Point Biserial & Phi Coefficient — Previous Year Questions

📊 Topic Overview

Metric	Value
Papers Appeared	11 of 27 (40.7%)
Priority	🟠 HIGH
Section Distribution	Sec A: ~6
Most Repeated	"Explain point biserial correlation and phi coefficient with suitable examples."

🎯 Must-Know Concepts

1. Point biserial correlation (r_pb): continuous variable vs genuinely dichotomous variable
2. Phi coefficient (φ): two genuinely dichotomous variables
3. Formulas and computation for each
4. When to use each vs Pearson's r and biserial r

---

🔥 Most Repeated Question

Common phrasings:

• "Explain point biserial correlation and phi coefficient with suitable examples."
• "Explain step by step the calculation of point biserial correlation and phi coefficient."
• "Describe point-biserial correlation and phi-coefficient."

Answer (Section A or B — ~350 words):

Special correlations are used when one or both variables are dichotomous (have only two categories). The choice between them depends on whether the dichotomy is genuine (true binary, e.g., pass/fail, male/female) or artificial (a continuous variable forced into two categories, e.g., above/below average).

Point Biserial Correlation (r_pb):

Used when one variable is continuous (interval/ratio) and the other is a genuine dichotomy (naturally two-category variable).

Examples: Correlation between exam score (continuous) and pass/fail status (genuine dichotomy); correlation between reaction time (continuous) and gender (genuine dichotomy).

Formula: $r_{pb} = \frac{\bar{X}_1 - \bar{X}_0}{SD_t} \sqrt{\frac{p \cdot q}{1}}$

Where:

• X̄₁ = mean of the continuous variable for one group (coded 1)
• X̄₀ = mean of the continuous variable for the other group (coded 0)
• SD_t = standard deviation of the total group on the continuous variable
• p = proportion in group 1, q = proportion in group 0 (p + q = 1)

Interpretation: r_pb ranges from 0 to 1 (always non-negative or negative depending on coding). Interpreted like Pearson's r.

Phi Coefficient (φ):

Used when both variables are genuine dichotomies.

Examples: Correlation between gender (male/female) and treatment outcome (success/failure); passing/failing two different tests.

Data is arranged in a 2×2 contingency table:

	Y = 1	Y = 0	Total
X = 1	a	b	a+b
X = 0	c	d	c+d
Total	a+c	b+d	n

Formula: $\phi = \frac{ad - bc}{\sqrt{(a+b)(c+d)(a+c)(b+d)}}$

Relationship to Chi-square: φ² = χ²/n (phi squared equals chi-square divided by n).

Interpretation: φ ranges from −1 to +1, interpreted like Pearson's r.

Choosing the Right Correlation:

Variables	Use
Both continuous, linear	Pearson's r
Continuous + genuine dichotomy	Point biserial r_pb
Continuous + artificial dichotomy	Biserial r_b
Both genuine dichotomies	Phi coefficient φ
Both artificial dichotomies	Tetrachoric r_t

---

🧠 Memory Aid

• Point biserial: one Point (continuous), one Binary = r_pb
• Phi: Phone (two buttons) = two genuinely binary variables

---

📚 Related Content

• Special Correlations: Point Biserial and Phi
• Biserial and Tetrachoric Correlations

✅ Self-Assessment

1. When is point biserial correlation used?
2. Write the formula for phi coefficient.
3. What is the relationship between phi and chi-square?
4. How is phi different from point biserial?
5. Construct the 2×2 table for phi computation.

💡 Exam Tips

• The table of "which correlation for which data type" earns marks in Section A.
• Know the phi formula (ad − bc in numerator) — frequently asked.
• Phi and chi-square relationship (φ² = χ²/n) is a standard exam point.
• Distinguish clearly: genuine vs artificial dichotomy determines which correlation to use.