Partial & Part (Semi-partial) Correlation — Previous Year Questions

📊 Topic Overview

Metric	Value
Papers Appeared	11 of 27 (40.7%)
Priority	🟠 HIGH
Section Distribution	Sec A: ~5
Most Repeated	"Differentiate between partial and part correlation with suitable example."

🎯 Must-Know Concepts

1. Partial correlation: correlation between X and Y with the effect of Z removed from both
2. Part (semi-partial) correlation: correlation between X and Y with the effect of Z removed from only one variable
3. When to use each
4. Formula and conceptual explanation

---

🔥 Most Repeated Question

Common phrasings:

• "Differentiate between partial and part correlation with suitable example."
• "Elucidate partial and part (semi-partial) correlation with the help of suitable examples."
• "Explain partial correlation. Compute Spearman's Rho for the following data."

Answer (Section A or B — ~350 words):

When studying the relationship between two variables, a third variable (control variable) may be influencing both, creating a spurious or inflated correlation. Partial and part correlations control for this third variable.

Partial Correlation (r_XY.Z):

Partial correlation is the correlation between variables X and Y after removing the effect of a third variable Z from both X and Y.

When to use: When Z is related to both X and Y and you want to know the "pure" relationship between X and Y independent of Z.

Example: The correlation between vocabulary (X) and reading comprehension (Y) may be inflated because both are related to age (Z). The partial correlation r_XY.Z gives the relationship between vocabulary and reading when age is controlled.

Formula (using zero-order correlations): $r_{XY.Z} = \frac{r_{XY} - r_{XZ} \cdot r_{YZ}}{\sqrt{(1 - r_{XZ}^2)(1 - r_{YZ}^2)}}$

Part (Semi-partial) Correlation (r_X(Y.Z)):

Part correlation (also called semi-partial correlation) is the correlation between X and Y after removing the effect of Z from only one variable (Y), not from both.

When to use: When you want to know how much unique variance in X is explained by Y after removing the influence of Z from Y alone. Commonly used in multiple regression to assess unique contribution of each predictor.

Example: What is the unique relationship between study time (Y) and exam score (X), after removing the effect of intelligence (Z) from study time — but not from exam score?

Key Difference:

Aspect	Partial Correlation	Part Correlation
Z removed from	Both X and Y	Only one variable
Question answered	Pure X–Y relationship	Unique variance in X explained by Y alone
Use	Controlling confounds	Multiple regression, unique contribution

Both partial and part correlations will be smaller than the zero-order (bivariate) correlation if Z is related to both variables.

---

🧠 Memory Aid

• Partial: Both variables partially cleaned of Z
• Part/Semi-partial: Only part of the pair is cleaned (one variable only)

---

📚 Related Content

• Partial Correlation: Controlling Third Variable
• Semi-partial and Multiple Correlation

✅ Self-Assessment

1. What is partial correlation and when is it used?
2. What is the difference between partial and part correlation?
3. Write the formula for partial correlation.
4. Give a psychology example where partial correlation is useful.
5. In multiple regression, which correlation (partial or part) gives unique contribution?

💡 Exam Tips

• The key distinction (partial = remove from both; part = remove from one) is the most important point.
• Always give a concrete example for each type.
• The comparison table is excellent for Section B answers.
• Questions sometimes combine "partial correlation" with a Spearman computation — prepare both independently.