Metacognitive biases in mental health are not explained away by acquiescence and inattention

In a recent study, Sarna, Dar and Mazor (2025) present an important example of how biases in responding (inattention and acquiescence) have the potential to alter estimated associations between self-reported mental health symptoms and metacognition. In this reply, we argue that their paper goes too far. We illustrate that their assertions do not hold in newer datasets that use different recruitment methods and have reduced rates of inattention and acquiescence. We also show that their analysis of new experimental data suffers from methodological problems (including low statistical power and omission of key covariates), which when corrected, substantially undermines their key claim and instead indicates that, even in their own data, metacognitive biases remain linked to mental health symptoms after controlling for inattention and acquiescence.