Cramér’s V

Definition:
Cramér’s V is a statistical measure of association strength between two categorical variables.
It is based on the Chi-square test statistic, but it normalizes the value so that the result always lies between 0 and 1.

• 0 → No association (completely independent variables).
• 1 → Perfect association (one variable fully explains the other).

---

Formula:

$V = \sqrt{\frac{\chi^2}{n \cdot (k – 1)}}$

Where:

• $\chi^2$ = Chi-square statistic
• $n$ = total sample size
• $k$ = the smaller of (number of rows, number of columns)

---

Interpretation (rule of thumb):

• 0.00 – 0.10 → Very weak association
• 0.10 – 0.30 → Weak association
• 0.30 – 0.50 → Moderate association
• > 0.50 → Strong association

(Note: thresholds can vary depending on the field.)

---

Example:
Imagine you survey 1,000 people about:

• Gender (Male/Female)
• Preference (Coffee/Tea)

After building a contingency table and calculating Chi-square, you compute Cramér’s V = 0.25.
This means there is a weak-to-moderate association between gender and drink preference.

---

Use in Data Science / ML:

• To detect categorical drift: compare category distributions over time.
• To check redundancy between features (e.g., if two categorical variables are too strongly associated, one may be dropped).
• To test feature-target association in classification tasks.