Check if your A/B test sample sizes show a statistically significant mismatch from the expected 50/50 distribution.