Student’s Name
Institution Affiliation
Date
Article Critique: Logistic Regression
The study by Tukur and Usman (2016) aimed to explore on whether gender and Jamp scores influences student admission in Kano University of Science and Technology. The researches wanted to develop a model to fit the data and determine the chances of a student being admitted or not. The author used the binary logistic regression approach because the independent variable was categorical. Besides, they claimed that traditional methods such as linear discriminate function analysis and ordinary least square (OLS) approach could not be ideal in this case since they have strict assumptions. That is, they need verification of multivariate normality with equal variance and covariance for discriminate analysis and linearity, normality and continuity for OLS. Hence, logistic regression analysis was most appropriate in the analysis according to the authors.
In my view, the method used by the authors was the most suitable for the study. The logistic regression method is well suited when establishing and testing the hypothesis about the association between a categorical outcome variable and one or more categorical or continuous explanatory variables. In this case, the dependent variable was a dichotomous outcome, admitted or not admitted, while the independent variables were gender and Jamp scores. Thus, I think the logistic was appropriate in developing a model for predicting the binary outcome.
The results in the study were presented in the form of tables. The authors presented the results in three tables including the logistic regression table, the goodness of fit test table and the table of observed and expected frequencies as indicated in Appendices. The tables could be used to tell whether the model developed from the data was appropriate, whether the two predictors were significant and their association with the dependent variable. For instance, It could be noted from Table 1 that the independent variable (Gender) is not statistically significance as its P-value is more than 0.05 that is (0.281>0.05) while the second independent variable (Scores) is statistically significance because its P-value is less than 0.05 that is (0.024<0.05). Based on Table 2 (Appendix), it is possible to claim that the model was a good fit for the data since the deviance test p-value was less than the level of significance, p = 0.009<0.05. Thus, the tables presented by authors are standalone and are easy to interpret the analysis.
Reference
Tukur, K. and Usman, A. U. (2016). Binary Logistic Regression Analysis on Admitting Students Using JAMP Score, International Journal of Current Research, 8(01), [identifier removed].
Appendix
Table 1
Logistic Regression Table
Predictor Coef SE Coef Z P Ratio Lower Upper
Constant 2.[identifier removed].[identifier removed].69 0.090
Gender
[identifier removed].[identifier removed].[identifier removed].08 0.[identifier removed].47 0.73 2.95
Score (% -0.[identifier removed].[identifier removed].25 0.[identifier removed].94 0.89 0.99
Log-Likelihood = -93.594
Test that all slopes are zero: G = 7.657, DF = 2, P-Value = 0.022
Goodness-of-Fit Tests
Method Chi-Square DF P
Pearson 74.[identifier removed].232
Deviance 96.[identifier removed].009
Hosmer-Lemeshow 6.[identifier removed].640
Table 2
Goodness-of-Fit Tests
Method Chi-Square DF P
Pearson 74.[identifier removed].232
Deviance 96.[identifier removed].009
Hosmer-Lemeshow 6.[identifier removed].640
Table 3
Table of Observed and Expected Frequencies
(See Hosmer-Lemeshow Test for the Pearson Chi-Square Statistic)
Group
Value [identifier removed] Total
Obs [identifier removed]
Exp 2.9 3.6 4.4 4.5 5.1 5.5 6.8 6.6 7.4 6.2
Obs [identifier removed]
Exp 12.1 12.4 11.6 10.5 9.9 9.5 10.2 8.4 7.6 4.8
Total [identifier removed]
Measures of Association:
(Between the Response Variable and Predicted Probabilities)
Pairs Number Percent Summary Measures
Concordant [identifier removed].3% Somers' D 0.27
Discordant [identifier removed].6% Goodman-Kruskal Gamma 0.27
Ties [identifier removed].1% Kendall's Tau-a 0.12
Total [identifier removed].0%
Use this resource to support learning, planning and original work. Follow your institution’s academic integrity requirements.