Journal of College & University Student Housing Volume 48, No. 1 - 53

the students are nested within different contexts-we
did not utilize hierarchical linear
modeling techniques for analyses for several
reasons. We computed the intraclass
correlation coefficients, an estimate of the
proportion of between-institution variance
compared to within-institution variance, and
discovered that the coefficients were less
than .001, suggesting greater independence
of observations in the different groups of institutions.
Scholars utilizing the MSL survey
in prior studies have similarly discovered
nominal between-institution differences in
their results (Dugan et al., 2013), suggesting
that ordinary least squares regressions are
sufficient analyses for the present project.
Finally, we analyzed the data using ordinary
least squares regression. We examined the
relationship between our independent variable
(participating as a resident assistant, yes
or no) and our dependent variable (leadership
efficacy).
RESULTS
After conducting the propensity score
matching analysis, we examined whether the
matching procedures balanced the distribution
of variables in both the treatment and
control groups by first reviewing the standardized
mean differences (the mean differences
between the two groups divided by
the standard deviation of the control group)
in the treatment and control groups before
and after matching. We detected no large imbalances
above .25 after matching, meeting
the threshold suggested by Rosenbaum and
Rubin (1985). We also examined the overall
imbalance test (Hansen & Bowers, 2008)
and found that no variables were significantly
unbalanced after matching. Additionally,
the measure developed by Iacus et al. (2009)
was smaller in the matched sample than in
the unmatched sample. Our visual inspections
of histograms of propensity scores
pre- and post-matching suggested that the
magnitude of standardized differences was
reduced, and histograms of standardized differences
of all terms pre- and post-matching
suggested that the standardized differences
post-matching were centered on zero and
that no systematic differences existed after
matching (Thoemmes, 2012). Therefore, the
covariates within the treatment and control
groups differed significantly before matching
procedures were implemented, and the
propensity score matching decreased bias by
making the observed and treatment groups
more similar with regard to their covariates.
After creating matched pairs of students,
we examined the differences in students' leadership
efficacy based upon whether they had
served as a resident assistant (yes or no). The
results of our analyses suggest that RAs had
significantly higher leadership efficacy than
did their matched group of peers who did not
participate as resident assistants (β = .265,
p < .001). Participation as a resident assistant
explained 7.0% of the variance in students'
leadership efficacy.
DISCUSSION
We found that resident assistants had significantly
higher leadership efficacy than
did a matched group of peers who did not
serve as resident assistants. In other words,
compared to their peers, RAs had significantly
higher confidence in their ability to
lead others, organize a group's tasks to accomplish
goals, take initiative to improve
something, and work with a team on a group
project. While we hypothesize that some of
the effects could be attributed to the intensive
mentorship with student affairs professionals
that often takes place in RA positions
Volume 48, No. 1 * 2021
53
Journal of College & University Student Housing Volume 48, No. 1

Table of Contents for the Digital Edition of Journal of College & University Student Housing Volume 48, No. 1
