Assessing prerequisites as a measure of success in a principles of finance course.

Blaylock, Alan ; Lacewell, Stephen K.

ABSTRACT

This paper seeks to determine if student success in courses that serve as prerequisites to the principles of finance course is carried over to success in the actual finance course. Also, additional variables, such as the gender of the student, the number of background courses taken, and the time since the courses were taken are analyzed to determine if they have a direct impact on final grades. The quantity of prerequisite courses and their timing are found to significantly influence student performance in the introductory finance course. This shows that adequate and timely exposure to prerequisite subjects are helpful in learning finance.

INTRODUCTION

Most principles classes of all disciplines have few, if any, formal prerequisite class requirements. The principles of finance class is usually the exception requiring prerequisites in both economics and accounting and often a math or statistics course. Since the principles of finance is by nature a very quantitative course, it would seem that a student's success in the prerequisite classes should carry over to the finance course.

Making this topic more interesting are the ever-changing requirements necessary for colleges of business to attain AACSB accreditation. Even a casual glance through the latest AACSB Eligibility Procedures and Standards for Business Accreditation (2003) reveals an increased emphasis on standards related to the assurance of learning. Thus, the design of courses used as prerequisites for the basic finance course as well as the design of finance curriculums in general have taken on an increased level of importance.

This paper seeks to determine if student success in courses that serve as prerequisites to the principles of finance course is carried over to success in the actual finance course. Also, additional variables, such as the gender of the student, the number of background courses taken, and the time since the courses were taken are analyzed to determine if they have a direct impact on final grades.

LITERATURE REVIEW

A study of the factors that determine performance in business related courses is not necessarily groundbreaking research. The area of economics, for example, includes papers by Schuhmann, McGoldrick, and Burrus (2005), Laband and Piette (1995), Anderson, Benjamin, and Fuss (1994), Bosshardt and Watts (1990), and Borg, Mason, and Shapiro (1989). Studies that examine factors related to performance in accounting classes include Gracia and Jenkins (2003), Drennan and Rohde (2002), Murphy and Stanga (1994), Graves, Nelson, and Deines (1993). Other studies related to student preparedness and student performance include one on business communications (Marcal and Roberts, 2000) and a study concerning performance on the Educational Testing Service Major Field Exam in Business (Bagamery, Lasik, and Nixon, 2005). However, the very few studies performed on the area of finance have mostly focused on self-reported qualitative factors such as student effort and test anxiety. Papers that study the quantitative relationship between student success in a principles of finance course and student success in the prerequisites needed for this class are very few. Only one paper by Didia and Hasnat (1998) touches on this subject. They found that a student's cumulative GPA has a statistically significant positive impact on success in the finance course. They also noted that a student's prior performance in accounting, economics, and math tended to carry over to success in finance. This study adds other variables not considered in the Didia and Hasnat study that may foretell success or failure in the basic finance course.

DATA AND METHODOLOGY

This study uses academic transcript data for students enrolled in six sections of the introductory finance course in the fall 2004 and spring 2005 semesters. All sections were taught by the same instructor so faculty influence is not an issue. Student performance is measured by the semester average for each student and is obtained from the instructor. All other information is obtained from student transcripts. One hundred forty out of 189 observations are usable.

Didia and Hasnat (1998) explain the grade received in a principles of finance class as function of maturity, background, aptitude, effort, and faculty contribution. They find all of these factors significantly influence student performance in the principles of finance course. Our study concentrates on the background component in detail. Didia and Hasnat use GPAs of prerequisite courses to measure student background. Specifically, they use the average GPA for the first two accounting courses, the average GPA for the first two economics courses (micro and macro), and the highest GPA of all math courses taken. Our present study expands on these variables. Essentially, our study includes the number and timing of prerequisites taken in addition to their GPAs.

The principles of finance course usually requires prerequisites courses in the areas of math, accounting, and economics. Our measures of student background relate to not only the academic performance as measured by grades, but also the quantity, and timing of these prerequisites courses. Similar to Didia and Hasnat, academic performance in prerequisite courses is measured by the average GPA for all accounting courses taken, the average GPA for all economics courses taken, and the average GPA for all math courses taken at the College Algebra level and above. Using only the GPAs in the math classes at or above the College Algebra level reduces any grade inflation from developmental math courses. To explain further, new students in the university that are determined to be weak in the area of math are enrolled in developmental math courses to prepare them for College Algebra. The weaker the student the more developmental math courses that are taken. High grades in developmental math courses are obviously not on par with high grades in College Algebra and above. A student may have many high developmental math grades but only mediocre algebra and calculus grades. Using all math grades to include the developmental math grades would not provide an accurate measure of a student's math background. However, as indicated below, a dummy variable measures if a developmental math course has been taken.

In addition to academic performance in the perquisite courses, the quantity of those prerequisite courses are also included as a measure of student background. Some students may have more exposure to these prerequisite areas than others which may lead to a better finance grade. For instance, students that have taken more than the first two accounting or economics courses, regardless of grade, may have a better grasp of these principle areas such that their performance in finance is enhanced. To incorporate this information, variables are added that use the number of math, accounting, and economics courses for which the student has taken. Also, although the GPAs of any developmental math courses are not included in the academic performance variables mentioned previously, a variable is included to indicate if any developmental math courses were ever taken.

A student's ability to harness prerequisites knowledge may be limited by the length of time since the prerequisites class was taken. Thus, the timing of prerequisite courses should be included in addition to the academic performance and quantity of prerequisite courses taken. Variables related to the timing of the prerequisites simply indicate the number of semesters since the last prerequisites course in each of the areas (math, accounting, and economics) was taken. A similar variable is used by Austin and Gustafson (2006).

As in Didia and Hasnat, we use cumulative GPA at the time of enrolling in the principles of finance course not only as a measure of student aptitude but also as a means of controlling the other GPA-related independent variables. Variables are also included to indicate gender and transfer status, and in addition to Didia and Hasnat, variables are included to indicate if the student is a finance, accounting, or economics major.

Although not usually a formal prerequisite, the area of statistics may also be useful in the finance class, especially when studying measures of risk such as variance and beta. Since statistics is not a formal prerequisite for the principles of finance course, the majority of students had not taken any statistics courses (only 71 out of 189). Using variables related to grades and timing of a student's statistics course severely limits the sample size. A variable that measures the quantity of statistics courses taken, however, would use the full sample size. Therefore, the variable measuring the quantity of statistics courses taken is used with the full sample, and the variables measuring the GPA and timing of a student's statistics course is used with the reduced sample. This results in 69 usable observations.

Two models are used. The first limits the measurements of statistics background to only the number of statistics courses taken while the second adds variables that measure the grades and timing of statistics courses. The first model is

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] Equation (1)

The second model is

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] Equation (2)

Where C is a constant; GENDER is a dummy variable that equals 1 if the student is male and 0 if the student is female; TRANSFER is a dummy variable that equals 1 if the student is a transfer student; DEVMATH is a dummy variable that equals 1 if the student has taken any developmental math course; ACC_MAJOR, ECON_MAJOR, and FIN_MAJOR are dummy variables that each equal 1 if the student is an accounting major, economics major, or finance major, respectively, and 0 otherwise; MATH_AGE, ECON_AGE, ACC_AGE, and STATS_AGE equal the number of semesters since, respectively, a math, economics, accounting, or statistics course was taken, MATH_GPA, ECON_GPA, ACC_GPA, and STATS_GPA equal the average GPA for the student's math, economics, accounting, and statistics courses, respectively, MATH_Q, ECON_Q, ACC_Q, AND STATS_Q equal the number of, respectively, math, economics, accounting, and statistics courses were taken, and GPA equals the cumulative GPA. Descriptive statistics for each of the variables are presented in Table 1.

RESULTS

The possible dependence between the prerequisite GPAs with each other as well as the cumulative GPA may result in multicollinearity. A rule of thumb suggested by Griffiths, Hill, and Judge (1993, page 435) is that multicollinearity may be a problem given a correlation coefficient greater than 0.8 or 0.9. Hair, Anderson, Tatham, and Black (1995, page 127) suggest harmful collinearity with correlations coefficients above 0.9. Hair, et. al. and Myers (1990, page 369) suggest multicollinearity may be problem with Variance Inflation Factors (VIFs) greater than 10.

The correlation matrix in Table 2 indicates relatively higher correlations are associated with the ages of the prerequisites courses. STATS_GPA has notable high correlations with MATH_AGE and ECON_AGE, and the high correlation of 0.85 with ACC_AGE indicates a possible problem with multicollinearity. GPA also has a relatively high correlation with STATS_AGE. Notice that this potential problem only impacts those variables in Equation 2. Table 3 reports Variance Inflation Factors (VIFs) for each non-binary variable. The correlations of the variables and their VIFs in Equation 1 (designated as Equation 1A) do not indicate any potential problem with multicollinearity. For Equation 2 (designated as Equation 2A), the correlations involving STATS_GPA may pose a problem. Although none of the VIFs break the threshold noted above, a few are dangerously high and should warrant concern during model selection. Table 3 also reports VIFs for reduced models to be discussed shortly.

Another concern is the large number of variables in each equation compared to the sample size. A stepwise regression could be performed, but Greene (1997, page 401) notes the possible faulty inference procedures associated with it. All of these variables are included due to the possible relationships with student performance in the principles of finance course. However, the inclusion of so many variables may cloud the results so that the marginal variable contributes little if any explanatory power. The solution should hold to the theoretic necessity of variable inclusion and at the same time use the mechanical nature of the stepwise regression process. The Akaike Information Criterion (AIC) and the Schwartz Criterion are used to arrive at a more parsimonious model. All combinations of variables are included in the equation to find the best AIC and Schwartz Criterion statistic.

The OLS coefficients from the two models are given in Table 4. In Equation 1 (designated as Equation 1A) three variables are significant, two of which, MATH_Q and ACC_AGE, are unique to this study. MATH_Q indicates that a student's final semester average in principles of finance increases by about 3 percentage points for each math class taken. ACC_AGE indicates that the semester average decreases by about 1 percentage point for every semester since the student has taken an accounting class. Not surprisingly, GPA contributes largely to the student's semester average.

The model of best fit as determined by the AIC and Schwartz Criterion is designated as Equation 1B. This equation also has the second highest adjusted R squared of all the possible models. MATH_Q, ACC-AGE, and GPA are all still statistically significant, although the significance of ACC_AGE is reduced. GENDER and MATH_AGE become significant in this model. A puzzling finding is the positive coefficient for MATH_AGE. This shows that student performance improves the longer the amount of time has elapsed since the last math class was taken. A possible explanation could be that students with better math ability complete math requirements early in their academic career, resulting in a longer period since the last math class was taken, while those with poorer math ability would have only recently completed a high level math course.

Equation 2 (designated as Equation 2A) adds STATS_AGE and STATS_GPA to the model with a cost of a reduced sample size. GPA still remains the largest significant contributor to student scores, but MATH_Q and ACC_AGE lose their significance. This could be a result of the reduced sample size, the inclusion of the two additional variables, or both. Notice that the new variable, STATS_GPA, is positively significant. STATS_Q was not significant in Equation 1A. This may show that a proficient statistics knowledge and not just an exposure to a statistics background is important to a student's performance.

A concern for Equation 2 was the possible problem with multicollinearity. Since the "age" variables evidenced the potential problem, these variables (MATH_AGE, ECON_AGE, ACC_AGE, and STATS_AGE) were all removed from the model. Although not reported here, all remaining variables kept their signs, and STATS_GPA and GPA retained their levels of significance. For this model, GPA was found to have a VIF of 5.88 with the second highest VIF of 3.03 associated with ECON_GPA.

The model of best fit for Equation 2 as determined by the AIC and Schwartz Criterion is designated as Equation 2B. This equation also has the highest adjusted R squared of all the possible models. STATS_GPA and GPA retain their levels of significance. Interestingly, the model fits the data better with ECON_GPA. Another puzzling finding is its negative coefficient. The authors can offer no reasonable explanation for this phenomenon. Notice that the VIFs reported in Table 3 indicate no evidence of multicollinearity.

An observation may be made concerning the sign change in several variables from Equation 1A to Equation 2A. For instance, ACC_GPA is positive in Equation 1A, yet negative in Equation 2A. This is because Equation 2A uses a different sample than Equation 1A. To fully appreciate the different characteristics of the sample of students who have not taken statistics and the sample of students who have, Equation 1A is run using both sub-samples. These new equations are designated as Equation 1C and 1D. STATS_Q is omitted from these regressions since STATS_Q would equal zero for all observations for the sub-sample of students who have not taken a statistics course. Although not reported here, STATS_Q was included when using the sub-sample of students having a statistics course, and the findings are similar to those that are reported in Table 4.

Equation 1C uses the sub-sample of students who have taken a statistics course. Notice that the coefficients have the same sign as those in Equation 2A. This is because both equations use the same sample. Notice also that this model does not fit the data quit as well as Equation 2A as indicated by the AIC, Schwartz Criterion, and adjusted R squared. Adding the STATS_GPA variable as in Equation 2A improves the model for this sample. Equation 1D uses the sub-sample of students who have not taken a statistics course. The variables that are significant here are also those that are significant in Equation 1A. Reviewing Equations 1C and 1D together with Equations 1A and 2A indicates that for those students who have taken a statistics course, better performance in the statistics course relates to better performance in the introductory finance course; for those students who have not taken a statistics course, more math courses and the more recent the last accounting course was taken results in better performance in the introductory finance course.

CONCLUSION

As in Didia and Hasnat (1998) cumulative GPA has been found to contribute significantly to a student's performance in the principles of finance class. This research identifies the quantity of math classes taken, the age of a student's last accounting class, and the GPA in a student's statistics classes as additional determinates of performance. Of course, different instructors teach in different ways, and what contributes to one instructor's students may not contribute to another's. However, instructors and program developers need to be aware that a student's performance in the principles of finance course may be a function of such factors as the timing and quantity of certain prerequisite courses, and not only the GPA in those courses.

REFERENCES

AACSB International, "Eligibility procedures and standards for business accreditation," AACSB International Business Accreditation Seminar (New Standards), November 2003, 156-239.

Anderson, Gordon, Dwayne Benjamin, and Melvyn A Fuss(1994). The Determinants of Success in University Introductory Economics Courses. Journal of Economic Education, 25 (2), 99-119.

Austin, M. Adrian and Leland Gustafson (2006). Impact of course length on student learning. Journal of Economics and Finance Education 5 (1), 26-37.

Bagamery, D. Bruce, John J. Lasik, and Don R. Nixon (2005). Determinants of success on the ETS business major field exam for students in an undergraduate multisite regional university business program. Journal of Education for Business, September/October, 55-63.

Borg, O. Mary, Paul M. Mason, and Stephen L. Shapiro (1989). The case of effort variables in student performance. Journal of Economic Education, 20 (3), 308-313.

Bosshardt, William and Michael Watts (1990). Instructor effects and their determinants in precollege economic education. Journal of Economic Education, 21 (3), 265-276.

Didia, Dal and Babnan Hasnat (1998). The determinants of performance in finance courses. Financial Practice and Education, 8 (1), 102-107.

Drennan, L.G. and F.H. Rohde (2002). Determinants of performance in advanced undergraduate management accounting: an empirical investigation. Accounting and Finance, 42, 27-40.

Gracia, Louis and Ellis Jenkins (2003). A quantitative exploration of student performance on an undergraduate accounting programme of study. Accounting Education, 12 (1), 15-35.

Graves, O. Finley, Irva Tom Nelson, and Dan S. Deines (1993) Accounting student characteristics: results of the 1992 Federation of Schools of Accountancy (FSA) Survey. Journal of Accounting Education, 11 (2), 221-225.

Greene, William H. (1997). Econometric analysis (Third Edition). Prentice Hall.

Griffiths, William E., R. Carter Hill, and George G. Judge (1993). Learning and practicing econometrics., John Wiley & Sons, Inc.

Hair, Joseph F. Jr., Rolph E. Anderson, Ronald L. Tatham, and William C. Black (1995). Multivariate analysis (Fourth Edition). Prentice Hall.

Laband, N. David and Michael J. Piette (1995). Does who teaches principles of economics matter. Papers and Proceedings of the American Economic Association, 85 (2), 335-338.

Murphy, P. Daniel and Keith G. Stanga (1994). The effects of frequent testing in an income tax course: an experiment. Journal of Accounting Education, 12 (1), 27-41.

Marcal, Leah and William W. Roberts (2000). Computer literacy requirements and student performance in business communications. Journal of Education for Business, May/June, 253-257.

Myers, Raymond H. (1990). Classical and modern regression with applications (Second Edition). Duxbury Press.

Schuhmann, W. Peter, Kim Marie McGoldrick, and Robert T. Burris (2005). Student quantitative literacy: importance, measurement, and correlation with economic literacy. American Economist, 49 (1), 49-65.

Alan Blaylock, Murray State University

Stephen K. Lacewell, Murray State University

Table 1: Descriptive Statistics

Variable Mean Standard
 Deviation

Semester average for the course (GRADE) 75.30 19.11
Number of males (GENDER) 35 --
Number of accounting majors (ACC_MAJOR) 11 --
Number of economics majors (ECON_MAJOR) 1 --
Number of finance majors (FIN_MAJOR) 8 --
Number of transfer students (TRANSFER) 42 --
Number of students that have taken 19 --
developmental math (DEVMATH)
Average GPA for all math courses taken
(at the college algebra level and above)
(MATH_GPA) 2.70 0.91
Number of semesters since a math course 7.19 6.58
was taken (MATH_AGE)
Number of math courses taken
(at the college algebra level and above)
(MATH_Q) 2.10 0.69
Average GPA for all economics courses taken 2.65 0.83
(ECON_GPA)
Number of semesters since an economics course 3.83 5.80
was taken (ECON_AGE)
Number of economics courses taken (ECON_Q) 2.45 0.58
Average GPA for all accounting courses taken 2.79 0.76
(ACC_GPA)
Number of semesters since an accounting 4.00 6.20
course was taken (ACC_AGE)
Number of accounting courses taken (ACC_Q) 2.04 0.67
Average GPA for all statistics courses taken 2.66 1.12
(STATS_GPA)
Number of semesters since a statistics course 3.14 5.30
was taken (STATS_AGE)
Number of statistics courses taken (STATS_Q) 1.36 0.51
Cumulative GPA (GPA) 2.80 0.60

Table 2: Correlation Coefficients Between the Variables

 GENDER ACC_MAJOR ECON_MAJOR

GENDER 1.0000
ACC_MAJOR -0.1984 1.0000
ECON_MAJOR -0.0086 -0.0588 1.0000
FIN_MAJOR 0.0117 -0.0443 -0.0491
TRANSFER -0.0064 0.0621 -0.0359
DEVMATH -0.0910 -0.1150 -0.0695
MATH_GPA -0.0550 0.2770 0.1268
MATH_AGE -0.0320 0.1846 -0.0499
MATH_Q -0.0961 0.0243 0.0646
ECON_GPA 0.0445 0.2358 0.1807
ECON_AGE -0.0413 0.1368 -0.0068
ECON_Q -0.0914 -0.1045 0.0437
ACC_GPA -0.1091 0.3181 -0.0233
ACC_AGE -0.0034 0.0246 0.0046
ACC_Q -0.2072 0.3794 0.0902
STATS_GPA 0.0382 0.1611 -0.0264
STATS_AGE -0.2218 0.1683 0.0369
STATS_Q -0.1389 -0.0594 -0.0292
GPA -0.1047 0.2809 0.1094

 FIN_MAJOR TRANSFER DEVMATH

GENDER
ACC_MAJOR
ECON_MAJOR
FIN_MAJOR 1.0000
TRANSFER 0.0913 1.0000
DEVMATH 0.0943 -0.0861 1.0000
MATH_GPA 0.1326 0.0120 -0.2928
MATH_AGE -0.0677 -0.1018 -0.1465
MATH_Q 0.0402 0.1040 0.1136
ECON_GPA -0.0546 -0.0070 -0.2123
ECON_AGE -0.0327 -0.1509 -0.0755
ECON_Q -0.1864 -0.0017 0.0338
ACC_GPA 0.1094 0.1711 -0.1600
ACC_AGE -0.1116 -0.0607 -0.0470
ACC_Q -0.0807 0.1582 -0.0360
STATS_GPA -0.0961 -0.1078 -0.0540
STATS_AGE 0.1103 -0.0436 -0.3967
STATS_Q -0.0990 -0.0334 0.0215
GPA 0.0623 -0.0266 -0.2365

 MATH_GPA MATH_AGE MATH_Q

MATH_GPA 1.0000
MATH_AGE 0.0644 1.0000
MATH_Q 0.0638 -0.1307 1.0000
ECON_GPA 0.4565 0.0528 0.0578
ECON_AGE 0.0239 0.5423 -0.1123
ECON_Q -0.0897 0.1863 0.0471
ACC_GPA 0.3757 0.2189 -0.0632
ACC_AGE 0.0765 0.3281 0.0130
ACC_Q 0.0186 0.0524 -0.0763
STATS_GPA 0.1593 0.7091 -0.0766
STATS_AGE 0.4968 0.1558 -0.0088
STATS_Q -0.0559 0.0018 0.0834
GPA 0.6261 0.0504 -0.0493

 ECON_GPA ECON_AGE ECON_Q

MATH_GPA
MATH_AGE
MATH_Q
ECON_GPA 1.0000
ECON_AGE 0.0697 1.0000
ECON_Q -0.1501 -0.0692 1.0000
ACC_GPA 0.4650 0.1359 -0.0530
ACC_AGE 0.1792 0.6117 0.0716
ACC_Q 0.0125 0.0817 0.0266
STATS_GPA 0.2219 0.7937 0.0739
STATS_AGE 0.5531 0.0120 -0.0456
STATS_Q -0.1269 -0.0034 0.2178
GPA 0.6694 0.0156 -0.0507

 ACC_GPA ACC_AGE ACC_Q

ACC_GPA 1.0000
ACC_AGE 0.0835 1.0000
ACC_Q 0.1415 -0.0702 1.0000
STATS_GPA 0.1289 0.8469 -0.0801
STATS_AGE 0.4927 0.0273 -0.0290
STATS_Q -0.0379 0.1306 0.0594
GPA 0.5949 0.0773 0.0511

 STATS_GPA STATS_AGE STATS_Q GPA

ACC_GPA
ACC_AGE
ACC_Q
STATS_GPA 1.0000
STATS_AGE 0.1217 1.0000
STATS_Q -0.1817 -0.0565 1.0000
GPA 0.2096 0.7397 -0.1196 1.0000

Table 3: Variance Inflation Factors

 Equation 1A Equation 1B Equation 2A Equation 2B

MATH GPA 1.85 -- 2.31 --
MATH AGE 1.69 1.15 2.49 --
MATH_Q 1.14 1.04 1.39 --
ECON GPA 2.18 -- 3.14 2.24
ECON AGE 2.33 -- 6.38 --
ECON_Q 1.28 -- 1.19 --
ACC GPA 1.91 -- 2.21 --
ACC AGE 1.86 1.13 9.36 --
ACC_Q 1.33 -- 1.64 --
STATS GPA -- -- 2.92 2.21
STATS AGE -- -- 5.63 --
STATS_Q 1.15 -- 1.41 --
GPA 3.08 1.02 6.01 3.44

Table 4: OLS Results
The dependent variable is the final grade received in the course.
The t-values are presented in parentheses.

 Independent Equation 1B
 Variables Equation 1A ([dagger]) Equation 1C

 13.862 15.9436 * 9.0852
C (1.3347) (1.7586) (0.4895)
 -4.4924 -5.5130 ** -4.1738
GENDER (1.6307) (2.2936) (1.0235)
 2.7105 2.122
ACC_MAJOR (0.6932) -- (0.3321)
 1.0356 2.0602
ECON_MAJOR (0.0915) -- (0.1204)
 0.2096 5.7957
FIN_MAJOR (0.0533) -- (0.8758)
 -1.2258 1.9768
TRANSFER (0.4220) -- (0.4737)
 3.2186 1.9786
DEVMATH (0.9977) -- (0.4022)
 1.0504 -0.9445
MATH_GPA (0.5317) -- (0.3065)
 0.2475 0.3145 *** 0.3426
Y (1.3831) (3.2456) (0.8571)
 3.1775 * 3.3122 * 3.671
MATH_Q (1.8646) (1.9153) (1.2019)
 -2.2232 -3.4389
ECON_GPA (0.9857) -- (0.8724)
 0.2413 0.2322
ECON_AGE (0.7846) -- (0.2904)
 -1.2427 1.2719
ECON_Q (0.5260) -- (0.3697)
 1.9838 -0.372
ACC_GPA (0.8386) -- (0.1052)
 -1.0226 *** -0.8208 * -0.7533
ACC_AGE (2.7338) (1.7632) (0.9605)
 -2.7260 -2.7057
ACC_Q (1.2705) -- (0.7732)

STATS_GPA -- -- --

STATS_AGE -- -- --
 2.6683
STATS_Q -1.4965 -- --
 20.0260 *** 19.6745 *** 25.9684 ***
GPA (5.4172) (8.5420) (3.9331)
Adjusted [R.sup.2] 0.4101 0.4276 0.3627
F 6.6851 21.7686 3.4191
N 140 140 69
Akaiki Information
Criterion 8.3855 8.2778 8.4976
Scwartz Criterion 8.7637 8.4039 9.0480

 Independent Equation 2B
 Variables Equation 1D Equation 2A ([dagger])

 23.8951 10.7505 26.5305 ***
C (1.5160) (0.6045) (3.0950)
 -5.7301 -0.708
GENDER (1.3584) (0.1779) --
 -1.8873 5.6129
ACC_MAJOR (0.3423) (0.9479) --
 -7.2943 0.0836
ECON_MAJOR (0.4203) (0.0053) --
 -4.7791 1.6759
FIN_MAJOR (0.8352) (0.2689) --
 -7.1869 3.0257
TRANSFER (1.5441) (0.7880) --
 3.9482 6.212
DEVMATH (0.7928) (1.3209) --
 1.777 -0.8033
MATH_GPA (0.5557) (0.2842) --
 0.2538 0.1464
MATH_AGE (1.1079) (0.3609) --
 4.7826 * 1.5646
MATH_Q (1.9669) (0.5409) --
 -0.8174 -4.1864 -5.8182
ECON_GPA (0.2664) (1.1572) (1.6241)
 0.163 0.4881
ECON_AGE (0.4324) (0.6626) --
 -3.6313 1.4322
ECON_Q (0.8931) (0.4517) --
 2.8819 -1.904
ACC_GPA (0.7616) (0.5758) --
 -2.8835*** -0.3391
ACC_AGE (2.9794) (0.4060) --
 -0.4594 -1.4706
ACC_Q (0.1397) (0.4568) --
 9.0958 *** 8.5344 ***
STATS_GPA -- (3.5219) (3.4837)
 -0.5006
STATS_AGE -- (0.6608) --
 3.182
STATS_Q -- (0.8150) --
 18.2497 *** 16.3821 ** 14.8288 ***
GPA (3.3605) (2.3821) (2.8214)
Adjusted [R.sup.2] 0.5623 0.4673 0.5424
F 4.1448 4.1397 27.8687
N 69 69 69
Akaiki Information
Criterion 8.5387 8.3459 8.0127
Scwartz Criterion 9.0891 8.9934 8.1422

([dagger]) indicates that White's corrected standard errors were used
due to the detection of heteroskedasticity.

*** Significance at the 0.01 level.

** Significance at the 0.05 level.

* Significance at the 0.10 level.