Evaluating the Effect of Graduate Record Exam Scores - Research Paper Example

Data Analysis Report April Table of Contents April Table of Contents 2 List of Tables 2 List of Figures 2 3 Objective 4 Hypothesis 4 Review Literature 4 Methodology 5 Data 5 Analysis 5 Conclusion 9 References 10 List of Tables Table 1: Descriptive Statistics for Admission 5 Table 2: Cross-tabulation for Admission Vs Class 6 Table 3: Chi-Square test for the Relationship between Admission and Class 6 Table 4: Logistics Regression Estimates 7 Table 5: Probability Prediction for Admission using Class 8 Table 6: Probability Prediction for Admission using GRE Scores 8 Table 7: Binomial Test 9 List of Figures Figure 1: Pie of Admission by Class 6 Abstract This research work was aimed at evaluating the effect of Graduate Record Exam (GRE) scores, Grade Point Average (GPA) and undergraduate program (whether from high class or low class) enrolled on the probability of being admitted (admitted and not admitted) into graduate school. The formulated hypothesis was to proof that high GRE scores, high GPA value and undergraduate program have no affects on the probability of admission into graduate school against the alternative that high GRE scores, high GPA value and undergraduate program affect the probability of admission. From the results, high GRE scores and GPA affect positively the probability of being admitted into graduate school while having been into a high class undergraduate program has no effect on the probability of being admitted into graduate school. The probabilities of admission increase with increase in GRE scores and GPA. Objective The objective of this research study was to evaluate the effect of Graduate Record Exam (GRE) scores, Grade Point Average (GPA) and undergraduate program (whether from high class or low class) enrolled on the probability of being admitted (admitted and not admitted) into graduate school. Hypothesis The research work indented to proof the following hypothesis; H0: High GRE scores, high GPA value and undergraduate program have no affects on the probability of admission into graduate school H1: High GRE scores, high GPA value and undergraduate program affect the probability of admission The null hypothesis assumes that GRE scores, GPA value and undergraduate program have no affects on the probability of being admitted. This is in agreement with research tenets. Review Literature Admission to graduate schools and colleges in the US depends on GRE scores and GPA among many other considerations. Student personal statements, referees and transcripts also form an important part of considerations during admission. According to Richmond (2006), high GPA and GRE does not mean high probability of admissions but shows how the student is likely to perform as a student and mostly many colleges allow a GPA of 2.5 as the minimum. However, according Barnes (2004), the consideration of GPA and GRE scores depends on the course one is applied for. For example, applying for a medical course one has to have very good scores in science and mathematics. Methodology Quantitative research technique was used in this research work as sought to determine the relationship between the dependent variable (admission) and the independent variables (GRE, GPA and undergraduate program). The use of this methodology is evident through the use of structured data collection tool where the 4 aspects are strictly recorded for each participant. According to Godin, G. (1993), the relationship between dependent variable sand independent variables can only be shown through the use of quantitative research approach. Data Secondary data is used in this research study and is available from http://www.ats.ucla.edu/stat/stata/dae/logit.dta. This data if of 400 participants collected from different applications in the US and participant’s Graduate Record Exam (GRE) scores, Grade Point Average (GPA) and undergraduate program enrolled were sought. Analysis The variable GRE was in form of marks ranging from 220 to 800 and with a mean of 587.7 and a standard deviation of 115.52. The mean of Grade Point Average is 3.39 with a standard deviation of .38 and ranged from 2.26 and 4. Out of the 400 participants, 31.75% (127) were admitted while 68.25% (273) were not admitted. Of the admitted 127 admitted to graduate school, 23.62% were from high class while 76.38% were not from high class. Table 1: Descriptive Statistics for Admission Figure 1: Pie of Admission by Class Table 2: Cross-tabulation for Admission Vs Class From table 2, it is evident that from the cross-tab there are no small values in the cells and thus, chi-square tests and logistics regressions can be used to show the effect of the within subject factors on the dependent variable; admission into graduate school. Table 3: Chi-Square test for the Relationship between Admission and Class From table 3, Pearson chi-square test is used since the expected values on the cells are all >5 (the minimum is 30 in the last cell). Otherwise, exact Chi-square test would have been performed. From the results, there is a statistically significant relationship between class and the probability of being admitted [Chi-Square = 7.43, P = 0.006]. Table 4: Logistics Regression Estimates From table 4, 400 observations have been included in the model and four iterations have been performed. A Likelihood Ratio (LR) of 21.85 is significant at 5% level of significance p = 0.0001 while the within subject factors (GRE, GPA and Class) explain 4.36% of the total variation in the probability of being admitted. Using the Wald z-statistic for the coefficients, then GRE and GPA are significant in the model [Z = 2.31, p = 0.021 and Z = 2.05, p = 0.040 respectively] while class is not significant in the model at 5% level of significance [Z = 1.50, p = 0.134]. The odds ratios shows that a 10 point increase in GRE scores, the odds of being admitted to graduate school increases by a factor of 1.0251(1.00248^10). Further, a unit increase in GPA increases the odd of being admitted by a factor of 1.949 while have been in a high class undergraduate school leads to the probability of being admitted increasing by a factor of 1.548. Using probabilities; Table 5: Probability Prediction for Admission using Class The probability of being admitted having belonged to a high class undergraduate program if 0.391 compared to 0.293 for belonging not to a high class undergraduate program. This probability holds if the GRE and GPA results are held at their mean. Table 6: Probability Prediction for Admission using GRE Scores The probability of admission against the GRE scores are provided and increases as the GRE score increases [220, Pr = 0.1516 and 800, Pr = 0.4291]. Table 7: Binomial Test The binomial test above shows that the undergraduate program sample distribution is not binomial with p Read More