Effect Size: Cohen's d · TechGradify

With enough users, even a tiny difference is "statistically significant". Effect size tells you whether it's big enough to matter. Cohen's d is the difference in means measured in pooled standard deviations:

d = \frac{\bar a - \bar b}{s_p}, \qquad s_p = \sqrt{\frac{(n_a - 1)s_a^2 + (n_b - 1)s_b^2}{n_a + n_b - 2}}

with sample variances (ddof=1). Implement cohens_d(a, b) and describe_effect(d), which returns "negligible" for $|d| < 0.2$ , "small" for $< 0.5$ , "medium" for $< 0.8$ and "large" otherwise. Raise ValueError if either group has fewer than 2 values or the pooled standard deviation is 0.