learning) on CSSES scores. That is, given the wealth
of research indicating that self-efficacy improves over
the course of relevant interventions, a pre- to postsemester increase in CSSES scores was hypothesized
for the service-learning students but not for students
who did not participate in service-learning.
Participants
Participants included 125 undergraduate students
(37 males, 89 females) who enrolled in Abnormal
Psychology during the 1996-1997 academic year.
The sample (age M = 20 years-6 months, SD = 3
years-5 months) consisted of 111 (89%) EuropeanAmericans, 6 (4.5%) African-Americans, 2 (1.5%)
Hispanic-Americans, 1 (less than 1%) Asian
American, and 5 (4%) students who did not disclose
their race/ethnicity. Regarding religious affiliation,
the sample included 86 (69%) Catholic students, 19
(15%) Protestant students, 2 (2%) Jewish students,
and 18 (14%) individuals reporting no religious
affiliation. As in Study 1, the demographic characteristics of this sample are representative of the population from which it was derived, as confirmed by
a set of one-sample t-tests.
Procedure
In an ongoing service-learning project in an
Abnormal Psychology course (Reeb, Isackson &
Sammon, 1997; Reeb, Isackson, Sammon &
Karban, 1997), the syllabus indicates an opportunity
to sign up for an additional service-learning credit.
Among 125 students who enrolled in Abnormal
Psychology during the 1996-1997 academic year, 32
pursued the service-learning opportunity, whereas
93 did not pursue the opportunity. Relative to those
who did not pursue service-learning, the servicelearning students tended to have slightly higher
grade point averages (M = 3.3, SD =.43 vs. M =
3.1, SD =.50), t(120) = 2.0, p =.05), and a slightly
higher number of past service experiences (M =
2.19, SD = 1.31 vs. M 1.63, SD = 1.30), t(123) =
2.10, p =.04. Those who pursued the opportunity
completed a Service-Learning Contract, which indicates an agreement to (a) participate in a series of
orientation meetings, (b) work approximately 3
hours per week at the service agency, (c) respond to
reflection questions via e-mail and group meetings,
and (d) write a term paper illustrating concepts from
Abnormal Psychology through experiential work.
The CSSES was administered to all students at preand post-semester.
Results and Discussion
Inter-Item Consistency. Study 2 replicated study 1
in demonstrating inter-item consistency for the
CSSES (coefficient alpha =.94). Consistent with
The Community Service Self-Efficacy Scale
study 1, none of the specific items detracted from
level of inter-item consistency (alpha coefficient),
and item-total correlation coefficients ranged from.68 to.79.
Test-Retest Reliability. For the students who did
not pursue service-learning, the pre- to post-semester changes in CSSES scores were nonsignificant (M
= 7.8, SD = 1.5 vs. M = 7.6, SD = 1.7), t(87) =
1.24, p =.22, and the test-retest reliability coefficient was high in magnitude and statistically significant (r =.62, p =.001). Since it was expected that
service-learning may influence CSSES scores,
scores for students who pursued service-learning
were not included in the computation of test-retest
reliability.
Criterion-Related Validity. Consistent with the
hypothesis derived from self-efficacy theory, the
pre-semester CSSES scores of service-learning students (M = 8.4, SD = 1.3) were significantly higher
than the corresponding scores of students who did
not pursue service-learning (M = 7.6, SD = 1.7),
t(123) = 2.50, p =.007, providing evidence of criterion-related validity for this instrument. Eyler, Giles,
& Braxton (1997) obtained a similar finding with a
different measure of personal efficacy.
Construct Validity. There was an attempt to examine construct validity by determining the effect of an
intervention (service-learning) on CSSES scores.
Among students who chose to pursue the servicelearning opportunity, CSSES scores were very high
at pre-semester; this might have produced a ceiling
effect, precluding an optimal test of the hypothesis
that service-learning increases students' self-efficacy for future service. In general, the change from
pre-semester (M = 8.4, SD = 1.3) to post-semester
(M = 8.2, SD = 1.3) in CSSES scores for the service-learning group was not statistically significant,
t(31) = -.80, p =.43, suggesting that a high sense of
self-efficacy was maintained over the course of the
semester. More specifically, there was evidence that
service-learning students maintained a high level of
self-efficacy for making meaningful contributions
(item 1), promoting social justice (items 3 and 7),
helping others in need (items 5 and 9), working
effectively with community professionals (item 6),
applying knowledge to "real-life" community problems (item 8), and committing to future service
(items 2 and 10).
However, on an item assessing one's general
power to make a major difference in the community
(item 4), service-learning students reported a very
high level at pre-semester (M = 8.8, SD = 1.8), but
then showed a small (but statistically significant)
decrease over the course of the semester (M = 8.1,
SD = 1.7), t(31) = -2.17, p =.04). Miller (1997)
obtained the same finding with a similar item, i.e., "I
53
