Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

1 Commit
 
 
 
 
 
 
 
 

Repository files navigation

Statistics Study Project

Self-Regulation and Academic Performance

Completed: 12/26/2025

A statistical analysis of the relationship between students' self-regulation, gender, and academic outcomes using multiple linear and logistic regression in R.

The project was completed as an empirical regression analysis assignment involving data preparation, exploratory data analysis, linear regression, logistic regression, model diagnostics, and interpretation of statistical results. This project was completed as part of the Statistics course at HSE University.

Disclaimer: The results are not intended to be scientifically significant, as some of the data has been generated. This project is intended for educational purposes in the study of statistics.

Research Objective

The study examines whether students' general level of self-regulation and gender predict academic outcomes.

Two regression models are used:

  • Multiple linear regression to predict test scores.
  • Logistic regression to predict the probability of passing the test.

Research Questions

  1. To what extent does general self-regulation predict students' academic performance?
  2. Are there gender differences in academic performance after controlling for self-regulation?
  3. Do self-regulation and gender predict the probability of passing the test?

Data

The dataset contains 204 observations of university students from Yaroslavl State University. The original self-regulation data were collected in 2023 using the Self-Regulation Profile of Learning Activity Questionnaire (SRPLAQ).

The analysis uses:

  • SR — general level of self-regulation, ranging from 0 to 58
  • Gender — gender of the student
  • Test_score — test score on a 0–100 scale
  • Pass — binary indicator of passing the test

The self-regulation measure is based on the SRPLAQ developed by Morosanova and colleagues. The questionnaire assesses conscious self-regulation of learning activity, with the general level calculated as an integrative indicator.

Synthetic Variables

The original dataset did not contain the variables required for the assignment in the necessary form. Therefore, Gender and Test_score were generated or transformed for the purposes of the statistical analysis. This project is considered to be an educational one, and the data do not claim to be of scientific significance.

Gender was randomly assigned to obtain approximately balanced groups.

The test score was derived from students' existing academic performance data. Original grades were converted from a 5-point scale to a 100-point scale using a predefined transformation and additional random variation. Missing grades for 47 participants were generated from a normal distribution based on the observed mean and standard deviation.

For logistic regression, a score of 61–100 was classified as Passed, while a score of 0–60 was classified as Not Passed.

The dataset used in this repository is a course-project dataset based on the original study data. The regression.xlsx file contains the calculations used to transform the academic performance variable and construct the analysis dataset. The R Markdown file serves as the analytical codebook, documenting variable preparation, transformations, and statistical procedures.

Analysis

The project includes:

Exploratory Data Analysis

  • Data structure and missing-value checks
  • Frequency tables
  • Descriptive statistics
  • Histograms
  • Boxplots
  • Normality checks

Multiple Linear Regression

The model predicts Test_score from SR and Gender.

Model diagnostics include:

  • Residuals vs. Fitted plot
  • Q-Q plot
  • Scale-Location plot
  • Shapiro-Wilk test
  • Breusch-Pagan test
  • Durbin-Watson test
  • Variance Inflation Factor (VIF)
  • R² and adjusted R²
  • F-test
  • Standardized coefficients

Logistic Regression

The model predicts Pass from SR and Gender.

The analysis includes:

  • Log-odds coefficients
  • Odds ratios
  • Marginal effects
  • Logit linearity assessment
  • Multicollinearity assessment
  • Predicted probabilities
  • Classification performance
  • ROC curve and AUC

Main Findings

Self-regulation was a significant positive predictor of academic performance in the linear regression model. A one-point increase in self-regulation was associated with an average increase of approximately 0.42 points in test score.

Gender was not a statistically significant predictor of test score after controlling for self-regulation.

The logistic regression showed that higher self-regulation was also associated with higher odds of passing the test. Each additional point in self-regulation increased the odds of passing by approximately 5%.

Gender was not a statistically significant predictor of passing the test.

The linear model explained approximately 8% of the variance in test scores, indicating that self-regulation and gender explain only a limited part of individual differences in academic performance.

Repository Contents

  • linear and logistic.Rmd — R Markdown file containing the complete analysis, statistical procedures, visualizations, and interpretation.
  • regression.xlsx — dataset and intermediate calculations used for the analysis.

References

  • Anazifa, R., Limiansi, K., & Pratama, A. (2023). Students’ Self-Regulated Learning based on Gender and Disciplinary Differences during Online Learning. Journal of Science Education Research, 7, 39–45. https://doi.org/10.21831/jser.v7i1.58174
  • Morosanova, V. I. (2016). Conscious self-regulation as a psychological resource for achieving educational and professional goals. Pedagogy, 10, 13–24.
  • Morosanova, V. I. (2022). Psychology of conscious self-regulation: From its origins to modern research. Theoretical and Experimental Psychology, 15(3), 57–83.
  • Morosanova, V. I., Vanin, A. V., & Tsyganov, I. Yu. (2011). Creating a new version of the questionnaire method "Self-Regulation Profile of Learning Activity Questionnaire (SRPLAQ)". Theoretical and Experimental Psychology, 4(1), 5–15.

About

Statistical analysis of the relationship between students’ self-regulation, gender, and academic outcomes using multiple linear and logistic regression in R. Course project for the Statistics course at HSE

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors