Creative Education
2012. Vol.3, No.1, 145-148
Published Online February 2012 in SciRes (
Copyright © 2012 SciRes. 145
Assessing Clinical and Academic Performance in a Master’s
Level Speech Language Pathology Program: A Path Analysis
Margaret M. Kjelgaard1, Anthony J. Guarino2
1Department of Communication Sciences and Disorders, MGH Institute of Health Professions, Boston, USA
2Center for Interprofessional Studies and Innovation, MGH Institute of Health Professions, Boston, USA
Received November 2nd, 2011; revised December 7th, 2011; accepted December 18th, 2011
This study assessed a path model to determine the direct and indirect effects on clinical and academic
success from students’ undergraduate and graduate performance. Astin’s I-E-O model served as the theo-
retical foundation for the hypothesized model. Input (I) data included students’ undergraduate information
(GPA, major, and GRE scores); Environment (E) data included grades in required graduate courses and
formative clinical evaluation; and Output (O) data included PRAXIS score and summative clinical
evaluation. The sample was 122 students who completed a SLP Program at a New England graduate
school. Results of the path analysis supported the efficacy of the proposed model in determining the direct
and indirect effects on professional competence. Graduate students with an undergraduate degree in SLP
were at a disadvantage upon entry to the program relative to students from other undergraduate majors.
Implications of the study are discussed.
Keywords: Graduate Program; Speech Language Pathology; Path Analysis; Learning Outcomes; Astin’s
I-E-O Model
The field of Speech and Language Pathology (SLP) has be-
come increasingly diverse in recent years. The current range of
settings, clinical populations and specialized areas of diagnosis
and treatment are widely varied. In 2011, the American Speech,
Language and Hearing Association (ASHA) website describes
18 independent divisions in the field ( With the
increasing number of special interests, graduate programs are
required to teach theory and practice on a wide variety of clini-
cal issues. Because of the breadth of competence necessary to
practice, many states require a minimum of a master’s degree to
practice. Therefore, a graduate program in SLP must admit and
then graduate highly competent practitioners. The graduate
program evaluated in the study, for example, requires six very
demanding semesters of course-work and at least three clinical
placements for degree completion.
Selecting students for admission is encumbered by the lim-
ited information available the applicant’s admissions file. In
addition to this limited information, it has been argued that, for
example, undergraduate GPA’s suffer from grade inflation since
the 1960’s and are, as a result, compressed at the upper end of
the scale (Dostoevsky & Hartley, 2002). Letters of recommen-
dation likewise, tend to be inflated (Rosovsky & Hartley, 2002).
It has also been shown that undergraduate GPA is less accurate
at predicting success in graduate school than standardized GRE
scores (Kuncel & Hezlett, 2007). GRE scores have themselves,
demonstrated mixed validity in predicting graduate student
success (Kuncel & Hezlet, 2007; Sacks, 1999; Sternberg &
Williams, 1997; Zwick, 2002).
Anecdotally, the authors have experienced surprise by the
students who excel as well as those who have difficulty, based
on information from their admissions file. For example, some
students with very high GRE scores and superior undergraduate
GPAs have failed to achieve to expectations, especially with
regard to clinical competence. Conversely, other students with
marginal scores have exceeded faculty expectations by demon-
strating superior clinical competence. The hypothesized model
in this study aimed to identify the direct and indirect effects on
clinical and academic success from students’ undergraduate and
graduate performance.
Prior research on the topic of academic success is rather lim-
ited with regard to SLP programs. One study by Forrest and
Naremore (1998) analyzed admissions data using a discrimi-
nant analysis to categorize students as either “top of the class
students” or “bottom of the class students”. These categories
were defined by scores on board exams and graduate school
GPA. Their results indicated that undergraduate GPA had a
93% overall accuracy rate, yet, GRE failed to achieve any sta-
tistically significant contribution. The limitations of this study
include a small number of students (N = 30) and few predictor
Other researchers (Astin, 1991, 1993; Bean, 1983; Tinto, 1993)
have proposed that understanding success in any academic
program requires assessing the interaction between the student
and the institution’s environment. The present study utilized
Astin’s (1991, 1993) I-E-O model. The premise of the I-E-O
model is that educational assessments must include information
on student inputs (I), the educational environment (E), and
student outcomes (O). Inputs include the students’ demograph-
ics, academic background, and previous experiences; Environ-
ment refers to the range of experiences students encountered
during their college tenure; and Outcomes are the knowledge,
beliefs, and values the student achieved at their institution. The
Inputs are presumed to shape outcomes directly and indirectly
with the institutional environment. Thus, academic success (the
Outcome), is a function of the interaction between the student
characteristics (Inputs) and the characteristics of the academic
setting (Environment).
The primary purpose of this study was to construct and as-
sess a path model to determine the direct and indirect effects on
two Outcome (O) variables (i.e., PRAXIS scores and Summa-
tive Clinical Evaluation). Input (I) data included the student’s
undergraduate information: Undergraduate major (Speech Ma-
jor), Undergraduate GPA (UG GPA), and GRE (GRE Q and
GRE V) while the Environment (E) data were grades in required
graduate courses (Course Average) and Formative Clinical
Evaluation. The hypothesized model is presented in Figure 1.
The sample for this study was 122 students who completed
the SLP Program in the 2007-2009 academic cycles at a New
England graduate school. The sample was comprised of mostly
women ranging in age from 23 to 65, (M = 27.67, SD = 5.37).
The GRE-Q scores ranged from 250 to 790 (M = 587.64, SD =
97.44) while the GRE-V scores ranged from 290 to 800 (M =
516.75, SD = 103.47). Just under half (40.3%) of the partici-
pants were SLP Majors as undergraduates with the remaining
58.9% representing a wide range of majors.
This study analyzed the following students’ Input variables,
1) major (SLP or other); 2) undergraduate GPA; and 3) GRE
scores (verbal and quantitative). The Environment variables
were the Average GPA from required graduate courses and
Formative Clinical Evaluation. The Formative Clinical Evalua-
tion was calculated from scores obtained from the 42 items on
the Clinical Practicum Competency Checklist. This checklist is
composed of eight subscales, 1) Clinical Excellence Across
Contexts; 2) Assessment Data Analysis; 3) Assessment Proce-
dures; 4) Intervention Planning; 5) Intervention Execution; 6)
Clinical Excellence in Writing; 7) Clinical Excellence as a Pro-
fessional; and 8) Self Evaluation. All items are scored on the
following scale: 1 = fails to demonstrate behavior consistently
regardless of amount of supervision or needs excessive and
repetitive instructions, 2 = demonstrates behavior with specific
instruction from supervisor, 3 = demonstrates behavior with
general guidance from supervisor, 4 = demonstrates behavior
with minimal guidance from supervisor, and 5 = demonstrates
behavior independently. The Output variables were PRAXIS
score and Summative Clinical Evaluation. The PRAXIS is de-
signed by experts from The American Speech Language Hear-
ing Association (ASHA), with the goal of testing the skills
necessary to enter professional practice as a new graduate. The
scores on the exam range from 250 - 990 with a passing score of
600 (Educational Testing Service, 2009). The Summative Clinical
Evaluation was comprised of performance scores ranging 1
(poor) to 5 (superior) by the department’s clinical practicum
coordinator. The practicum coordinator oversees all students in
at least three settings outside of the academic setting. This
process involves multiple site visits, counseling, and consulting
between the site supervisors and the students and therefore the
coordinator has a comprehensive understanding of how stu-
dents perform in professional clinical settings. The Summative
Clinical Evaluation was based on the written and verbal as-
sessments of the supervisors in outplacement settings, students’
academic performance, the level of independence exhibited by
the student, as well as commonly-identified characteristics of
good clinicians (e.g., flexibility, proactivity, and critical think-
Figure 1.
Hypothesized model.
Copyright © 2012 SciRes.
The proposed model was comprised of seven structural equa-
tions. In the first two equations, PRAXIS and Summative Clini-
cal Evaluation was regressed on Course Average and Formative
Clinical Evaluation. The next two structural equations had
GRE-Quantitative, GRE-Verbal, and Undergraduate GPA pre-
dict Course Average and Clinical Evaluation. The last three
equations had GRE-Quantitative, GRE-Verbal, and Under-
graduate GPA regressing on Undergraduate SLP Major (coded:
“1” as “Yes”, “0” as “No”). The proposed model illustrates the
linear causal relationships among variables from the three time
points. This allows for understanding the variables from the
admissions file most related to course and clinic performance
during the program and which variables from graduate school
performance are most related to scores at graduation.
Data Analysis
A Hotelling’s MANOVA was conducted to detected differ-
ences between SLP majors and non-SLP majors on 1) under-
graduate GPA; 2) GRE verbal and quantitative scores; 3) gradu-
ate GPA; and 4) formative clinical evaluation. Independent sam-
ples t-tests were conducted as univariate follow-ups. A path
analysis, an application of multiple regression in conjunction
with causal theory, was used to analyze the causal models in
this inquiry.
Prior to analysis, all variables were examined through vari-
ous IBM SPSS programs for accuracy of entry, missing values
and assumptions of normality and linearity resulting that data
met assumptions. Results of the Hotelling’s MANOVA re-
ported statistically significant differences between SLP (Speech
Pathology) majors and non-SLP majors, F(5, 117) = 8.97, p <
.001, η2 = .28. Although SLP majors entered the program with
statistically significantly greater GPAs (M = 3.64, SD = .26)
than non-SLP majors (M = 3.50, SD = .31), F(1, 121) = 6.75, p
= .011, η2 = .05; SLP majors had poorer GRE quantitative
scores (M = 550.80, SD = 103.07) than non-SLP majors (M =
612.88, SD = 86.03), F(1, 121) = 13.13, p < .001, η2 = .10 as
well as GRE verbal scores (M = 467.60, SD = 84.65) relative to
their peers from other undergraduate majors (M = 550.41, SD =
102.89), F(1, 121) = 22.11, p < .001, η2 = .16. These students
who entered with SLP majors continued to be the weaker aca-
demically in the program, having statistically significantly
lower GPA’s at the graduate level, (M = 3.69, SD = .26) than
those entering as non-SLP majors (M = 3.76, SD = .17), F(1,
121) = 4.25, p = .041, η2 = .03.
Figure 2 presents the path coefficients as well as the coeffi-
cient of determination for each of the endogenous variables.
Undergraduate Speech Major was negatively associated with
both GRE Q and GRE V, but was positively related to UG GPA.
All three predictors for Course Average (GRE Q, GRE V, and
UG GPA) achieved significant coefficients, while GRE Q and
UG GPA were moderately correlated to Formative Clinical
Evaluation; GRE V failed to predict Formative Clinical Evalua-
tion. Course Average achieved a statistically significant path
coefficient for PRAXIS; however, Formative Clinical Evalua-
tion did not. Both Course Average and Formative Clinical
Evaluation achieved significant path coefficients for Summa-
tive Clinical Evaluation.
Findings from the present study have implications for future
theoretical work in the area of clinical and academic perform-
ance in graduate SLP programs. The results indicated a better
understanding of success can be derived by assessing the inter-
Figure 1.
Hypothesized model.
Copyright © 2012 SciRes. 147
action between the student and the institution’s environment.
The path analysis identified the direct and indirect effects on
Summative Clinical Evaluation and PRAXIS score. Students
with the highest successful PRAXIS scores were those who
were able to perform in the highly demanding courses in terms
of content and time-management. For success in those courses,
GRE Q and Undergraduate GPA were significant predictors.
The results also seem to suggest that students entering the pro-
gram with an undergraduate degree in SLP were at a disadvan-
tage upon entry to the program. Although SLP majors entered
the program with greater GPAs than non-SLP majors; they had
poorer GRE-Q and GRE-V than non-SLP majors. Furthermore,
these students who entered with SLP majors continued to be
weaker academically in the program. This finding has implica-
tions for not only admissions decisions, but also for under-
graduate programs in SLP. Undergraduate SLP programs need
to better prepare their students for the rigors of graduate school
by offering more research inquiry courses.
The Input (I) data used to select students for a Masters pro-
gram in SLP is related indirectly to academic success in the
program. In consideration of academic aptitude, the GRE-Q,
GRE-V, and UG-GPA demonstrated predictive validity, while
GRE-Q and UG-GPA were predictive of clinical aptitude. The
institution’s courses (Course Average, i.e., E data) prepared
students well for both academic (PRAXIS) and clinical (Sum-
mative Clinical Evaluation, i.e., O data), whereas the clinical
training (Formative Clinical Evaluation) is related to the clini-
cal outcome measure, but not to the PRAXIS scores achieved
by students.
The generalizability of the findings to other institutions is to
be approached cautiously. As indicated by Astin (1993), the
patterns underlying clinical and academic success may vary by
type of institution, the setting, and the composition of the stu-
dent body. Institutions may use this model as a starting point in
investigating success at their respective campuses.
Astin, A. (1991). Assessment for excellence: The philosophy and prac-
tice of assessment and evaluation in higher education. New York:
Astin, A. (1993). What matters in college? Four critical years revisited.
San Francisco: Jossey-Bass.
Bean, J. (1983). The application of a model of turnover in work or-
ganizations to the student attrition process. The Review of Higher
Education, 6, 129-148.
Educational Testing Service (2009). The Praxis Series. Princeton, NJ:
Forrest, K., & Naremore, R. C. (1998). Analysis of criteria for graduate
admissions in speech-language pathology: Predictive utility of appli-
cation materials. American Journal of Speech-Language Pathology,
7, 57-61.
Kuncel, N. R., & Hezlett, S. A. (2007). Standardized tests predict gra-
duate students’ success. Science, 315, 1080-1081.
Rosovsky, H., & Hartley, M. (2002). Evaluation and the Academy: Are
we doing the right thing? Cambridge, MA: American Academy of
Arts and Sciences.
Sacks, P. (1999). Stan dardized minds. Cambridge, MA: Perseus Books.
Sternberg, R. J., & Williams, W. M. (1997). Does the graduate record
examination predict meaningful success in the graduate training of
psychologists? A case study. American Psychologist, 52, 630-641.
Tinto, V. (1993). Leaving college: Rethinking the causes and cures of
student attrition. Chicago: The University of Chicago Press.
Zwick, R. (2002). Fair game? The use of standardized admissions in
higher education. New York: RoutledgeFalmer.
Copyright © 2012 SciRes.