Questionable necessity effects on suicidal ideation: a commentary on

Marchetti et al. recommended researchers of mental health to use necessary condition analysis (NCA). In an empirical application, they found statistically significant necessity effects of perceived burdensomeness and thwarted belonging on suicidal ideation. However, necessity effects in NCA do not prove that X is necessary for Y, as the effect could, for example, be due to a correlation between X and Y and this correlation could, in turn, be due to a confounding impact by a third variable Z. We scrutinized the findings and conclusions by Marchetti et al. by estimating "ranges of spuriousness," i.e., 95% confidence intervals of necessity effects that could be expected due to the correlations between suicidal ideation and perceived burdensomeness and thwarted belonging, respectively. The necessity effects fell within the range of spuriousness, meaning that they could be accounted for by the correlations between the variables. Hence, conclusions by Marchetti et al., that perceived burdensomeness and thwarted belonging are necessary for suicidal ideation, may be challenged. It is important for researchers using NCA to be aware of the method's severe limitations. We recommend scrutinizing findings by estimating ranges of spuriousness and to require that necessity effects are above this range if concluding that X is necessary for Y.