Analyzing dependent data as if independent biases effect size estimates and increases the risk of false-positive findings.