Learning and Individual Differences 37 (2015) 210–216

Contents lists available at ScienceDirect

Learning and Individual Differences

journal homepage: www.elsevier.com/locate/lindif

Hope, self-efficacy, optimism, and academic achievement: Distinguishing

constructs and levels of specificity in predicting college
grade-point average☆
David B. Feldman ⁎, Maximilian Kubota
Santa Clara University, United States

a r t i c l e i n f o a b s t r a c t

Article history: Research shows that Snyder's (1994) goal-directed hope construct predicts college GPA. However, studies have
Received 20 May 2014 documented relatively weak relationships between these variables, possibly because hope was measured regard-
Received in revised form 12 November 2014 ing goals generally, not academic-specific goals. Additionally, most studies have not compared variance
Accepted 13 November 2014
accounted for in GPA by hope relative to other expectancy constructs. In a cross-sectional sample of 89 college
students, we administer the Hope Scale, Domain Specific Hope Scale (academic subscale), General Self-Efficacy
Keywords:
Hope Theory
Scale, Academic Self-Efficacy Scale, Life Orientation Test-Revised (optimism), among others. We test a path-
GPA analytic model where academic-specific expectancies (e.g., academic hope, academic self-efficacy) have direct
Academic achievement paths to GPA, and generalized expectancies (e.g., general hope, general self-efficacy) have paths to these
Optimism academic-specific variables. A modified version of this hypothesized model demonstrated good fit. Generalized
Self-efficacy hope predicted academic-specific hope and academic self-efficacy, both of which then predicted GPA. Optimism
and general self-efficacy did not predict academic-specific expectancy variables nor GPA.
© 2014 Elsevier Inc. All rights reserved.

Grade-point average (GPA) is one of the most important indicators cifically. The present study measures hope both in this general way and
of college success, potentially influencing both financial aid in the an academic-specific way. Second, few past studies have examined
short term and career prospects in the long term. This study seeks to hope's ability to account for variance in GPA relative to other related
examine the relationship between hope and GPA in college students. and widely used expectancy constructs. Namely, self-efficacy and opti-
Two decades of research show that the cognitive, goal-directed con- mism have both been shown to predict GPA (Chemers, Hu, & Garcia,
ceptualization of hope developed by Snyder (1994) predicts a variety of 2001; Segerstrom & Nes, 2006). Though past studies have examined
outcomes in college students, including GPA. We detail this conceptual- combinations of hope with one of these additional constructs at a
ization of hope in the subsequent section. The notion that hope predicts time, to our knowledge no study has analyzed all three (i.e., hope,
academic achievement is not a new one. Research supporting this rela- self-efficacy, and optimism) simultaneously in predicting GPA. When
tionship has been both cross-sectional (Buckelew, Crittendon, Butkovic, all three variables are modeled together, it is unknown to what degree
Price, & Hurst, 2008; Davidson, Feldman, & Margalit, 2012) and prospec- they overlap or which are better predictors of GPA. Thus, we assess
tive (Rand, Martin, & Shea, 2011; Snyder et al., 2002). self-efficacy and optimism in addition to hope.
The present study augments this literature in two ways: First, the Before detailing the present study, it may be helpful to explain Hope
aforementioned studies have documented only relatively weak rela- Theory in greater detail, review past hope research, and discuss the
tionships between hope and GPA, even using cross-sectional data. Cor- relationships between hope and other expectancy constructs.
relations ranged between .18 and .29. One explanation for the lack of a
stronger relationship may be the very general way hope is typically 1. Hope Theory
measured. In most studies, hope was assessed using the Adult Hope
Scale (AHS; Snyder et al., 1991), which measures hope regarding re- The model of hope used throughout this article is based on Snyder's
spondents' life goals in general, not necessarily their academic goals spe- (1994) Hope Theory, probably the most researched conceptualization
of hope during the past two decades. Accordingly, hope is the cognitive
process allowing individuals to plan for and execute the pursuit of goals.
☆ This research was presented, in part, at the Society for Personality and Social Psychology
Goals are anything that an individual desires to get, do, be, experience,
2014 annual meeting in Austin, Texas.
⁎ Corresponding author at: Department of Counseling Psychology, Santa Clara University,
or create. They can vary substantially in terms of scope as well as
500 El Camino Real, Santa Clara, CA 95053, United States. Tel.: +1 408 554 4462. among people and life domains (i.e., academics, family, romantic life,
E-mail address: [email protected] (D.B. Feldman). etc.).

http://dx.doi.org/10.1016/j.lindif.2014.11.022
1041-6080/© 2014 Elsevier Inc. All rights reserved.
 D.B. Feldman, M. Kubota / Learning and Individual Differences 37 (2015) 210–216 211

According to Snyder et al. (1991), hope is “a cognitive set that is based Empirically, research has shown that hope accounts for variance
on a reciprocally-derived sense of successful agency (goal-directed over and above these variables in pain tolerance (Snyder et al., 2005),
determination) and pathways (planning to meet goals)” (p. 571). The problem-focused coping (Snyder et al., 1991), and mental illness symp-
first of these components, agency, consists of “the thoughts that people toms (Snyder et al., 1991) in college students, academic effort in children
have regarding their ability to begin and continue movement on selected (Lackaye & Margalit, 2006), and quality of life in substance-abusing
pathways toward those goals” (Snyder, Michael, & Cheavens, 1999). As homeless veterans (Irving, Seidner, Burling, Pagliarini, & Robbins-Sisco,
in Watty Piper's The Little Engine That Could, agency thoughts such as 1998), among others. Moreover, measures of hope have been shown to
“I think I can,” provide the motivation to initiate and sustain the goal- load on different factors than optimism (Bryant & Cvengros, 2004;
pursuit process (Piper, 1978; for empirical support, see Snyder, Gallagher & Lopez, 2009; Magalleta & Oliver, 1999) and self-efficacy
LaPointe, Crowson, & Early, 1998). The second component, pathways (Magalleta & Oliver, 1999).
thinking, consists of cognitions regarding planning routes to reach With regard to GPA, Rand et al. (2011) assessed both hope and opti-
goals. Because some plans may not succeed, hopeful people are theo- mism at the beginning of students' first semester in law school. Path
rized to produce multiple pathways in order to circumvent possible ob- analysis showed that hope but not optimism predicted first-semester
stacles (Snyder, 2002). It is worth noting, however, that the subjective GPA. In another study, Rand (2009) demonstrated that the shared vari-
experience of hope does not depend on the existence of concrete path- ance between hope and optimism predicted grades in undergraduates.
ways, but rather upon a belief or expectancy that such effective path- In addition, hope maintained a unique, indirect effect on grades through
ways exist (Snyder et al., 1991). grade expectancy, whereas optimism did not. Levi, Einav, Ziv, Raskind,
Researchers have generated a sizable literature demonstrating rela- and Margalit (2013) similarly found that hope indirectly predicted
tionships between hope and such variables as athletic performance grades through grade expectancies in tenth graders. They further
(Curry & Snyder, 2000; Curry, Snyder, Cook, Ruby, & Rehm, 1997), psy- found that hope contributed to grade expectancies, whereas self-
chological adjustment (Chang & DeSimone, 2001; Feldman & Snyder, efficacy did not. No research to date models the comparative contribu-
2005), coping with physical illness (Irving, Snyder, & Crowson, 1998), tions of all three expectancy constructs (optimism, self-efficacy,
life meaning (Feldman & Snyder, 2005), finding benefit in adversity and hope) together in predicting GPA. Given the past findings just
(Tennen & Affleck, 1999), and general goal achievement (Feldman, cited demonstrating that hope often accounts for variance in out-
Rand, & Kahle-Wrobleski, 2009), among others. comes (particularly GPA) over and above both optimism and self-
It is not difficult to see why hope may be related to GPA. School is a efficacy, we expected hope to be the best predictor of GPA in the present
goal-directed activity (Alkharusi, 2010). To succeed likely requires set- study.
ting goals (e.g., getting an A on a test, giving an excellent presentation),
determining pathways for achieving those goals (e.g., attending class,
reading course material, studying), and generating agency to move 3. Generalized versus specific expectancies
the process forward.
In the Rand (2009) study just mentioned, neither hope nor optimism
2. Hope, self-efficacy, and optimism directly predicted students' grades. However, hope uniquely predicted
grade expectancies, which, in turn, predicted grades. This points to the
Hope is a cousin of other expectancy constructs like self-efficacy and potential importance of assessing expectancies at the appropriate level
optimism. All three of these constructs concern expectancies regarding —general vs. academic domain-specific.
the attainment of future positive states. As such, there have been calls Hope can be assessed at both the general as well as the domain-
for theoretically and empirically distinguishing hope (Aspinwall & specific levels. The Adult Hope Scale (AHS; Snyder et al., 1991) assesses
Leaf, 2002; Tennen, Affleck, & Tennen, 2002). hope for goals in general, without reference to a particular life domain.
Contrasting hope with self-efficacy, Snyder (1995, 2002) has written In contrast, the Domain-Specific Hope Scale (DSHS; Sympson, 1999,
that self-efficacy is concerned principally with the expectancy that one 2000) measures hope in six domains: social, academic, family–home,
can perform behaviors, whereas hope is concerned with expectancies romantic, work, and leisure.
that one can attain goals. Consistent with this assertion, Bandura Historically, many researchers have assessed self-efficacy for partic-
(1977) has contrasted self-efficacy with expectancies of goal outcomes ular domains or behaviors. Bandura (1997) has advocated this ap-
in the following way: “Outcome and efficacy expectations are differen- proach, writing that “efficacy beliefs should be measured in terms of
tiated, because individuals can believe that a particular course of action particularized judgments of capability that may vary across realms of
will produce certain outcomes, but if they entertain serious doubts activity, under different levels of task demands within a given activity
about whether they can perform the necessary activities, such informa- domain, and under different situational circumstances” (p. 42). With
tion does not influence their behavior.” (p. 193). Self-efficacy is largely regard to academics, the Academic Self-Efficacy Scale (ASES) assesses
agnostic regarding whether an action will lead to goal outcomes, where- such domain-specific self-efficacy. However, research suggests
as hope concerns expectancies that one can achieve goals through the that self-efficacy beliefs also can be generalized across domains
combination of goal-directed planning (pathways) and motivation (e.g., Scholz, Doña, Sud, & Schwarzer, 2002). Thus, researchers have de-
(agency). veloped instruments such as the General Self-Efficacy Scale (GSES;
Optimism, on the other hand, is a type of outcome expectancy. Schwarzer & Jerusalem, 1995) to assess individuals' perceptions of
Contrasting hope with optimism, Snyder (1995, 2002) has written their ability to perform overall, across situations. Of note, there is dis-
that optimism is primarily concerned with the expectancy that positive agreement among self-efficacy researchers regarding the utility of
outcomes will occur without regard to one's actions, whereas hope is the general self-efficacy construct, with some asserting that self-
explicitly concerned with expectancies that attainment of positive out- efficacy in principle should always be measured in a more specific
comes will occur through one's own planning (pathways) and motiva- manner.
tion (agency). Scheier, Carver, and Bridges (2001) define optimism as Optimism, by definition, is the generalized expectancy of positive
“expectancies that are generalized—expectancies that pertain more or outcomes (Carver & Scheier, 2002). Thus, no domain-specific measure
less to the individual's entire life space” (p. 190). This difference in the has been widely used. Scheier and Carver (1985) write, “Optimists
necessity of one's personal involvement in bringing about goals has often appear to be optimistic ‘in general,’ in that their positive expecta-
led Rand et al. (2011) to suggest that, in highly-controllable situations tions are not limited to a particular behavioral domain.” (p. 220). The
like school, hope may be a stronger predictor of performance and Life Orientation Test, Revised (LOT-R, Scheier, Carver, & Bridges, 1994)
well-being than optimism. is the most frequently used measure of general optimism.
212 D.B. Feldman, M. Kubota / Learning and Individual Differences 37 (2015) 210–216

4. Present study Last, participants completed a packet of questionnaires. These ques-

tionnaires are the source of data for the hypotheses being tested in the
As mentioned previously, the present study augments past research present article (see Section 5.3).
in two ways: First, past studies have not examined hope's ability to
account for variance in GPA simultaneously with optimism and self- 5.3. Measures
efficacy. Given that these are related constructs, but nonetheless em-
phasize different aspects of positive expectancy, such an analysis 5.3.1. Adult Hope Scale (AHS)
would be useful. Second, most past work has documented relatively The AHS (Snyder et al., 1991) is a measure of hope for goals in
weak relationships between GPA and hope, possibly because hope has general, rather than for a specific goal or domain. It contains 4 items
been measured in a general way rather than a way specific to academic tapping pathways thinking, 4 tapping agency thinking, and 4 distracters.
goals. Thus, we utilize both the AHS and academic subscale of the DSHS Respondents rate each item on a 1 (definitely false) to 8 (definitely true)
to measure hope. We similarly measure self-efficacy at both the general scale. Sample pathways and agency items, respectively, are “There are
level (using the GSES) as well as the academic-specific level (using the lots of ways around any problem,” and “I energetically pursue my
ASES). Because optimism has been conceptualized primarily at the goals.” Researchers have provided evidence supporting the reliability
general level, only a generalized measure was used (the LOT-R). and validity of the AHS (Snyder et al., 1991). In the present sample,
Using path analysis, we test a model with the following characteris- the AHS had a Cronbach's alpha of .79.
tics: Academic-specific expectancy variables (academic hope, academic
self-efficacy) were hypothesized to predict GPA directly, whereas gen- 5.3.2. Domain-Specific Hope Scale (DSHS)
eralized expectancy variables (general hope, general self-efficacy) The DSHS (Sympson, 1999) is a measure of hope in different life
were hypothesized to predict their academic-specific counterparts. domains. This 48-item measure contains 6 domain scales: social, aca-
General optimism was hypothesized to directly predict GPA. Though demic, family-home, romantic, work, and leisure (Snyder, Cheavens, &
expectancy variables at each level (general and academic-specific) Sympson, 1997). Only the 8-item scale for the academic domain
were expected to correlate with one another, it was expected that the was used in the present study. Respondents rate each item on a 1
best predictor of GPA would be academic hope. (definitely false) to 8 (definitely true) scale. Sample academic pathways
and agency items, respectively, are “There are lots of ways to meet the
challenges of any class,” and “I energetically pursue my school work.”
5. Method Researchers have provided evidence supporting the reliability and
validity of the DSHS (Sympson, 1999; see Magyar-Moe, 2009). In the
5.1. Participants present sample, the academic hope scale had a Cronbach's alpha of .85.

Participants were 89 college students (27 males, 62 females) from a 5.3.3. General Self-Efficacy Scale (GSES)
Northern California university who took part in the study as one means The GSES (Scholz et al., 2002; Schwarzer & Jerusalem, 1995) is a
of fulfilling the requirements of their introductory psychology courses. measure of general self-efficacy. It consists of 10 items. Respondents
The mean age was 19.35 (SD = 2.25). Most participants were freshmen rate each item on a 1 (not true at all) to 4 (exactly true) scale. Sample
(62%), with 30% sophomores, 6% juniors, and 2% seniors. The sample items include “I am confident that I could deal efficiently with unexpected
was primarily Caucasian (60%), with fewer participants identifying as events,” and “I can usually handle whatever comes my way.” Researchers
Asian or Asian American (12%), Latino (9%), African American (2%), have provided evidence supporting the reliability and validity of the GSES
mixed ethnicity (12%), or “other” (5%). (Scholz et al., 2002). In the present sample, the GSES had a Cronbach's
alpha of .82.

5.2. Procedure 5.3.4. Academic Self-Efficacy Scale (ASES)

The ASES (Chemers et al., 2001) is an 8-item measure of self-efficacy
Data were collected as part of a larger study examining the relation- in the domain of academics, reflecting students' confidence in their
ship between hope and non-conscious goals. The full procedure is ability to perform well in school. Respondents rate each item on a 1
reported here, though results related to much of the full study will be (very untrue) to 7 (very true) scale. Sample items include “I am good
reported elsewhere. at research and writing papers,” and “I usually do very well in school
Participants reported to lab rooms one at a time. After signing the and at academic tasks.” Researchers have provided evidence supporting
informed consent form, they completed a scrambled word task in the reliability and validity of the ASES (Chemers et al., 2001). In the
which they made sentences from 30 sets of words. This was an experi- present sample, the ASES had a Cronbach's alpha of .83.
mental manipulation unrelated to the hypotheses in the present article.
For participants in the experimental condition, some of these scrambled 5.3.5. Life Orientation Test-Revised (LOT-R)
sentences contained words related to academics, whereas for partici- The LOT-R (Scheier et al., 1994) is a 10-item scale of optimism, 4
pants in the control condition, these sentences were composed of items of which are “fillers” and not scored. Respondents answer each
neutral words. The manipulation was designed to prime the goal of item on a Likert-type scale ranging from 1 (strongly agree) to 5 (strongly
academic achievement in half of participants. To determine whether disagree). Sample items include “In uncertain times, I usually expect the
condition had any effect on study measures, a series of between- best,” and “I hardly ever expect things to go my way” (reverse scored).
subjects one-way ANOVAs were performed with condition serving as Researchers have provided evidence of the reliability and validity of the
the independent variable and each of the study measures serving as LOT-R (Scheier et al., 1994). In the present sample, the LOT-R had a
the dependent variables (i.e., AHS, DSHS, GSES, ASES, LOT-R, and GPA). Cronbach's alpha of .85.
All analyses were non-significant; thus, for the hypotheses being tested
in the present article, there were no effects of condition. 5.3.6. GPA
Next, all participants completed a computer-aided implicit attitude Cumulative grade point average was assessed through self-report.
measure (Ferguson, 2008). Again, this portion of the procedure is unre- On the demographic survey, we asked students, “What is your GPA at
lated to the hypotheses being tested here. Participants viewed a number Santa Clara University?” We did not think it necessary to access
of words and indicated their personal evaluation of each word by pressing students' GPAs through the university registrar because research shows
keys marked “good” or “bad” as quickly as possible. high correlations between self-reported GPA and GPA assessed through
 D.B. Feldman, M. Kubota / Learning and Individual Differences 37 (2015) 210–216 213

student records (Cassady, 2001). In a recent meta-analysis of 29 studies Table 1

including 56,265 college and high school students, Kuncel, Credé, and Correlations among study variables.

Thomas (2005) found a mean sample-size-weighted correlation of .84 Variables 1 2 3 4 5 6 M (SD)

between self-reported and registrar-reported GPA. The mean correlation 1. General hope (AHS) – .57⁎⁎ .67⁎⁎ .55⁎⁎ .49⁎⁎ .32⁎⁎ 50.52 (6.40)
for the 12 studies containing only college students (N = 12,089) was 2. Academic hope (DSHS) – .39⁎⁎ .66⁎⁎ .23⁎ .69⁎⁎ 50.20 (7.41)
even higher, at .90. 3. General Self-Efficacy – .39⁎⁎ .35⁎⁎ .31⁎⁎ 31.36 (3.83)
Scale (GSES)
4. Academic Self-Efficacy – .21⁎ .59⁎⁎ 43.48 (6.14)
5.4. Analyses
Scale (ASES)
5. Optimism (Life – .18⁎ 14.39 (4.66)
Path analysis was conducted using AMOS 21 (Arbuckle, 2012), Orientation Test-Revised)
which enables testing of theoretical causal models based on the 6. GPA – 3.24 (.47)
observed pattern of relationships among variables. Consistent with Note.
accepted path-analytic practices (Byrne, 2011), our model contains ⁎⁎ p b .01.
two types of hypothesized paths: those symbolized by single-headed ⁎ p b .05.

arrows on the path diagram, and those symbolized by two-headed ar-

rows (see Fig. 1). We will use the terminology of “prediction” to refer (.32) was significantly smaller than the correlation between academic
to the single-headed-arrow paths and the terminology of “correlation” hope and GPA (.69), Z = 4.42, p b .01; as were the correlations between
to refer to the double-headed-arrow paths. However, it should be general self-efficacy and GPA (.31), Z = 5.12, p b .01, and between
noted that the design of the study is cross-sectional; these terms are optimism and GPA (.18), Z = 5.36, p b .01.
used as a convention to refer to the types of modeled paths and do
not imply a prospective study. We address this further in Section 7. 6.2. Hypothesized model
We used three goodness-of-fit statistics to evaluate the hypothe-
sized model: chi-square, root mean square error of approximation Using path analysis, we tested a model with the following character-
(RMSEA), and comparative fit index (CFI). Chi-square measures the ab- istics (see Fig. 1): Academic-specific expectancy variables (academic
solute fit between the hypothesized model and the observed pattern of hope, academic self-efficacy) were given direct paths to GPA, whereas
relationships. Acceptable fit is indicated by a non-significant chi-square, generalized expectancy variables (general hope, general self-efficacy)
which denotes no difference between the hypothesized and observed were given paths to each of their academic-specific counterpart
patterns of relationships. RMSEA adjusts the measure of absolute fit ac- variables. General optimism was given a direct path to GPA, because op-
cording to the complexity of the hypothesized model, with simpler timism is by definition a generalized expectancy. Consistent with com-
models being favored. Lower values of RMSEA indicate better fit, with mon path analytic practices, all exogenous variables were allowed to
values of .06 or less representing acceptable fit (Hu & Bentler, 1999). correlate.
Finally, the CFI is an index of incremental fit comparing the hypothe- The hypothesized model displayed inadequate fit to the data, χ2
sized model to a null model in which all variables are uncorrelated. (6) = 26.29 (p b .001), RMSEA = .20, CFI = .90. In Amos, modification
Higher values of CFI indicate better fit, with values of .95 or greater indices can be requested, suggesting the addition of paths to improve fit.
representing acceptable fit (Hu & Bentler, 1999). One path was suggested from general hope to academic self-efficacy,
leading to the following fit: χ2 (5) = 7.90 (p = .16), RMSEA = .08,
6. Results CFI = .99. We conducted a nested model chi-square test, which showed
a significant decrease in chi-square over the hypothesized model, indi-
6.1. Descriptive statistics and zero-order correlations cating improved fit, χ2diff (1) = 18.39, p b .001. Nonetheless, the value
of RMSEA was still inadequate, though modification indices did not
All study variables were correlated with one another. Means, stan- suggest additional paths.
dard deviations, and correlations are presented in Table 1. The variables RMSEA favors simpler models, so eliminating unnecessary paths can
most strongly correlated with GPA were academic hope (.69) and aca- increase this fit index. In Amos, the “specification search” procedure can
demic self-efficacy (.59), both of which are domain-specific expectan- be used to estimate fit indices if paths are fixed at zero (i.e., dropped
cies. A Steiger's Z-test (Meng, Rosenthal, & Rubin, 1992) indicated that from the model). Because two paths displayed non-significant beta
these two relatively large correlations are not significantly different weights near zero (.01 from general optimism to GPA, and .03 from
from one another, Z = 1.52, p N .05. The correlations between the gen- general self-efficacy to academic self-efficacy), we requested fit indices
eralized expectancy variables and GPA, however, were somewhat if either or both were removed. The best fit was achieved by removing
smaller. Specifically, the correlation between general hope and GPA both of these paths.

Fig. 1. Hypothesized path model. Single-headed arrows represent “causal” paths. Double-headed arrows represent correlations. Circles represent error variances.
214 D.B. Feldman, M. Kubota / Learning and Individual Differences 37 (2015) 210–216

Fig. 2. Final modified path model. Single-headed arrows represent “causal” paths, for which standardized regression coefficients (β) are presented. Double-headed arrows represent
correlations. Circles represent standardized error variances. All coefficients are significant at p b .05.

The modified model showed very good fit to the data, χ2 (7) = 8.00 As mentioned earlier, there has been a call in the literature (Snyder,
(p = .33), RMSEA = .04, CFI = .995 (see Fig. 2). We again conducted a 1995, 2002) to clarify relationships and differences among hope, self-
nested model chi-square test, which showed no significant difference in efficacy, and optimism. The current study has shed further light on
chi-square values between the current model and the previous one, this issue, particularly with reference to GPA. Overall, we found hope
χ2diff (2) = .10, p = .95. Lack of significant difference is generally to be the most consistent predictor of GPA across the generalized and
taken to indicate that the more parsimonious model should be favored domain-specific levels. That is, general hope predicted GPA through
over the more complex one (Schermelleh-Engel, Moosebrugger, & academic-specific hope.
Müller, 2003). The current model is the more parsimonious one One possible explanation for why optimism may not have predicted
(i.e., fewer estimated paths). GPA pertains to its generalized nature. Recall that we hypothesized that
All paths in this model were significant. Generalized hope predicted domain-specific expectancies, in contrast to general expectancies,
academic-specific hope (β = .57, p b .001) and academic self-efficacy should be better predictors of domain-specific outcomes like grades.
(β = .55, p b .001). In turn, academic-specific hope predicted GPA Optimism was the only construct in the study for which an academic-
(β = .54, p b .001), as did academic self-efficacy (β = .23, p = .02). specific measure was not used, given the generally accepted definition
The combination of academic-specific hope and academic-specific of optimism as a generalized expectancy (Scheier et al., 2001). Addition-
self-efficacy accounted for 51% of the variance in GPA. The total indirect ally, optimism is an expectancy of positive outcomes irrespective of
effect of general hope on GPA, through academic-specific hope and aca- one's actions. This fact has led some theorists to suggest that, in highly
demic self-efficacy, was .43.1 As mentioned, general self-efficacy failed behaviorally controllable situations such as college, hope may be a
to predict academic-specific self-efficacy, and thus this path was elimi- stronger predictor of performance than optimism (Rand et al., 2011).
nated from the model. Also as mentioned, general optimism failed to Next, we turn our discussion to self-efficacy. Because the Academic
predict GPA, and thus this path was eliminated. Self-Efficacy Scale (ASES; Chemers et al., 2001) assesses expectancies
To test whether the beta weight of the path between academic- regarding specific behaviors in the domain of academics, it makes
specific hope and GPA was significantly different than the beta weight sense that it predicted GPA. It is curious, however, that general self-
between academic self-efficacy and GPA, we performed a nested efficacy did not predict academic self-efficacy in the path model. In-
model chi-square test in which our final modified model was compared stead, general hope predicted academic self-efficacy, which in turn pre-
to an identical model where these two paths were constrained to be dicted GPA. This may, of course, be a measurement artifact of the two
equal. The resulting model also had good fit to the data, χ2 (8) = 9.57 scales used to measure general and academic-specific self-efficacy
(p b .30), RMSEA = .05, CFI = .99. There was no significant difference (GSES and ASES). These particular scales rarely have been used in
between the chi-square values of this model and the one in which studies together. However, it is worth noting that some self-efficacy
both paths were estimated separately, indicating that these two paths researchers have criticized the construct of generalized self-efficacy on
are not statistically different from one another, χ2diff (1) = 1.57, p = .21. the basis that it is theoretically inconsistent with the notion of self-
efficacy itself. Self-efficacy was originally defined as a behavior- or
7. Discussion domain-specific expectancy (Bandura, 1977). As mentioned previously,
Bandura (1997) has written that self-efficacy should be measured only
The results largely support our hypotheses, showing that general- within a given domain or with reference to a particular behavior. The
ized hope predicted academic hope, and academic hope in turn directly results from this study seem to confirm Bandura's thoughts.
predicted GPA. However, the findings failed to show similar results for Finally, we turn to hope. The results of the present study are consis-
self-efficacy. Namely, there was no significant path between general tent with Rand's (2009) findings showing that general hope predicted
self-efficacy and academic self-efficacy. Instead, general hope was grades in a college course indirectly through specific grade expectan-
found to predict academic self-efficacy, and academic self-efficacy in cies. In this case, general hope did so through domain-specific hope.
turn predicted GPA. Lastly, consistent with past research (Rand et al., Hope concerns the cognitive process of pursuing a goal through path-
2011), optimism was not found to predict GPA. ways (planning-related thoughts) and agency (motivation-related
thoughts). It makes sense that being successful in the academic domain
1
Because research not assessing academic hope has demonstrated a correlation be- requires this combination of planning and motivation. School-related
tween general hope and GPA (Snyder et al., 2002), and because a zero-order correlation tasks such as writing research papers, taking notes, and scheduling
of .32 was found in the present study, we re-ran the final path model with a direct path time require putting plans in motion and motivating oneself.
from general hope to GPA. Although this path was not suggested by modification indices,
it holds theoretical interest. Though this model manifested excellent fit to the data, χ2
Of note, the beta weight of the path between academic self-efficacy
(6) = 4.28 (p = .64), RMSEA = .00, CFI = 1.00, the additional direct path, which had a and GPA was not significantly different from the beta weight of the
beta weight of −.18, did not reach significance. path between academic hope and GPA. As mentioned previously,
 D.B. Feldman, M. Kubota / Learning and Individual Differences 37 (2015) 210–216 215

these two constructs are theoretically somewhat different, with self- References
efficacy emphasizing positive expectancies for the performance of
instrumental behaviors, and hope emphasizing positive expectations Alkharusi, H. (2010). Literature review on achievement goals and classroom goal structure:
Implications for research. Electronic Journal of Research in Educational Psychology, 8,
regarding goal-directed planning and motivation. Nonetheless, what 1363–1386.
both academic self-efficacy and academic hope have in common is Arbuckle, J.L. (2012). Amos (Version 25) [computer program]. Armonk, NY: IBM Corp.
their domain-specific nature. This result again points to the importance Aspinwall, L.G., & Leaf, S.L. (2002). In search of the unique aspects of hope: Pinning our
hopes on positive emotions, future-oriented thinking, hard times, and other people.
of assessment at the domain-specific level. Psychological Inquiry, 13, 276–288.
It is important to note some limitations of the present study. First, Bandura, A. (1977). Self-efficacy: Toward a unifying theory of behavioral change.
the data were drawn from a larger study (detailed earlier). To help Psychological Reports, 84, 191–215.
Bandura, A. (1997). Self-Efficacy: The exercise of control. New York: Freeman.
evaluate whether the experimental manipulation of that larger study af- Bryant, F.B., & Cvengros, J.A. (2004). Distinguishing hope and optimism: Two sides of a
fected the measures reported here, we tested for between-group differ- coin, or two separate coins? Journal of Social and Clinical Psychology, 23, 273–302.
ences, and no significant differences were noted. Thus, the manipulation Buckelew, S.P., Crittendon, R.S., Butkovic, J.D., Price, K.B., & Hurst, M. (2008). Hope as a
predictor of academic performance. Psychological Reports, 103, 411–414.
did not differentially affect participants' answers to the present study Byrne, B.M. (2011). Structural equation modeling with MPLUS: Basic concepts, applications,
measures. Nonetheless, it is impossible to know whether all participants and programming. New York: Routledge.
were affected by the mere fact that they were completing other tasks Carver, C.S., & Scheier, M.F. (2002). The hopeful optimist. Psychological Inquiry, 13,
288–290.
and measures.
Cassady, J.C. (2001). Self-reported CPA and SAT: A methodological note. Practical
Secondly, the present study was cross-sectional, limiting our ability Assessment, Research & Evaluation, 7 (Retrieved July 12, 2013 from http://
to determine causality or temporality. Although we utilized path analy- PAREonline.net/getvn.asp?v=7&n=12).
sis, a procedure that allows researchers to specify causal or predictive Chang, E.C., & DeSimone, S.L. (2001). The influence of hope on appraisals, coping, and dys-
phoria: A test of hope theory. Journal of Social and Clinical Psychology, 20, 117–129.
relationships among modeled variables even when data are collected Cheavens, J.S., Feldman, D.B., Gum, A., Michael, S.T., & Snyder, C.R. (2006). Hope therapy in
cross-sectionally, lack of either a longitudinal or experimental design a community sample: A pilot investigation. Social Indicators Research, 77, 61–78.
limits any strong ability to evaluate hope's causal or prospective rela- Chemers, M.M., Hu, L., & Garcia, B.F. (2001). Academic self-efficacy and first year college
student performance and adjustment. Journal of Educational Psychology, 93, 55–64.
tionship to GPA. Thus, it would be helpful for future studies to test the Curry, L.A., & Snyder, C.R. (2000). Hope takes the field: Mind matters in athletic perfor-
present model using causal or prospective designs. mances. In C.R. Snyder (Ed.), Handbook of hope: Theory, measures, and applications
A third limitation of this study concerns its sample, which consists of (pp. 243–260). San Diego, CA: Academic Press.
Curry, L.A., Snyder, C.R., Cook, D.L., Ruby, B.C., & Rehm, M. (1997). The role of hope in
undergraduates from a private Northern California university. It is un- student-athlete academic and sport achievement. Journal of Personality and Social
known to what degree these results would generalize to other groups Psychology, 73, 1257–1267.
such as graduate students or public university students. On a related Davidson, O.B., Feldman, D.B., & Margalit, M. (2012). A focused intervention for first-year
college students: Promoting hope, sense of coherence and self-efficacy. Journal of
note, the original hypothesized path model was modified based on fit
Psychology, 146, 333–352.
indices. Although this is a common practice, whenever a model is ad- Feldman, D.B., & Dreher, D.E. (2012). Can hope be changed in 90 minutes? Testing the ef-
justed post-hoc, there is a risk of modifying the model based on random ficacy of a single-session goal-pursuit intervention for college students. Journal of
Happiness Studies, 13, 745–759.
error or the characteristics of the particular sample. As such, future
Feldman, D.B., Rand, K.L., & Kahle-Wrobleski, K. (2009). Hope and goal attainment:
research should test our revised model in additional samples. It also Testing a basic prediction of hope theory. Journal of Social and Clinical Psychology,
may be useful to replicate the present study in samples of primary and 28, 479–497.
secondary school students. Past research (Gilman, Dooley, & Florell, Feldman, D.B., & Snyder, C.R. (2005). Hope and the meaningful life: Theoretical and
empirical associations between goal-directed thinking and life meaning. Journal of
2006) has examined the relationship between hope and GPA among Social and Clinical Psychology, 24, 401–421.
adolescents but has not modeled the relative contributions of hope, Ferguson, M.J. (2008). On becoming ready to pursue a goal you don't know you have:
self-efficacy, and optimism simultaneously. Effects of nonconscious goals on evaluative readiness. Journal of Personality and
Social Psychology, 95, 1268–1294.
A last limitation is that we obtained GPA via self-report, a fact that Gallagher, M.W., & Lopez, S.J. (2009). Positive expectancies and mental health: Identifying
potentially introduces social desirability bias. As mentioned previously, the unique contributions of hope and optimism. The Journal of Positive Psychology,
we chose not to access students' records because research shows high 4(6), 548–556.
Gilman, R., Dooley, J., & Florell, D. (2006). Relative levels of hope and their relationship
correlations between self-reported GPA and registrar-reported GPA with academic and psychological indicators among adolescents. Journal of Social
(Cassady, 2001; Kuncel et al., 2005). It should be noted, however, that and Clinical Psychology, 25(2), 166–178.
the meta-analysis conducted by Kuncel et al. (2005) found that self- Green, L.S., Oades, L.G., & Grant, A.M. (2006). Cognitive–behavioral, solution-focused life
coaching: Enhancing goal striving, well-being, and hope. The Journal of Positive
reported GPA accurately reflects registrar GPA in most cases, but not Psychology, 1, 142–149.
all. Therefore, future studies should obtain GPA directly, if possible. Hu, L., & Bentler, P.M. (1999). Cutoff criteria for fit indexes in covariance structure analy-
Even given these limitations, we believe that the present study offers sis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6,
1–55.
useful evidence of the relationships among hope, self-efficacy, opti-
Irving, L.M., Seidner, A.L., Burling, T.A., Pagliarini, R., & Robbins-Sisco, D. (1998). Hope and
mism, and academic achievement. It also highlights the importance of recovery from substance dependence in homeless veterans. Journal of Social and
measuring expectancy variables at domain-specific levels in addition Clinical Psychology, 17, 389–406.
to the more usual generalized level. Irving, L.M., Snyder, C.R., & Crowson, J.J., Jr. (1998). Hope and the negotiation of cancer
facts by college women. Journal of Personality, 66, 195–214.
Given the relatively strong relationship between hope and GPA, it Klausner, E.J., Clarkin, J.F., Spielman, L., Pupo, C., Abrams, R., & Alexopoulos, G.S. (1998).
may be beneficial to explore interventions to increase academic hope. Late-life depression and functional disability: The role of goal-focused group psycho-
Hope-based interventions have been shown to affect depressed mood therapy. International Journal of Geriatric Psychiatry, 13, 707–716.
Kuncel, N.R., Credé, M., & Thomas, L.L. (2005). The validity of self-reported grade point av-
and anxiety (Cheavens, Feldman, Gum, Michael, & Snyder, 2006; erages, class ranks, and test scores: A meta-analysis and review of the literature.
Klausner et al., 1998), life satisfaction (Green, Oades, & Grant, 2006), Review of Educational Research, 75, 63–82.
and goal achievement (Feldman & Dreher, 2012). But these interven- Lackaye, T.D., & Margalit, M. (2006). Comparisons of achievement, effort, and self-
perceptions among students with learning disabilities and their peers from different
tions teach skills related to hope in general, rather than with reference achievement groups. Journal of Learning Disabilities, 39, 432–446.
to a particular domain such as academics. Given that the present study Levi, U., Einav, M., Ziv, O., Raskind, I., & Margalit, M. (2013). Academic expectations and
highlights the importance of conceptualizing hope at the academic- actual achievements: The roles of hope and effort. European Journal of Psychology of
Education, http://dx.doi.org/10.1007/s10212-013-0203-4.
specific level for predicting GPA, interventions could be created to foster Magalleta, P.R., & Oliver, J.M. (1999). The hope construct, will, and ways: Their relations
the development of academic hope in particular. with self-efficacy, optimism, and general well-being. Journal of Clinical Psychology,
With continued work and a better understanding of how hope 55, 539–551.
Magyar-Moe, J.L. (2009). Therapist's guide to positive psychological interventions. New
impacts GPA, perhaps new tools could be developed to help improve
York: Academic Press.
students' school performance and, ultimately, maximize their future Meng, X., Rosenthal, R., & Rubin, D.B. (1992). Comparing correlated correlation coefficients.
opportunities. Psychological Bulletin, 111, 172–175.
216 D.B. Feldman, M. Kubota / Learning and Individual Differences 37 (2015) 210–216

Piper, W. (1978). The little engine that could. New York: Grosset & Dunlap. Snyder, C.R. (2002). Hope theory: Rainbows in the mind. Psychological Inquiry, 13,
Rand, K.L. (2009). Hope and optimism: Latent structures and influences on grade expec- 249–275.
tancy and academic performance. Journal of Personality, 77, 231–260. Snyder, C.R., Berg, C., Woodward, J.T., Gum, A., Rand, K.L., Wrobleski, K.L., et al. (2005).
Rand, K.L., Martin, A.D., & Shea, A.M. (2011). Hope, but not optimism, predicts academic Hope against the cold: Individual differences in trait hope and acute pain tolerance
performance of law students beyond previous academic achievement. Journal of on the cold pressor task. Journal of Personality, 73, 287–312.
Research in Personality, 45, 683–686. Snyder, C.R., Cheavens, J., & Sympson, S.C. (1997). Hope: An individual motive for social
Scheier, M.F., & Carver, C.S. (1985). Optimism, coping, and health: Assessment and commerce. Group Dynamics: Theory, Research and Practice, 1, 107–118.
implication of generalized outcome expectancies. Health Psychology, 4, 219–247. Snyder, C.R., Harris, C., Anderson, J.R., Holleran, S.A., Irving, L.M., Sigmon, S.T., et al.
Scheier, M.F., Carver, C.S., & Bridges, M.W. (1994). Distinguishing optimism from neurot- (1991). The will and the ways: Development and validation of an individual-
icism (and trait anxiety, self-mastery, and self-esteem): A reevaluation of the Life differences measure of hope. Journal of Personality and Social Psychology, 60,
Orientation Test. Journal of Personality and Social Psychology, 67, 1063–1078. 570–585.
Scheier, M.F., Carver, C.S., & Bridges, M.W. (2001). Optimism, pessimism, and psychological Snyder, C.R., LaPointe, A.B., Crowson, J.J., Jr., & Early, S. (1998). Preference of high- and
well-being. In E.C. Chang (Ed.), Optimism & pessimism: Implications for theory, research, low-hope people for self-referential input. Cognition and Emotion, 12, 807–823.
and practice (pp. 189–216). Washington, D.C.: American Psychological Association. Snyder, C.R., Michael, S.T., & Cheavens, J.S. (1999). Hope as a psychotherapeutic founda-
Schermelleh-Engel, K., Moosebrugger, H., & Müller, H. (2003). Evaluating the fit of tion of nonspecific factors, placebos, and expectancies. In M.A. Huble, B. Duncan, &
structural equation models: Tests of significance and descriptive goodness-of-fit S. Miller (Eds.), Heart and soul of change (pp. 205–230). Washington, D.C.: APA.
measures. Methods of Psychological Research, 8, 23–74. Snyder, C.R., Shorey, H.S., Cheavens, J., Pulvers, K.M., Adams, V.H., III, & Wiklund, C. (2002).
Scholz, U., Doña, B.G., Sud, S., & Schwarzer, R. (2002). Is general self-efficacy a universal Hope and academic success in college. Journal of Educational Psychology, 94, 820–826.
construct? Psychometric findings from 25 countries. European Journal of Psychological Sympson, S. (1999). Validation of the Domain Specific Hope Scale: Exploring hope in life
Assessment, 18, 242–251. domains. Unpublished doctoral dissertation, University of Kansas, Lawrence.
Schwarzer, R., & Jerusalem, M. (1995). Generalized self-efficacy scale. In J. Weinman, S. Sympson, S. (2000). Rediscovering hope: Understanding and working with survivors of
Wright, & M. Johnston (Eds.), Measures in health psychology: A user's portfolio trauma. In C.R. Snyder (Ed.), Handbook of hope: Theory, measures, and applications
(pp. 35–37). Windsor, UK: Nfer-Nelson. (pp. 285–300). San Diego, CA: Academic Press.
Segerstrom, S.C., & Nes, L.S. (2006). When goals conflict but people prosper: The case of Tennen, H., & Affleck, G. (1999). Finding benefits in adversity. In C.R. Snyder (Ed.), Coping:
dispositional optimism. Journal of Research in Personality, 40, 675–693. The psychology of what works (pp. 279–304). New York: Oxford Press.
Snyder, C.R. (1994). The psychology of hope. New York: Free Press. Tennen, H., Affleck, G., & Tennen, R. (2002). Clipped feathers: The theory and measurement
Snyder, C.R. (1995). Conceptualizing, measuring, and nurturing hope. Journal of Counseling of hope. Psychological Inquiry, 13, 311–317.
and Development, 73(3), 355–360.