In social science research, it’s often not enough to simply observe differences between groups. We also want to understand which variables cause those differences and predict which group a new individual might belong to. This is where discriminant analysis becomes useful. In this article, we’ll explain what discriminant analysis is, when to use it, and how to interpret it—clearly and with examples.

1. What Is Discriminant Analysis?

Discriminant analysis is a multivariate statistical method used to assign individuals or observations to predefined groups. It also identifies which variables best distinguish between those groups.

Example: Predicting whether university students are in face-to-face or distance education based on their achievement, motivation, and stress levels.

2. When to Use It?

• When the dependent variable is categorical (e.g., group membership) with two or more groups
• When independent variables are continuous (numerical)
• When you want to identify variables that differentiate between groups
• When you want to predict group membership for new individuals

3. Key Concepts

Discriminant Function
Linear combinations that best separate the groups. Each function can be thought of as an axis.

Wilks’ Lambda
Tests whether the differences between groups are statistically significant. The smaller the value, the stronger the discrimination.

Classification Matrix
Shows how accurately the model classifies individuals. The higher the percentage, the more successful the model.

4. How to Run Discriminant Analysis in SPSS

• Go to: Analyze > Classify > Discriminant
• Select the group variable (categorical) and independent variables (continuous)
• Under “Statistics,” check options like Wilks’ Lambda and Eigenvalues
• Under “Classify,” define classification settings
• Click “OK”

5. How to Interpret Results

• Wilks’ Lambda (p < 0.05): Significant difference between groups
• Standardized discriminant coefficients: Show which variables contribute most to group separation
• Classification accuracy: Indicates model performance (e.g., 82% correct classification)

6. How to Report in Your Thesis

“Discriminant analysis revealed that achievement and motivation significantly distinguish students by education type (Wilks’ Lambda = 0.72, χ² = 18.45, p < 0.01). The model correctly classified 81% of students.”

7. Don’t Confuse Discriminant Analysis With

8. Conclusion

Discriminant analysis not only reveals differences between groups but also identifies the variables responsible for those differences. With its ability to predict group membership, it serves as a powerful analytical tool in your thesis.