Predictors of critical thinking skills of incoming business students.
Whitten, Donna ; Brahmasrene, Tantatape
INTRODUCTION
The promotion of critical thinking ranks among the primary goals
for educators today (Elder, 2004). As reported in a review of literature
by the Office of Outcomes Assessment of the University of Maryland in
2006, critical thinking as an outcome of postsecondary education was
made explicit by several recent national reports (Association of
American Colleges and Universities, 1985; National Education Goals
Panel, 1991; National Institute of Education Study Group, 1984). As
such, the topic of critical thinking is of interest to educators.
Various definitions of critical thinking have been offered. They all
share a common set of meanings. Critical thinking refers to the use of
cognitive skills or strategies and involves solving problems,
formulating inferences, calculating likelihoods, and making decisions.
According to the manual for the California Critical Thinking Skills Test
(CCTST) developed by Peter and Noreen Facione, an important consensus
with regard to the concept of critical thinking was announced in 1990 by
a panel of theoreticians drawn from throughout the United States and
Canada representing several academic fields. These experts characterized
critical thinking as the process of purposeful, self-regulatory judgment
(Facione, 1990). Critical thinking, so defined, is the cognitive engine
which drives problem-solving and decision-making. At the core of
critical thinking are the cognitive skills of reasoning, evaluation,
analysis and inference.
RELATED LITERATURE
Previous studies have used the CCTST to measure critical thinking
(Williams, 2003; Zettergren, 2004; Colucciello, 2005; Yang, 2008).
Scores are included on the following skills: inductive and deductive
reasoning, evaluation, analysis and inference. Inductive reasoning and
deductive reasoning were scored on the CCTST. Induction is usually
described as moving from the specific to the general while deduction
begins with the general and ends with the specific. In the case of a
strong inductive argument it is unlikely or improbable that the
conclusion would actually be false and all the premises true, but it is
logically possible that it might. Arguments based on experience or
observations are expressed inductively since inductive reasoning is
based on making a conclusion based on a set of empirical data. If it is
observed that something is true many times, concluding that it will be
true in all instances is a use of inductive reasoning. Deductive
reasoning allows proof the hypothesis is true. For valid deductive
arguments, it is not logically possible for the conclusion to be false
and all the premises true. For example, deductive reasoning begins with
a general rule, which is known to be true. From that general rule a
conclusion is made about something specific.
The evaluation score on the CCTST measures the results of an
individual's reasoning. The justification of that reasoning in
terms of the evidential, conceptual, methodological, criteriological and
contextual considerations is also measured. Evaluation involves
examining, appraising and judging something carefully. It is the process
of examining a system or system component to determine the extent to
which specified properties are present. Evaluation is the systematic
determination of merit, worth, and significance of something or someone.
The CCTST score on Analysis measures the ability to identify the
intended and actual inferential relationships intended to express
beliefs, judgments, experiences, reasons, information or opinions. This
includes the sub-skills of examining ideas, detecting arguments, and
analyzing arguments into their component elements. Analysis is the
process of breaking a complex topic or substance into smaller parts to
gain a better understanding of it. Perhaps, in its broadest sense, it
might be defined as a process of isolating or working back to what is
more fundamental by means of which something, initially taken as given,
can be explained or reconstructed. This process is a method of studying
the nature of something or of determining its essential features and
their relations. Analysis involves detailed examination of the elements
to understand them, separation of those elements to examine the
individual parts and assessment based on careful consideration of those
elements.
Inference is the process of arriving at some conclusion that,
though it is not logically derivable from the assumed premises,
possesses some degree of probability relative to the premises. The CCTST
score on Inference measures the ability to draw a conclusion or making a
logical judgment based on circumstantial evidence and prior conclusions
rather than on the basis of direct observation. In other words,
inference is the act or process of deriving a conclusion based solely on
what one already knows. It is the act of deriving one idea from another.
Inferences can be valid or invalid and can proceed through either
deductive reasoning or inductive reasoning.
Much research has been conducted on predictors of success in
academia. Burton and Ramist published a report in 2001 on some of the
studies predicting the success of students in college. The conclusion
was that scholastic aptitude test (SAT) scores and high school records
of grade point average (GPA) and high school rank in class were the most
common predictors (Burton & Ramist, 2001).
Previous studies included Class/Year in School (Kealey, Holland
& Watson, 2005; Lampert, 2007). Kealey, Holland and Watson (2005)
observed that class/year in school was not significant in predicting
performance in a course while in Lampert (2007)'s research this
variable was significant in predicting critical thinking scores.
Bridgeman, Burton and Pollack found in 2008 that High School GPA
was significant as a predictor of college GPA. Ventura (2005) determined
that High School Rank was significant as an academic predictor and Baron
determined in 1992 that it was a significant predictor of college
grades. Troutman (1978) indicated that high school rank was a predictor
of performance in freshman mathematics. Math is logic based and
therefore draws on various types of critical thinking.
SAT Scores were significant as academic predictors in studies by
Ventura (2005), Bridgeman, Burton and Pollack (2008) and Baron (1992).
Osana, Lacroix, Tucker, Idan and Jabbour (2007) indicated that verbal
ability is strongly related to syllogistic reasoning, which is
evaluating whether a conclusion necessarily follows from two premises.
This is consistent with findings by Quinlin (1989) which discussed the
relationship between mathematics and reflective thinking and inference.
In addition, in 2007 Cavanagh discovered that students who scored high
on the math portion of the SAT had greater career accomplishments in
fields related to science, technology, engineering and mathematics.
Finally, Stylianides and Stylianides (2008) discuss the link between
deductive reasoning and mathematics. Fields like engineering rely
heavily on the type of processes involved in critical thinking
(Niewoehner, 2008; Ceylon & Lang, 2003).
Gender was included in studies by Ventura in 2005 and Kealey,
Holland and Watson in 2005 and was not significant as an academic
predictor. Race was included as a variable to contribute information not
collected in previous studies. Major was significant as a predictor of
performance in a course by Kealey, Holland and Watson (2005) however,
Lampert ascertained it was not significant in predicting critical
thinking scores.
The idea that critical thinking can and should be taught was not
embraced by all initially, including Glaser (1984) and other skeptics,
because many believed it was a misguided effort. They argued that
thinking skills were context-bound and do not transfer across academic
domains. However, a paper published by Halpern in 1999 noted studies of
successful instruction in critical thinking that included work conducted
by Rubinstein and Firstenberg (1987), Lochhead and Whimby (1987), and
Wood (1987). Successful methods of teaching critical thinking include
practice and teaching critical thinking for transfer from one situation
to another (Van Gelder, 2005; Brahmasrene & Osisek, 2003;
Willingham, 2007).
HYPOTHESIS
The above literature review leads to the hypothesis that the total
critical thinking score and its components such as inductive reasoning,
deductive reasoning, evaluation, analysis, and inference are affected by
college classification (class/year in school), high school GPA, high
school rank, SAT verbal scores, SAT mathematical scores (math), gender,
race and major. These scores are directly proportional to all
independent variables except gender and major. Gender is a dummy
variable where 0 and 1 represent male and female, respectively. This
means the likelihood of being male increases the critical thinking
scores, and vice versa. Major is a dummy variable where 0 represents
business major and 1 for non-business major.
For empirical analysis, the models have been constructed as shown
below:
CT = CONSTANT + [b.sub.1] CLASS + [b.sub.2] HSGPA + [b.sub.3]
HSRANK + [b.sub.4] VERBAL + [b.sub.5] MATH + [b.sub.6] GENDER +
[b.sub.7] RACE + [b.sub.8] MAJOR + [u.sub.i]
CT represents the total critical thinking, evaluation, analysis,
inference, inductive reasoning and deductive reasoning scores.
Description of the variables is summarized in Table 1. [u.sub.i] is a
stochastic error term or disturbance term.
DATA
The California Critical Thinking Skills Test (CCTST) developed by
Insight Assessment was administered to students during 2004-2006
academic years in an introductory accounting course. The dispersed
target group helps eliminate selection bias. The test completion rate
was about 77.5 percent or 483 forms completed out of 623 students,
resulting in 300 usable forms after collecting high school and SAT data.
Table 2 provides descriptive statistics of the total critical thinking,
evaluation, analysis, inference, inductive reasoning and deductive
reasoning scores. These are scale variables where differences between
values are comparable. A mean total critical thinking of 15.24 out of 34
possible points or 44.82 percent suggests the respondents in this study
are performing at a similar level as the group means of 15.89 provided
by Insight Assessment Technical Report number 4, which makes the CCTST
available (Facione, 1990). In addition, a previous study by Williams in
2003 reported means of 15.38 and 16.62. The mean scores of inductive and
deductive reasoning are 8.42 or 49.53 percent and 6.82 or 40.12 percent,
respectively. Evaluation, analysis and inference scores show the
averages of 3.93 or 28.07 percent, 3.92 or 43.55 percent, and 7.39 or
67.18 percent, respectively. Most students in this study are freshmen
and sophomore (averaged 1.8 out of 4 classifications). Their average
high school GPA was 2.96 with 0.584 of high school ranked. This means
participants were right above the upper half of their class. The SAT
verbal scores were 468.96 compared with the 2006 Indiana average of 498
and national average of 503. The SAT mathematical scores were 492.96
compared with the 2006 Indiana average of 509 and national average of
518. Gender, race and major are nominal variables where the variable
values do not have a natural ranking. Their frequencies are reported in
Table 3. About 52.7 percent or 158 out of 300 participants are female
while 47.3 percent (142/300) are male. When asked how they identify
themselves, 82 percent or 246 out of 300 valid cases indicated Anglo
American or Caucasian while 18 percent (54/300) are others. Regarding
major, 56 percent or 168 out of 300 are business majors. The rest, 44
percent (132/300) are non-business majors.
METHODOLOGY
The ordinary least square (OLS) method was employed to test the
above hypotheses. One of the tasks in performing regression analysis
with several independent variables was to calculate a correlation matrix
for all variables. Table 4 reports the Pearson Correlations for all
about herein dependent variables. There were no particularly large
intercorrelations among independent variables except for high school GPA
and high school rank. High school GPA was eliminated to avoid
multicollinearity problem. However, a measure of multicollinearity among
independent variables would be performed.
RESULTS
The assumption of linear multiple regression and the fitness of the
model was tested. According to the computed values of a multiple
regression model, the null hypothesis was rejected at a significant
level of less than 0.01 (F test) in all models as shown in Table 5 and
6. This means that among these estimated equations, there existed a
relationship between critical thinking scores (total, inductive
reasoning, deductive reasoning, evaluation, analysis and inference) and
the explanatory variables: years in college (classification), high
school GPA, high school rank, SAT verbal scores, SAT mathematical
scores, gender, race and major. The coefficient of multiple
determination (R Square) of models in Table 5 varied from 0.22 to 0.41
which are comparable to similar studies. Kealey, Holland and Watson
studied the association between critical thinking and performance in
principles of accounting and reported an adjusted R Square of .31
(Kealey, Holland & Watson, 2005). Lampert used analysis of variance
in a study of critical thinking dispositions as na outcome of
undergraduate education and reported an adjusted R Squared of .05
(Lampert, 2007).
Note that R Square is a measure of goodness of fit. R Square of
zero does not mean that there is no association among the variables. Two
variables, gender and major, had no significant influence on all six
critical thinking models. Therefore, gender and major were omitted. Only
significant variables were included in the final models. The final
results are shown in Table 6. The F test shows significant level of less
than 0.01 in all six models. The variance inflation factor (VIF) is also
presented to detect multicollinearity among independent variables. A
value of VIF less than 10 generally indicates no presence of
multicollinearity. It appears that the observed dependencies did not
affect their coefficients.
Furthermore, significant test (t-test) for all critical thinking
models show the coefficients of independent variables with varying
degrees of significant t-value ([alpha] < 0.01, 0.05 and 0.10), all
with expected positive signs. Therefore, the null hypothesis of years in
college (class), high school rank, SAT verbal, mathematical scores, and
race was rejected as shown in Table 6. Class variable was significant
([alpha] < 0.05) for total scores, deductive reasoning and inference,
but marginally significant ([alpha] < 0.10) for inductive reasoning.
Class had no significant impact on evaluation and analysis scores. High
school rank highly and significantly ([alpha] < 0.01) affected total
and inference scores while significantly ([alpha] < 0.05) affected
inductive, deductive reasoning, evaluation and analysis scores. SAT
verbal scores had a highly significant t-value ([alpha] < 0.01) on
all models except deductive reasoning where it was significant with
t-value ([alpha] < 0.05). However, SAT mathematical scores had highly
significant t-value ([alpha] < 0.01) on the total critical thinking
score, inductive reasoning, deductive reasoning while having significant
t-value ([alpha] < 0.05) on analysis and marginally significant
t-value ([alpha] < 0.10) on evaluation. It had no effect on inference
scores. With respect to the race variable, race had a significant
t-value ([alpha] < 0.05) only on the total critical thinking score
while a marginally significant t-value ([alpha] < 0.10) on inductive
reasoning.
CONTRIBUTIONS
This paper makes three important contributions and supports
previous studies. First, since Class/Year in School was significant for
all measures of critical thinking except Evaluative and Analytical,
different measures of critical thinking may develop over a
student's academic career. It may be that these measures of
critical thinking develop later in students academic experiences.
Therefore studies of students further in their academic careers may be
of interest. High school rank was significant for all measures of
critical thinking. This is consistent with previous studies by Ventura
(2005) and Baron (1992). They included high school rank and determined
that it is positively significant as an academic predictor.
The second important contribution was a result of examining the SAT
math and verbal scores separately. In doing so it revealed that
inferential thinking was predicted by verbal scores but not math scores.
Verbal score on the SAT were significant for all measures of critical
thinking. Students that have high verbal skills may be able to perform
better on critical thinking skills test due to their high verbal
ability. The math score on the SAT was significant for total score,
inductive, deductive, evaluative and analytical reasoning. It would be
interesting to do further research on why math skills are significant
for only these measures of critical thinking and not inferential
thinking.
The third important contribution was the inclusion of race which
was significant for total score and inductive thinking but not for other
measure of critical thinking. It may be due to Caucasian students have
greater access to academic resources such as home computers and the
internet. This may influence performance in terms of total score.
However, further research may need to be conducted to determine reasons
for the significance in terms of inductive reasoning versus other
measures of critical thinking.
PRACTICAL IMPLICATIONS
Measures of critical thinking skills are used for assessment
purposes, including self-assessment for accreditation. In addition,
information on critical thinking skills is useful to designers of
curriculum. While critical thinking is difficult to teach, there is a
need to teach thinking skills at all levels of education. As Carr (1990)
and Willingham (2007) state in their articles on teaching critical
thinking, it should not be taught on its own or by relying on special
courses and text. Instead, every teacher should create an atmosphere
where students are encouraged to read deeply, question, engage in
divergent thinking, look for relationships among ideas, and grapple with
real life issues (Carr, 1990).
A study by Yazici (2004) indicated that collaborative learning
enhances critical thinking skills. Therefore studies that include group
work as a teaching strategy for critical thinking may be valuable.
Ishiyama, McClure, Hart and Amico in 1999 found no significant
difference in critical thinking disposition and evaluation of teaching
strategy lending support to utilizing methods of instruction the enhance
critical thinking skills. Williams (2003) indicated that critical
thinking was the strongest indicator of multiple-choice examination
performance. This has implications for educators as they develop
curriculum and measurement instruments, especially for assessment
purposes. As Peach, Mukherjee, and Hornyak (2007) noted, critical
thinking is recognized as important but difficult to assess. However, it
is an essential component for assessment and institutions must
participate in assessing and developing it.
CONCLUSION
This article provides identification of several important
directions for future research. A focus on different evaluation methods
may reveal of interest. In addition, measuring infusion of critical
thinking would be valuable. Demonstrating critical thinking skills in
the classroom and then observing the effectiveness would be of interest
to educators since it is understood that teaching critical thinking it
is difficult. Finally, research that focuses on conveying the importance
and power of critical thinking to students may determine whether it
generates interest. This may provide information educators can use to
improve their ability to teach critical thinking skills.
REFERENCES
Baron, J. (1992). SAT's, Achievement tests and High-school
rank as Predictors of College Performance. Educational &
Psychological Measurement, 52 (4) 1047-1055.
Brahmasrene , T. & Osisek, P. (2003). Infusing Marginal
Thinking into a Principles of Economics Course. Midwestern Business and
Economic Review, 31, 22-23.
Bridgeman , B., Burton, N. & Pollack, J. (2008). Predicting
Grades in College Courses. Journal of College Admission, Spring, 20-25.
Burton , N. & Ramist, L. (2001). Predicting Success in College:
SAT Studies of Classes Graduating Since 1980. The College Board,
Research Report No. 2001-2, 1-2.
Carr, K. (1990). How can we Teach Critical Thinking? ERIC Digest
adapted from Childhood Education, Winter 1988, 69-73.
Cavabagh, S. (2007). SAT Scores Linked to Adult Success. Education
Week, 27 (4).
Ceylon , T. & Lang, W.L. (2003). Critical Thinking and
Engineering Education. Conference Paper Presented at American Society
for Engineering Education at Valparaiso University, IN. April 4-5, 2003,
41-43.
Colucciello, M. (2005). Relationships Between Critical Thinking
Dispositions and Learning Styles. Journal of Professional Nursing, 15
(5) 294-301.
Elder, L. (2004). The State of Critical Thinking Today: The need
for a Substantive Concept of Critical Thinking. http://
www.criticalthinking.org/resources/articles/the-state-ct-today.shtml.
Facione, P. (1990). Critical Thinking: A Statement of Expert
Consensus for Purposes of Educational Assessment and Instruction. The
Complete American Philosophical Association Delphi Report, The
California Academic Press, ERIC Doc. No.: ED 315 423, 1-3.
Facione, P. (1990). The California Critical Thinking Skills Test:
College Level Technical Report #4--Interpreting the CCTST, Group Norms
and Sub-Scores. The California Academic Press, ERIC Doc. No.: ED
327-566, 4-11.
Glaser, R. (1984). Education and Thinking: The Role of Knowledge.
American Psychologist, 3, 93-104.
Halpern, D. (1999). Teaching for Critical thinking: Helping College
Students Develop the Skills and Dispositions of a Critical Thinker. New
Directions for Teaching and Learning, 80 Winter 1999, 69-70.
Ishiyama, J., McClure, M., Hart, H., & Amico, J. (1999).
Critical Thinking Disposition and Locus of Control as Predictors of
Evaluations of Teaching Strategies. College Student Journal, 33 (2)
272-273.
Kealey, B., Holland, J., & Watson, M. (2005). Preliminary
Evidence on the Association Between Critical Thinking and Performance in
Principles of Accounting. Issues in Accounting Education, 20 (1) 39-44.
Lampert, N. (2007). Critical Thinking Dispositions as an Outcome of
Undergraduate Education. Journal of General Education, 56 (1) 21-28.
Lockhead , J. & Whimby, A. (1987). Teaching Analytic Reasoning
Through Think-Aloud Pair Problem Solving. New Directions for Teaching
and Learning, no. 30. San Francisco: Jossey-Bass.
Niewoehner, R.J. (2008). Applying a Critical Thinking Model for
Engineering Education. The Thinker's Guide to Engineering
Reasoning, United States Naval Academy, 341-344.
Office of Outcomes Assessment University of Maryland University
College. (2006). Critical Thinking as a Core Academic Skill: A Review of
Literature, Spring 1-11.
Osana, H., Lacroix. G., Tucker, B., Idan, E. & Jabbour, G.
(2007), Development of Interplay Between Automatic Processes and
Cognitive Resources in Reading. Journal of Educational Psychology, 99
(4) 898-899.
Peach, B., Mukherjee. A., & Hornyak, M. (2007), Assessing
Critical Thinking: A College's Journey and Lessons Learned. Journal
of Education for Business, 22, (July/Aug) 313-318.
Quinlin, C. (1989). Inferential Thinking for Mathematics Teachers.
The Australian Mathematics Teacher, 45 (4) 2022.
Rubinstein, M. F., & Firstenberg, I. R. (1987), Tools for
Thinking. New Directions for Teaching and Learning, no. 30. San
Francisco: Jossey-Bass.
Stylianides, G., & Stylianides. A (2008), Proof in School
Mathematics: Insights from Psychological Research into Students'
Ability for Deductive Reasoning. Mathematical Thinking and Learning, 10
(2) 103-105.
Troutman, J. (1978). Cognitive Predictors of Final Grades in Finite
Mathematics. Educational & Psychological Measurement, 38 (2)
401-404.
Van Gelder, T. (2005). Teaching Critical Thinking: Some lessons
from Cognitive Science. College Teaching, 53 (1) 42-43.
Ventura, P. (2005). Identifying Predictors of Success for an
Objects-First CS1. Computer Science Education, 15 (3) 227-240.
Willingham, D. (2007). Critical Thinking: Why is it so Hard to
Teach. American Educator, Summer 2007 10-11.
Williams, R. (2003). Critical thinking as a Predictor and Outcome
Measure in a Large Undergraduate Educational Psychology Course.
Conference Presentation, The American Education Research Association,
Chicago Il 7-19.
Williams, R. (2003). Thinking Skills and Work Habits: Contributors
to course Performance. Journal of General Education, 51 (3) 200-201.
Yang, Y. (2008). A Catalyst for Teaching Critical Thinking in a
Large University Class in Taiwan: Asynchronous Online Discussions with
the Facilitation of Teaching Assistants. Educational Technology Research
& Development, 56 (3) 241-264.
Yazici, H. (2004). Student Perceptions of Collaborative Learning in
Operations Management Classes. Journal of Education for Business,
November/December 111-117.
Zettergren, K. (2004). Changes in Critical Thinking Scores: An
Examination of One Group of Physical Therapist Students. Journal of
Physical Therapy Education, Fall 1-2.
Donna Whitten, Purdue University North Central
Tantatape Brahmasrene, Purdue University North Central
Table 1: Description of Variables
Dependent variables
TOTALCT Total score on all 34 questions. Total score on inductive
and deductive reasoning questions. Total score on
evaluation, analysis and inference questions.
INDUCT Inductive Reasoning--Starting with a specific hypothesis
and moving to a general rule by making a conclusion based
on a set of empirical data. An example of inductive
reasoning is scientific confirmation and experimental
disconfirmation.
DEDUCT Deductive Reasoning--Begins with a general rule, which we
know to be true, and ends with the specific conclusion. An
example of deductive reasoning is geometric proofs in
mathematics.
EVAL Evaluation--The systematic determination of merit, worth,
and significance of something or someone.
ANALY Analysis--Analysis involves detailed examination of the
elements to understand them, separation of those elements
to examine the individual parts and assessment based on
careful consideration of those elements.
INFER Inference--The process of deriving a conclusion based
solely on what one already knows.
Independent variables
CLASS College classification, 1-4
HSGPA High school GPA
HSRANK High school rank
VERBAL SAT verbal score
MATH SAT math score
GENDER 0 = male
1 = female
RACE 0 = other
1 = Anglo American, Caucasian
MAJOR 0 = business majors
1 = non-business majors
Table 2: Descriptive Statistics
Std.
N Minimum Maximum Mean Deviation
TOTALCT 300 5 28 15.24 4.399
INDUCT 300 2 15 8.42 2.457
DEDUCT 300 1 14 6.82 2.570
EVAL 300 0 10 3.93 1.957
ANALY 300 1 6 3.92 1.200
INFER 300 1 13 7.39 2.510
CLASS 300 0 4 1.80 .804
HSGPA 300 1.14 4.31 2.96 .800
HSRANK 292 .032 1.000 .584 .234
VERBAL 259 230 760 468.96 83.181
MATH 259 270 760 492.96 86.461
Valid (listwise) 252
Table 3: Frequency
Valid Cumulative
GENDER Frequency Percent Percent Percent
Valid 0 142 47.3 47.3 47.3
1 158 52.7 52.7 100.0
Total 300 100.0 100.0
RACE
Valid .00 54 18.0 18.0 18.0
1.00 246 82.0 82.0 100.0
Total 300 100.0 100.0
MAJOR
Valid .00 168 56.0 56.0 56.0
1.00 132 44.0 44.0 100.0
Total 300 100.0 100.0
Table 4: Correlations
CLASS HSGPA HSRANK
CLASS Pearson Correlation 1 -.089 -.086
Sig. (2-tailed) .123 .144
N 300 300 292
HSGPA Pearson Correlation -.089 1 .590(**)
Sig. (2-tailed) .123 .000
N 300 300 292
HSRANK Pearson Correlation -.086 .590(**) 1
Sig. (2-tailed) .144 .000
N 292 292 292
VERBAL Pearson Correlation .015 .293(**) .390(**)
Sig. (2-tailed) .806 .000 .000
N 259 259 252
MATH Pearson Correlation -.021 .253(**) .434(**)
Sig. (2-tailed) .741 .000 .000
N 259 259 252
GENDER Pearson Correlation .067 -.167(**) -.192(**)
Sig. (2-tailed) .246 .004 .001
N 300 300 292
RACE Pearson Correlation .047 .092 .132(*)
Sig. (2-tailed) .414 .113 .024
N 300 300 292
MAJOR Pearson Correlation -.168(**) .116(*) .271(**)
Sig. (2-tailed) .004 .045 .000
N 300 300 292
VERBAL MATH GENDER
CLASS Pearson Correlation .015 -.021 .067
Sig. (2-tailed) .806 .741 .246
N 259 259 300
HSGPA Pearson Correlation .293(**) .253(**) -.167(**)
Sig. (2-tailed) .000 .000 .004
N 259 259 300
HSRANK Pearson Correlation .390(**) .434(**) -.192(**)
Sig. (2-tailed) .000 .000 .001
N 252 252 292
VERBAL Pearson Correlation 1 .655(**) .051
Sig. (2-tailed) .000 .413
N 259 259 259
MATH Pearson Correlation .655(**) 1 .087
Sig. (2-tailed) .000 .162
N 259 259 259
GENDER Pearson Correlation .051 .087 1
Sig. (2-tailed) .413 .162
N 259 259 300
RACE Pearson Correlation .088 .012 .008
Sig. (2-tailed) .157 .842 .895
N 259 259 300
MAJOR Pearson Correlation .167(**) .135(*) -.195(**)
Sig. (2-tailed) .007 .029 .001
N 259 259 300
RACE MAJOR
CLASS Pearson Correlation .047 -.168(**)
Sig. (2-tailed) .414 .004
N 300 300
HSGPA Pearson Correlation .092 .116(*)
Sig. (2-tailed) .113 .045
N 300 300
HSRANK Pearson Correlation .132(*) .271(**)
Sig. (2-tailed) .024 .000
N 292 292
VERBAL Pearson Correlation .088 .167(**)
Sig. (2-tailed) .157 .007
N 259 259
MATH Pearson Correlation .012 .135(*)
Sig. (2-tailed) .842 .029
N 259 259
GENDER Pearson Correlation .008 -.195(**)
Sig. (2-tailed) .895 .001
N 300 300
RACE Pearson Correlation 1 .031
Sig. (2-tailed) .596
N 300 300
MAJOR Pearson Correlation .031 1
Sig. (2-tailed) .596
N 300 300
Note
** Correlation is significant at the 0.01 level (2-tailed).
* Correlation is significant at the 0.05 level (2-tailed).
Table 5: Coefficients of Preliminary Models
TOTALCT INDUCT DEDUCT
Constant -2.978 *** .198 -3.177
(-1.953) (.207) (-3.464)
CLASS .594 ** .280 .315 *
(2.115) (1.586) (1.861)
HSRANK 3.278 *** 1.639 ** 1.639 ***
(2.978) (2.373) (2.475)
VERBAL .012 *** .007 *** .005 **
(3.360) (3.204) (2.242)
MATH .018 *** .006 *** .012 ***
(5.050) (2.800) (5.475)
GENDER .203 (.436) -.253 (-.867) .456 (1.629)
RACE .981 * .716 ** .265
(1.738) (2.021) (.781)
MAJOR -.402 -.416 .014
(-.860) (-1.420) (.051)
R Square 0.409 0.271 0.372
F Statistics 24.140 *** 12.987 *** 20.637 ***
EVAL ANALY INFER
Constant -1.856 .375 -1.498
(-2.351) (.794) (-1.585)
CLASS .188 .044 .362 **
(1.289) (.509) (2.080)
HSRANK 1.325 ** .701 ** 1.252 *
(2.324) (2.055) (1.836)
VERBAL .006 *** 004 *** .002 ***
(3.008) (3.865) (.975)
MATH .004 ** .002 ** .012
(1.977) (2.036) (5.479)
GENDER -.017 (-.071) -.005 (-.033) .225 (.779)
RACE .440 -.121 .662 *
(1.505) (-.692) (1.893)
MAJOR -.192 .075 -.285
(-.795) (.521) (-.984)
R Square 0.215 0.250 0.301
F Statistics 9.528 *** 11.581 *** 14.976 ***
Notes
t statistics are in parentheses.
Significant level : * 0.10, ** 0.05, ***0.01
Table 6: Coefficients of Final Models
TOTALCT INDUCT DEDUCT
Constant -2.978 ** .122 -2.904 ***
(-1.960) (.127) (-3.280)
CLASS .639 ** .304 * .338 **
(2.303) (1.744) (2.019)
1.012 1.012 1.012
HSRANK 2.949 *** 1.614 ** 1.394 **
(2.829) (2.461) (2.232)
1.287 1.287 1.271
VERBAL .012 *** .007 *** .005 **
(3.313) (3.081) (2.335)
1.867 1.867 1.860
MATH .018 *** .006 *** .012 ***
(5.214) (2.725) (5.760)
1.948 1.948 1.939
RACE .982 * .700 **
(1.745) (1.977)
1.020 1.020
R Square 0.406 0.264 0.363
F Statistics 33.688 *** 17.684 *** 35.198 ***
EVAL ANALY INFER
Constant -1.177 .366 .922
(-1.652) (.864) (1.029)
CLASS .372 **
(2.033)
1.010
HSRANK 1.267 ** .703 ** 1 995 ***
(2.364) (2.208) (3.011)
1.260 1.260 1.192
VERBAL .006 *** 004 *** .010 ***
(3.130) (3.943) (5.184)
1.854 1.854 1.182
MATH .003 * .002 **
(1.859) (2.096)
1.937 1.937
RACE
R Square 0.200 0.250 0.200
F Statistics 20.494 *** 27.043 *** 20.046 ***
Notes
t statistics are in parentheses.
Significant level : * 0.10, ** 0.05, ***0.01
Number underneath parentheses is variance inflation factor,
a measure of collinearity.
