Despite Leventhal et al. (2016) providing a 50-year overview of their Common-Sense Model of Self-Regulation (CSM), they fail to cite the null findings in relation to the CSM and adherence, as found by at least two reviews (Brandes & Mullan, 2014; Law et al., 2014).

Brandes and Mullan (2014) meta-analysed 23 datasets from 30 studies in chronically ill populations (26 studies concerned medication adherence) and assessed the CSM with adherence as the outcome. The results were stark, with effect sizes (r+) ranging from −0.02 [causal (95 % CI −0.17 to 0.16) and emotional (95 % CI −0.07 to 0.03) representations] to only 0.12 [treatment control (95 % CI 0.05–0.19) and personal control (95 % CI 0.06–0.18)]. Moderate to high heterogeneity was also evident for all dimensions apart from timeline, coherence and emotional representations, with funnel plots indicative of bias. These results are not supportive of the CSM for predicting adherence, in contrast to the conclusions of Leventhal et al., which cites other meta-analyses, but not this evidence or indeed that of Law et al. (2014).

Psychology has significant reproducibility issues (Open Science Collaboration, 2015), with substantial evidence of biased literatures (e.g. Donnelly et al., 2015; Ferguson & Heene, 2012; Open Science Collaboration, 2015). Ignoring the findings of well-conducted systematic reviews, in favour of selected, supportive studies, does not provide sufficient support for any theory (Ferguson & Heene, 2012; Ioannidis, 2005). It also reduces our credibility with other professions (Johnston, 2016; Open Science Collaboration, 2015). The CSM should be robust to meta-analytic investigations.