Lilienfeld (2007, Psychological treatments that cause harm. Perspectives on Psychological Science, 2, 53) identified a list of potentially harmful therapies (PHTs). Given concerns regarding the replicability of scientific findings, we conducted a meta‐scientific review of Lilienfeld’s PHTs to determine the evidential strength for harm. We evaluated the extent to which effects used as evidence of harm were as follows: (a) (in)correctly reported; (b) well‐powered; (c) statistically significant at an inflated rate given their power; and (d) stronger compared with null effects of ineffectiveness or evidence of benefit, based on a Bayesian index of evidence. We found evidence of harm from some PHTs, though most metrics were ambiguous. To enhance provision of ethical and science‐based care, a comprehensive reexamination of what constitutes evidence for claims of harm is necessary.