Responses to Daly's (2022) Criticisms of Nobes et al. (2019) In his recent Commentary, Daly (2022) makes many criticisms of Nobes et al. (2019). In our reply (Nobes, 2023) we focus on the substantive issues on which Daly and we agree and disagree regarding the magnitude and explanation of the Cinderella effect. Here we respond to his criticisms, first by discussing errors in Nobes et al. (2019), and then errors and misrepresentation in Daly (2022). In the third section we consider Daly's often combative language.1. Errors in Nobes et al. (2019) 1.1. Regarding whether the Canadian and US homicide data included non-cohabiting stepfathers (which would exaggerate the increased risk to stepchildren), we wrote that a "problem with all four studies suggests that increased risk might have been overestimated… the perpetrator data includes both cohabiting and noncohabiting fathers" (p. 1093, para. 5, emphases added). We should have said that, while with Daly and Wilson's (1994)
and ourBritish data this is definitely the case, for the same reasons the Canadian and US perpetrator data probably include both cohabiting and non-cohabiting stepfathers. Similarly, our point that "none of the previous studies (Daly & Wilson, 1994;Harris et al., 2007;Weekes-Shackelford & Shackelford, 2004) took account of the high proportions of nonresidential perpetrators" (p. 1099, para. 3) was overstated, as at this stage we cannot be certain. We discuss Daly's (2022) claim that the Canadian homicide data definitely did not include noncohabiting stepfathers below (2.2.1). 1.2. "Despite their extensive output on this issue over three decades, Daly and Wilson tested for only three possible confounding variables in only one study (1985)" (p. 1093, para. 8). In fact, they tested for four in that study (the other being personality traits, which they mentioned in the Discussion), and also for marital status in Daly and Wilson (2001). We should therefore have said that they have tested for only five possible confounding variables in only two studies. See the section "Confounding Factors" in the Reply.1.3. Another error was our claim that "There seem to be no selectionist grounds for predicting any substantive change in stepparents' increased risk with child age. Indeed, the only explanation they offer implies that the increased risk to older children would be even greater than to 0-to 4-year-old children" (p. 1099, para. 1). However, as Daly (2022Daly ( , p. 2973) points out, he and colleagues have suggested another possible explanation of the reduction in increased risk to stepchildren as they grow older, and so our statement is