Critical thinking: change during medical school and
relationship to performance in clinical clerkships
Jane N Scott, Ronald J Markert & Margaret M Dunn
Departments of Anatomy, Medicine and Surgery, School of Medicine, Wright State University, Dayton, OH, USA
SUMMARY
The development of critical thinking, the ability to solve
problems by assessing evidence using valid inferences,
abstractions, and generalizations, is one of the global
goals advocated by most medical schools. This study
determined changes in critical thinking skills between
entry and near the end of the third year of medical
school, assessed the predictive ability of a test of critical
thinking skills, and assessed the concurrent validity of
clerkship components and ®nal grade. The Watson±
Glaser Critical Thinking Assessment (WGCTA) was
administered to one class of students at entry to med-
ical school and near the end of year 3. Performance
data for those students who completed their clinical
clerkships on schedule were also recorded. Critical
thinking improved modestly but signi®cantly from en-
try to medical school to near the end of year 3. The
ability of a critical thinking test to predict clerkship
performance was limited; the correlation between
WGCTA total score at entry and the components and
®nal grade of ®ve major clerkships ranged from near 0
to 0á34. The concurrent validity of clerkship compo-
nents and ®nal grade was also limited; correlations with
WGCTA total score near the end of year 3 ranged
between 0á08 and 0á49. The correlation between
WGCTA total score and United States Medical Li-
censing Examination Step 2 was higher at year 3 than at
medical school entry. Critical thinking skills improve
moderately during medical school. Used alone, tests of
critical thinking may be of limited value in predicting
which students will be successful in clinical clerkships.
Clerkship evaluation components and ®nal grade have
limited concurrent validity when a test of critical
thinking is the criterion.
Keywords
*Clinical clerkship; *education measurement; *educa-
tion, medical, undergraduate; *problem solving; *stu-
dents, medical
INTRODUCTION
Critical thinking skills depend on the ability to ask
discriminating questions based on searches for better
ideas or decisions. These skills can be acquired or
enhanced through an active process of learning and
practice (Browne & Keeley 1990). The clinical
component of medical school education may provide
an opportunity to become more skilled in critical
thinking. Students use knowledge from the basic and
clinical sciences to formulate questions, and based on
the answers make a correct diagnosis and treatment
plan.
The Watson±Glaser Critical Thinking Assessment
(WGCTA) has been psychometrically validated and is
content neutral for medical students in that questions
are not related to the basic or clinical sciences of
medicine. For example, some items deal with weather,
basic scienti®c facts, or subjects about which people do
not have strong feelings. Other WGCTA items of a
political, economic, or social nature can evoke strong
feelings and introduce biases into thinking processes
(Watson & Glaser 1980). The WGCTA provides a
method of measuring `raw' or basic critical thinking
skills that are not in¯uenced by a speci®c knowledge
base or training process.
Critical thinking as measured by the WGCTA has
been shown to correlate moderately with student
academic success during the preclinical years. The most
impressive correlations were with courses or testing
instruments designed to integrate or use basic infor-
mation to resolve clinical or research problems (Miller
et al. 1993; Scott & Markert 1994). To extend our
previous observations (Scott & Markert 1994), the aims
of the present study were (1) to assess changes in crit-
ical thinking skills; (2) to investigate the relationship
between critical thinking measured at medical school
entry and clinical clerkship performance; and (3) to
examine the concurrent validity of clerkship evaluation
components using critical thinking near the end of year
3 as the criterion.
Correspondence: Dr Jane N Scott, Department of Anatomy, Wright State University School of Medicine, Dayton, Ohio 45435, USA
14 MEDICAL EDUCATION 1998, 32, 14±18 Ó 1998 Blackwell Science Ltd

METHODS
This study involved Wright State University School of
Medicine class of 1994 students who matriculated for
the 1990±91 academic year. The WGCTA was ad-
ministered at the outset of year 1 and near the end of
year 3. In addition, performance data for those students
who completed their clinical clerkships on schedule
during the 1992±93 academic year were recorded.
Sample size varied for the three analyses of the study:
(1) change in WGCTA scores from entry to near the
end of year 3 …n ˆ 68†; (2) correlations between entry
WGCTA scores and clerkship performance (n = 82);
and (3) correlations between end of year 3 WGCTA
scores and clerkship performance …n ˆ 79†.
There were 41 women and 41 men: 66 whites,
9 African-Americans, 5 Asians, 1 from India, and 1
from mainland Puerto Rico. The WGCTA consists of
®ve subtests that assess the following critical thinking
abilities:
Inference: discriminating between true and false in-
ferences drawn from data.
Recognition of assumptions: recognizing assumptions
or presumptions in statements.
Deduction: deciding if conclusions follow from given
information.
Interpretation: weighing evidence and deciding if
conclusions based on the evidence are warranted.
Evaluation of arguments: distinguishing strong, rele-
vant arguments from weak or irrelevant arguments re-
garding an issue (Watson & Glaser 1980).
A total score and ®ve subtest scores were calculated
for each student.
Evaluation components consisting of `in-house'
clerkship tests, clinical ratings, National Board of
Medical Examination (NBME) subject examinations
and ®nal clerkship grades were recorded for Internal
Medicine, Obstetrics/Gynaecology, Paediatrics, Psy-
chiatry, and Surgery. Clerkship tests were written ex-
aminations prepared by the staff. Clinical ratings were
subjective evaluations of student performance made by
staff members and residents. All clerkships except
paediatrics included the NBME subject examination in
their evaluation components.
The paired T-test was used to compare WGCTA
total score at orientation and year 3. Pearson correla-
tion was used to determine the relationship of the
WGCTA with the clerkship components and United
States Medical Licensing Examination (USMLE) Step
2 scores which assess student understanding of basic
clinical science including health promotion and pre-
vention of disease. Inferences were made at the 0á05
level of signi®cance.
RESULTS
Table 1 shows that the increase in the mean WGCTA
total score for the class (64á4±66á2) was signi®cant
…P ˆ 0
026†. The increases in total scores for women
(65á9±67á0) and men (62á8±65á3) were not signi®cant.
For the subtests only the mean score for evaluation of
arguments improved signi®cantly between entry at year
1 and by the end of year 3. The correlation between
entry and year 3 WGCTA scores was 0á55 (n ˆ 68,
P < 0
001).
Table 2 presents the correlations between clerkship
evaluation components and WGCTA total score at entry
(an evaluation of the predictive ability of the WGCTA)
and near the end of year 3 (an assessment of the con-
current validity of clerkship evaluation components).
Table 1 Changes in class of 1994 mean score on WGCTA from medical school entry to near the end of year 3 clerkships …n ˆ 68†
Entry Year 3
WGCTA Mean
1
SD
2
Mean
1
SD
2
P
Total score …n ˆ 68† 64á4 6á7 66á2 6á8 0á026
Female …n ˆ 35† 65á9 6á8 67á0 7á3 0á259
Male …n ˆ 33† 62á8 6á3 65á3 6á3 0á054
Subject scores
Inference 10á9 2á4 11á4 2á1 0á074
Recognition of assumptions 13á8 2á0 13á8 2á8 0á938
Deductions 13á0 2á0 13á1 2á2 0á881
Interpretation 14á2 1á5 14á3 1á4 0á355
Evaluation of arguments 12á5 3á0 13á5 2á8 0á040
1
Maximum score ˆ 78;
2
SD ˆ standard deviation:
15 MEDICAL EDUCATION 1998, 32, 14±18 Ó 1998 Blackwell Science Ltd
Critical thinking after clerkships J N Scott et al.