Matched-names analysis reveals no evidence of name-meaning effects: a collaborative commentary on Silberzahn and Uhlmann (2013).

Commentary In an article recently published in this journal (Silberzahn & Uhlmann, 2013), two of the authors of the present commentary found that Germans whose last name has a noble meaning, henceforth referred to as a noble surname , such as Kaiser (" emperor ") or König (" king "), were more likely to hold managerial positions than Germans with other last names. However, further data collection and new analyses, reported in this collabora-tive commentary, indicate that the apparent name-meaning effect is more likely attributable to name frequency. That is, these findings suggest that the effects reported previously should not be interpreted as evidence of a causal effect of names on career outcomes. Silberzahn and Uhlmann compared the share of Germans with noble surnames who are managers with the share of managers among Germans with the 100 most common surnames. This control group was selected to include surnames used throughout Germany, but this meant that noble surnames were compared with more frequent (i.e., common) names. Name frequency could affect the estimated relationship between name meaning and career outcomes if name frequency were correlated with variables that were, in turn, correlated with career outcomes, such as socioeconomic status, urban/rural residency, and religion. More important, it turned out that an idiosyncratic feature of the social-networking Web site from which the data originated, www.xing.com, created a mechanical channel by which the relative share of managers with low-frequency last names was overestimated (see Supplement 1 in the Supplemental Material available online). Silberzahn and Uhlmann sought to account for name frequency and other possible confounding variables using generalized-estimating-equation regression (in the original submission) and hierarchical linear modeling (in the final published article). Both approaches, however, assume that the effect of name frequency on the probability of being a manager is linear. The skew in the distribution of frequency of German last names, and the nonlinear mechanism by which the database overestimates the share of individuals with low-frequency surnames who are managers, allowed the relationship between name meanings and career outcomes to survive linear controls. Because of these issues, a matched-names analysis, in which each noble surname is compared with a set of similarly frequent names, is a better analytic approach than controlling for name frequency using regression or hierarchical linear modeling. For each noble surname, we identified a comparison set of 50 names that were similar in frequency (see Supplement 2 in the Supplemental Material). …