Regression Analysis on

Student Performance

A comprehensive regression analysis examining how social, domestic, and personal factors influence secondary school students' academic performance.

role

Project Lead

aim

Analyse factors that influence a student’s grades in high school

duration

January 2024 - May 2024

research questions

  1. Does a student’s social life (romantic relationship status, going out with friends, weekend alcohol consumption) affect their final grade?

  2. How does the way students spend their weekday time (study hours, travel time, free time, workday alcohol consumption) relate to academic performance?

  3. What is the correlation between students' personal well-being (health status, absences, study time) and their final grades?

  4. Do parental education levels impact students' academic outcomes?

  5. Is there a correlation between family life (parents' occupations and family relationships) and student academic performance?

data

my research focus

Each team member focused on one research question. Mine was: What is the correlation between students' personal well-being (health status, absences, study time) and their final grades?

Hypothesis:

Diagnostics:

  • Breusch Pagan Test: p-value = 0.00151 so model had non-constant errors

  • Shapiro-Wilk Test​: p-value = 2.807e-10​ so model had non-normal errors

  • VIF: No multicollinearity issues found

  • Box-cox Transformation: Used lambda of 1.07 to transform Y and updated Breusch Pagan Test p-value to 0.9537​ and Shapiro-Wilk Test p-value to <2.2e-16

  • K-Fold Cross Validation​: RMSE for transformed model was smaller (5.534) than the RMSE for the original model (22.73)​

Since p-value is less than alpha (< 2.2e-16) and F-statistic is greater than the critical value (588.2 > 0.480454), it can be concluded that a student's well-being and engagement have a linear impact on their final grade.

hypothesis tests

key findings

  • Students' social life, including relationship status, frequency of social outings, and alcohol consumption on weekends, significantly impacts their academic performance.

  • The way students allocate their weekday hours, particularly regarding study time and free time, shows a significant linear correlation with their final grades.

  • Personal well-being (health status and absences) and engagement levels strongly correlate with student grades.

  • Family factors, notably parental education levels and job statuses, also significantly relate to student academic outcomes, confirming that parental education and domestic environments are critical influencers.