Acute Performance Estimation of Students Using Quantile Regression Approach (A Case Study of Lahore)

Sunaina Ishtiaq, Yasar Mahmood and Hina Khan

Extreme behavior (Performance) of students is inclined by number of
factors which are must be painted for important policy implications. This
study states that CGPA is the most important system to detect student
performance. Data on CGPA has been collected from B.A/B.Sc (Hons.) of
32 private and public universities of Lahore. Generally, researchers
investigate an average performance of the students with classical methods
of simple linear regression. This approach does not give complete picture
of different variables influencing student performance from corner to
corner. Quantile regression introduces information across the whole
distribution of the student’s achievements. Study furnishes that students
performance strongly affected by father’s education. Student’s gender,
passion for fashion, and mother’s job are significant factors. Class
participation is found as a magical variable that has positive impact on
student performance at all quantiles. The quantile estimate of student
performance shows that effect of the urban-rural difference is significant
factor. The study clearly shows for high performance students, factors like
mother occupation, father education, gender and area become significant
at high quantiles. The results highlight that quantile regression model is a
useful technique to examine information than ordinary least squares. It also depicts that ordinary least squares underestimated and overestimated the Quantile regression at different quantiles.

Permanent link to this article: