table 6
Hierarchical Linear Regression of College GPA
on ACT and Academic Self-Confidence in Study 2

Dependent
Variable Step Variable B SE B b R R2d F
a
First semester 1 ACT composite 0.086 0.003 .378 .453 .205 1,089.28
  a
GPA 2 ASC 0.005 0.001 .049 .455 .002 18.46
a
Second semester 1 ACT composite 0.091 0.004 .374 .453 .205 676.97
  a
GPA 2 ASC 0.004 0.002 .041 .454 .002 6.81

Note. GPA = grade point average; ASC = Academic Self-Confidence. Standardized and
nonstandardized beta estimates reflect values for the final regression model. Reported R, R2Δ,
and F values correspond to the regression steps. Step 2 of the regression contains ACT composite
scores + ASC scores in all analyses.
a. Significant values at p < .05.

significantly improved the model describing first- to second-semester retention

(p < .01). Furthermore, a model containing ASC and ACT composite scores was
superior to one that included ACT scores alone. The same pattern of findings was
obtained in models predicting 1st- to 2nd-year persistence. Specifically, the best
fitting model was one that specified that both ACT composite scores and ASC
scores predict 1st- to 2nd-year persistence.

DISCUSSION

The present studies were conducted to further describe the relationships

between academic self-efficacy beliefs and college outcomes. In an effort to
extend existing literature, we used multiple predictor and criterion variables and
 Gore / ACADEMIC SELF-EFFICACY AND COLLEGE OUTCOMES    109

Table 7
Hierarchical Logistic Regression of College Persistence
on ACT and Academic Self-Confidence in Study 2

Dependent
Variable Step Variable B SE Exp(B) G df Gdiff dfdiff
S1-S2 0 Intercept 0.175 0.062 1.191 4,378.28 22
  persistence 1 ACT composite 0.074 0.054 1.076 4,320.37 23 57.91a 1
2 ASC 0.011 0.045 1.011 4,314.53 24 5.83a 1
S2-S3 0 Intercept –0.674 0.234 0.509 5,347.95 15
  persistence 1 ACT composite 0.071 0.010 1.073 5,269.67 16 78.28a 1
2 ASC 0.008 0.004 1.008 5264.93 17 4.74a 1

Note. ASC = Academic Self-Confidence. S1-S2 denotes first- to second-semester persistence, S2-
S3 denotes second- to third-semester persistence (also referred to as 1st- to 2nd-year persistence).
B is the nonstandardized logistic regression coefficient for the final model, and SE is the standard
error of that coefficient. Exp(B) is the odds ratio for each predictor in the final model. G refers to
the log likelihood ratio. Gdiff and dfdiff refer to the differences between (a) the intercept model and
the intercept + ACT Composite model, and (b) the ACT composite model and the full model
with ACT Composite + ACS.
a. Significant values at p < .05.

adopted an analytic strategy that would permit us to evaluate the incremental

validity of academic self-efficacy beliefs in predicting college success.
Results from the present studies suggest that, at least when measured at the
beginning of the first semester of college, academic self-efficacy beliefs are rela-
tively weak predictors of academic performance. This finding was consistent
across measures, academic semesters, and studies. Bivariate effects between CSEI
scores and GPAs ranged from .00 to .13 in Study 1. Relations between ASC scores
and GPA were only slightly higher (.14-.20) in Study 2. Academic self-efficacy
beliefs failed to account for a significant proportion of variance in GPA beyond
that accounted for by ACT composite scores. Similar findings were observed in
our logistic regression analyses of the relationships between academic self-efficacy
beliefs and college persistence. Models that included academic self-efficacy
beliefs in predicting students’ persistence across academic semesters either failed
to improve on a model that included only ACT composite scores (Study 1) or
improved on that model to an extent that may be of little practical importance
(Study 2).
A very different picture emerged when academic self-efficacy beliefs were
measured at the end of the students’ first semester in college. Correlations
between end-of-semester CSEI course subscale scores and GPA, for example,
ranged from .21 (third semester) to .35 (second semester). These relationships are
two to three times the magnitude of relationships observed between beginning-of-
semester CSEI score and GPA. End-of-semester CSEI scores accounted for
between 4% and 10% of the variance in students’ GPAs. Furthermore, CSEI
110    JOURNAL OF CAREER ASSESSMENT / February 2006

scores accounted for variance in academic performance beyond that accounted

for by ACT composite scores. These findings are consistent with previous reports.
For example, Gore et al. (in press) reported significant relationships between
GPAs and end-of-semester CSEI scores but not between GPA and CSEI scores
measured at the start of students’ first semester in college. Kahn and Nauta (2001)
reported similar findings. In a study of the performance and persistence of 1st-year
college students, these authors found stronger relationships between students’
academic self-efficacy beliefs and college performance when those efficacy
beliefs were measured during the second semester of college. A closer compari-
son between Kahn and Nauta’s bivariate correlations and those from the present
study reveals almost perfect correspondence.
A similar pattern was observed when academic self-efficacy beliefs measured at
the end of the first college semester were entered into the logistic regression equa-
tion predicting 1st- to 2nd-year persistence. The addition of CSEI scores resulted
in statistically and practically significant improvements in the model. Inspection
of the odds ratios, for example, reveal that for every 1-point increase in CSEI
course self-efficacy score, a student’s odds of returning the 2nd year increase by
22%. These findings conflict with those reported by Kahn and Nauta (2001).
These authors found that outcome expectations and performance goals, but not
academic self-efficacy beliefs, predicted students’ return in the 2nd year. Several
differences between the two studies may account for the discrepant findings. First,
Kahn and Nauta included more than one predictor variable in their equation. It
is possible that the roles of outcome expectations and performance goals over-
shadowed the relationship between academic self-efficacy and persistence in
Kahn and Nauta’s study. Second, Kahn and Nauta used a modified version of a
self-efficacy measure of broad academic milestones (Lent, Brown, & Gore, 1997).
This instrument assesses students’ confidence in their ability to complete core
academic requirements, attain a certain GPA, and succeed in college. In contrast,
the CSEI measures students’ confidence in their ability to successfully perform
more discrete college success behaviors such as research a term paper, take class
notes, or talk with professors.
That self-efficacy beliefs of experienced college students are more strongly
related to college performance and persistence than are the efficacy beliefs of
“college-naïve” students is consistent with the tenants of social cognitive theory
and previous research. According to Bandura (1986), self-efficacy beliefs develop
as a result of personal performance accomplishments, vicarious learning, persua-
sion, and the interpretation of physiological states. Bandura suggested that per-
sonal performance accomplishments are the most influential of the four sources,
and this hypothesis has gained empirical support (Lent, Lopez, & Bieschke,
1991; Lent, Lopez, Brown, & Gore, 1996). Students’ academic efficacy beliefs
are likely to be more accurate to the extent to which the students have experience
in the academic arena. Given that the CSEI measures students’ confidence in
their ability to successfully perform college-specific academic and pro-academic
social activities, students’ scores on this instrument are likely to change as stu-
 Gore / ACADEMIC SELF-EFFICACY AND COLLEGE OUTCOMES    111

dents acquire college experience. Because the ASC scale was administered only
once, and at the beginning of the semester, the generalizability of these conclu-
sions to that scale remains to be determined.
In addition to the temporal effects described above, we also observed an inter-
esting pattern of relationships between CSEI subscale scores and the college
outcome variables. Although small in magnitude and not always statistically sig-
nificant, course-related academic self-efficacy beliefs were more strongly related
to GPA, whereas CSEI social scores were more strongly related to college persis-
tence. For example, the odds ratios displayed in Table 4 suggest that students’
odds of returning to school in the second or third semester are increased by 22%
and 33%, respectively, for every 1-point change on the CSEI social subscale.
Differential relationships between CSEI subscale scores and outcome measures
were also reported by Gore et al. (in press). In that study, CSEI subscales differ-
entially related to students’ expectations of using the library (CSEI Course and
Social subscales); establishing relationships with people on campus (CSEI Social
and Roommate subscales); engaging in reading, writing, quantitative, and scien-
tific activities (CSEI Course subscale); and expected college GPA (CSEI Course
and Social subscales). Furthermore, Robbins and his colleagues (2004) found
that social support and social involvement were related to student retention but
not to student performance.
The issue of measurement specificity and correspondence between predictor
and criterion measurement permeates the self-efficacy literature. Bong and
Skaalvik (2003) suggested that the measurement of self-efficacy beliefs should be
related to the corresponding target performance of interest. Some authors have
even argued for the use of identical items to assess the relations between self-efficacy
and performance (Pajares & Kranzler, 1995; Pajares & Miller, 1994, 1995),
although the dangers of this practice have also been articulated (Lent & Hackett
1987; Marsh, Roche, Pajares, & Miller, 1997).
The two measures of academic self-efficacy used in the present study represent
similar, albeit not identical, levels of measurement specificity. Evidence support-
ing this conclusion comes from two observations. First, the two scales were only
moderately correlated with each other. Different measures of the same construct
would be expected to correlate more strongly even in light of attenuation attribut-
able to measurement error. More convincing of the distinct nature of these two
measures is the fact that they did not share similar correlations with ACT composite
scores. These findings are consistent with previous observations that different
measures of academic self-efficacy beliefs differently correlate with standardized
test scores (Lent et al., 1997). Given that the ASC scale includes more general
academic self-efficacy items, it is not surprising that it would correlate more
highly with an achievement indicator like ACT composite score. ACT scores
indicate the extent to which students have mastered high school academic con-
tent. Those mastery (or failure) experiences would also inform students’ general
academic self-efficacy estimates. In contrast, because items on the CSEI are more
112    JOURNAL OF CAREER ASSESSMENT / February 2006

specific to behaviors that are likely to occur in college, students’ efficacy estimates
on this instrument are unlikely to correlate highly with past measures of aca-
demic achievement.
Our findings contribute to the growing body of literature suggesting that aca-
demic self-efficacy beliefs can be used to predict college students’ academic per-
formance and persistence. However, our results reinforce the work of previous
researchers and suggest that the utility of academic self-efficacy as a predictor of
college success may be partially dependent on (a) when self-efficacy beliefs are
measured, (b) what aspect of self-efficacy is being measured, and (c) what college
outcome one wishes to predict. Taken together, these observations have impor-
tant practical implications.
Our results suggest that students need feedback on their performance (both
social and academic) before they can realistically assess their ability to achieve
academic goals. This finding argues against the use of academic self-efficacy
belief instruments as preadmission enrollment management tools or with recent-
ly enrolled 1st-year students. Rather, our findings suggest that college and univer-
sity personnel are better off assessing self-efficacy beliefs at the end of the first
semester or beginning of the second semester of college. Academic self-efficacy
belief assessment might be used to identify students who could benefit from aca-
demic interventions such as tutoring, Supplemental Instruction, advising, or
study skills workshops. Alternatively, academic self-efficacy beliefs measures such
as the ASC scale or CSEI could be used to evaluate the effectiveness of these or
other student success programs (Gore et al., in press).
Our results suggest that the first semester of college is a critical time for pro-
moting the academic self-efficacy beliefs of incoming 1st-year students. First-year
interventions such as Supplemental Instruction (Ramirez, 1997) and the First-
Year Experience (Gardner, 1986) are designed to promote success among 1st-year
students. The First-Year Experience (FYE) program in particular focuses on pro-
moting both the academic and pro-academic social skill development of fresh-
man students. FYE programs orient students to their new academic milieu (cam-
pus social and support structures) by providing students with opportunities to
apply newly learned self-regulatory and study skills and by creating a safe environ-
ment in which academic success behaviors can be modeled and practiced. By
encouraging students to become involved in student organizations and campus
activities and by helping students establish a professional mentor relationship
with a faculty member, FYE programs may serve to further bolster students’ aca-
demic and pro-academic social self-efficacy beliefs.
Finally, programs such as the FYE offer researchers an excellent platform with
which to further study the role of academic self-efficacy beliefs in college stu-
dents’ success. Researchers may use instruments such as the CSEI and ASC scale
to further investigate why entering students’ efficacy estimates are so poor at pre-
dicting their outcomes. Moderators such as high school achievement or personal-
ity characteristics present themselves as obvious candidates. Alternatively, these
 Gore / ACADEMIC SELF-EFFICACY AND COLLEGE OUTCOMES    113

instruments could be added to other dependent measures and used to launch a

process research program designed to uncover exactly which components of the
FYE are most efficacious in preparing successful students.

References

ACT. (1997). ACT assessment program technical manual. Iowa City, IA: Author.
Astin, A. W. (1999). Student involvement: A developmental theory for higher education. Journal of
College Student Development, 40, 518-529.
Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change. Psychological
Review, 84, 191-215.
Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory. Englewood
Cliffs, NJ: Prentice Hall.
Betz, N. E., & Hackett, G. (1981). The relationship of career-related self-efficacy expectations to
the perceived career options in college women and men. Journal of Counseling Psychology, 28,
399-410.
Boldt, R. F. (1986). Generalization of SAT validity across colleges (College Board Rep. No. 86-3).
New York: College Entrance Examination Board.
Bong, M. (1997). Generality of academic self-efficacy judgments: Evidence of hierarchical rela-
tions. Journal of Educational Psychology, 89, 696-709.
Bong, M. (2004). Academic motivation in self-efficacy, task value, achievement goal orientations,
and attributional beliefs. Journal of Educational Research, 97, 287-297.
Bong, M., & Skaalvik, E. M. (2003). Academic self-concept and self-efficacy: How different are they
really? Educational Psychology Review, 15, 1-40.
Brown, S. D., Lent, R. W., & Larkin, K. C. (1989). Self-efficacy as a moderator of scholastic apti-
tude-academic performance relationships. Journal of Vocational Behavior, 35, 34-75.
Chemers, M. M., Hu, L., & Garcia, B. F. (2001). Academic self-efficacy and first-year college stu-
dent performance and adjustment. Journal of Educational Psychology, 93, 55-64.
Elias, S. M., & Loomis, R. J. (2002). Utilizing the need for cognitive and perceived self-efficacy to
predict academic performance. Journal of Applied Social Psychology, 32, 1687-1702.
Gardner, J. N. (1986). The freshman year experience. College and University, 61, 261-274.
Gore, P. A., Jr., Leuwerke, W. C., & Turley, S. E. (in press). A psychometric study of the College
Self-Efficacy Inventory. Journal of College Student Retention.
Hackett, G., & Betz, N. E. (1981). A self-efficacy approach to the career development of women.
Journal of Vocational Psychology, 18, 326-339.
Harvey, V., & McMurray, N. (1994). Self-efficacy: A means of identifying problems in nursing
education and career progress. International Journal of Nursing Studies, 31, 471-485.
Juang, L., & Vondracek, F. (2001). Developmental patterns of adolescent capability beliefs: A per-
son approach. Journal of Vocational Behavior, 59, 34-52.
Kahn, J. H., & Nauta, M. M. (2001). Social-cognitive predictors of first-year college persistence:
The importance of proximal assessment. Research in Higher Education, 42, 633-652.
Le, H., Casillas, A., Robbins, S. B., & Langley, R. (2005). Motivational and skills, social, and self-
management predictors of college outcomes: Constructing the Student Readiness Inventory.
Educational and Psychological Measurement, 62, 1-28.
Leach, C. W., Queirolo, S. S., DeVoe, S., & Chemers, M. (2003). Choosing letter grade evalu-
ations: The interaction of students’ achievement goals and self-efficacy. Journal of Vocational
Behavior, 28, 495-509.
114    JOURNAL OF CAREER ASSESSMENT / February 2006

Lent, R. W., Brown, S. D., & Gore, P. A., Jr. (1997). Discriminant and predictive validity of aca-
demic self-concept, academic self-efficacy, and mathematics-specific self-efficacy. Journal of
Counseling Psychology, 44, 307-315.
Lent, R. W., Brown, S. D., & Larkin, K. C. (1986). Self-efficacy in the prediction of academic per-
formance and perceived career options. Journal of Counseling Psychology, 33, 265-269.
Lent, R. W., Brown, S. D., & Larkin, K. C. (1987). Comparison of three theoretically derived
variables in predicting career and academic behavior: Self-efficacy, interest congruence, and
consequence thinking. Journal of Counseling Psychology, 34, 293-298.
Lent, R. W., & Hackett, G. (1987). Career self-efficacy: Empirical status and future directions.
Journal of Vocational Behavior, 30, 347-382.
Lent, R. W., Lopez, F. G., & Bieschke, K. J. (1991). Mathematics self-efficacy: Sources and rela-
tions to science-based career choice. Journal of Counseling Psychology, 38, 424-430.
Lent, R. W., Lopez, F. G., Brown, S. D., & Gore, P. A., Jr. (1996). Latent structure of the sources
of mathematics self-efficacy. Journal of Vocational Behavior, 49, 292-308.
Lopez, F. G., Lent, R. W., Brown, S. D., & Gore, P. A., Jr. (1997). Role of social-cognitive expec-
tations in high school students’ mathematics-related interest and performance. Journal of
Counseling Psychology, 44, 44-52.
Marsh, H. W., Roche, L. A., Pajares, F., & Miller, D. (1997). Item-specific efficacy judgments
in mathematical problem solving: The downside of standing too close to trees in a forest.
Contemporary Educational Psychology, 22, 363-377.
Mathiasen, R. E. (1984). Predicting college academic achievement: A research review. College
Student Journal, 18, 380-386.
Mone, M. A. (1994). Comparative validity of two measures of self-efficacy in predicting academic
goals and performance. Educational and Psychological Measurement, 54, 516-529.
Mone, M. A., Baker, D. D., & Jeffries, F. (1995). Predictive validity and time dependency of self-
efficacy, self-esteem, personal goals, and academic performance. Educational and Psychological
Measurement, 55, 716-727.
Multon, K. D., Brown, S. D., & Lent, R. W. (1991). Relation of self-efficacy beliefs to academic
outcomes: A meta-analytic investigation. Journal of Counseling Psychology, 38, 30-38.
Pajares, F., & Kranzler, J. (1995). Self-efficacy beliefs and general mental ability in mathematical
problem solving. Contemporary Educational Psychology, 20, 426-443.
Pajares, F., & Miller, M. (1994). Role of self-efficacy and self-concept beliefs in mathematical prob-
lem solving: A path analysis. Journal of Educational Psychology, 86, 193-203.
Pajares, F., & Miller, M. D. (1995). Mathematics self-efficacy and mathematics performances: The
need for specificity of assessment. Journal of Counseling Psychology, 42, 190-198.
Pintrich, P. R., & De Groot, E. V. (1990). Motivational and self-regulated learning components of
classroom academic performance. Journal of Educational Psychology, 82, 33-40.
Ramirez, G. M. (1997). Supplemental instruction: The long-term impact. Journal of Developmental
Education, 21, 2-10.
Robbins, S., Allen, J., Casillas, A., & Peterson, C. (2005). Unraveling the differential effects of
standardized achievement and psychosocial predictors of college outcomes: A prospective study.
Unpublished manuscript.
Robbins, S. B., Lauver, K., Le, H., Davis, D., Langley, R., & Carlstrom, A. (2004). Do psychosocial
and study skills factors predict college outcomes? A meta-analysis. Psychological Bulletin, 130,
261-288.
Roeser, R. W., Midgley, C., & Urdan, T. C. (1996). Perceptions of the school psychological envi-
ronment and early adolescents’ psychological and behavioral functioning in school: The mediat-
ing role of goals and belonging. Journal of Educational Psychology, 88, 408-422.
Schunk, D. H. (1991). Self-efficacy and academic motivation. Educational Psychology, 26, 207-
231.
 Gore / ACADEMIC SELF-EFFICACY AND COLLEGE OUTCOMES    115

Sechrest, L. (1963). Incremental validity: A recommendation. Educational and Psychological

Measurement, 23, 153-158.
Siegel, R. G., Galassi, J. P., & Ware, W. B. (1985). A comparison of two models for predicting
mathematics performance: Social learning versus math aptitude-anxiety. Journal of Counseling
Psychology, 32, 531-538.
Solberg, V. S., Gusavac, N., Hamann, T., Felch, J., Johnson, J., Lamborn, S., et al. (1998). The
Adaptive Success Identity Plan (ASIP): A career intervention for college students [Monograph].
Career Development Quarterly, 47, 48-95.
Solberg, V. S., O’Brien, K., Villarreal, P., Kennel, R., & Davis, B. (1993). Self-efficacy and Hispanic
college students: Validation of the College Self-Efficacy Instrument. Hispanic Journal of
Behavioral Sciences, 15, 80-95.
Solberg, V. S., & Villarreal, P. (1997). Examination of self-efficacy, social support, and stress as pre-
dictors of psychological and physical distress among Hispanic college students. Hispanic Journal
of Behavioral Sciences, 19, 182-201.
Tinto, V. (1993). Leaving college: Rethinking the causes and cures of student attrition (2nd ed.).
Chicago: University of Chicago Press.
Wood, R. E., & Locke, E. A. (1987). The relation of self-efficacy and grade goals to academic per-
formance. Educational and Psychological Measurement, 47, 1013-1024.
Zimmerman, B. J. (1986). Development of self-regulated learning. Which are the key subprocesses?
Contemporary Educational Psychology, 16, 307-313.
Zimmerman, B. J. (1990). Self-regulated learning and academic achievement: An overview.
Educational Psychologist, 25, 3-17.
Zimmerman, B. J., Bandura, A., & Martinez-Pons, M. (1992). Self-motivation for academic attain-
ment: The role of self-efficacy beliefs and personal goal setting. American Educational Research
Journal, 29, 663-676.
Zimmerman, B. J., & Martinez-Pons, M. (1990). Student differences in self-regulated learning:
Relating grade, sex, and giftedness to self-efficacy and strategy use. Journal of Educational
Psychology, 82, 51-59.