Friday 14 March 2025
A statistical analysis of student marks distribution reveals that the traditional assumption of normality is often misguided. Researchers at BITS Pilani, Goa Campus, have demonstrated that many datasets exhibit skewness, a characteristic that can lead to inaccurate conclusions.
When analyzing educational data, it’s common to assume that student marks follow a normal distribution. This assumption is based on the idea that marks are randomly distributed around a mean grade. However, this notion may not always hold true. In reality, marks often display asymmetrical patterns, with a concentration of students receiving either high or low grades.
To investigate this phenomenon, researchers applied the classical Shapiro-Wilk test to two datasets from the Math F113 course. The results showed that both datasets rejected normality, suggesting that the data does not conform to a traditional bell-curve shape. However, this conclusion may be too hasty, as it fails to account for skewness.
A modified version of the Shapiro-Wilk test was developed to address this limitation. This new approach transforms the data into an approximately normal form, allowing for a more accurate assessment of skewed datasets. The results of this test revealed that both datasets are consistent with a skew-normal distribution, which is characterized by a slightly asymmetrical pattern.
The implications of this study are significant. By recognizing the presence of skewness in educational datasets, researchers and educators can gain a better understanding of student performance. This knowledge can inform more effective grading policies and assessment design. Moreover, the methodology presented here has broader applications across various domains where skewed data is prevalent.
In order to validate these findings, future research should focus on extending this analysis to other courses and institutions. Comparative studies with alternative methods for detecting skew-normality could also provide valuable insights. Furthermore, simulations can be used to evaluate the power and robustness of the modified test under varying degrees of skewness and sample sizes.
This study highlights the importance of considering skewness when analyzing educational data. By adopting a more nuanced approach to statistical analysis, researchers and educators can gain a deeper understanding of student performance and make more informed decisions about assessment design and grading policies.
Cite this article: “Skewed Marks: A Statistical Analysis of Student Performance”, The Science Archive, 2025.
Student Marks Distribution, Statistical Analysis, Normality Assumption, Skewness, Educational Data, Shapiro-Wilk Test, Dataset Analysis, Grading Policies, Assessment Design, Skew-Normal Distribution
