More confusion about deliberate practice: commentary on Miller et al. (2018)

Twenty-five years ago, Ericsson, Krampe, and Tesch-Römer (1993) introduced the concept of deliberate practice (DP), arguing “individual differences in ultimate performance can largely be accounted for by differential amounts of past and current levels of practice” (p. 392). In a meta-analysis (Macnamara, Hambrick, & Oswald, 2014), we found DP did not explain even most of the individual differences in performance. We concluded DP is important, just not as important as Ericsson et al. argued. In a High Ability Studies article, Miller and colleagues (Miller et al., 2018) claim that “although all 88 studies in Macnamara et al. (2014) were ‘interpreted’ by the researchers as DP, in reality, they were not” (p. 5). Miller et al. reanalyzed our dataset and report performance correlated more strongly with DP (.40; our correlation was .38) than with activities they deemed nonDP (.21). We credit Miller et al. (2018) for their efforts. However, it is unclear what the criteria for DP were in their reanalysis. Furthermore, Miller et al. miss the mark on a critical methodological point. We discuss these problems in turn.