Reader Comments

Post a new comment on this article

Reliability and validity of the Big-5 personality variables included in this manuscript

Posted by phanel152 on 18 Mar 2017 at 12:40 GMT

Recently, Ludeke and Larsen (2017, http://dx.doi.org/10.1016...) have found that the 10-item Big Five Inventory provided in the World Values Survey (WVS), and which we used in this study, is "quite unreliable". For this, they correlated the two items of each dimension across all 25 countries and found that the correlations differed more or less unsystematically across countries and were on average around zero.
While conducting the analysis we noticed low reliabilities of this personality measure, but ignored them, because some of the overall statistics were in the line with the literature. For example, students were less conscientiousness than the general public (d = -.25, see Table 1 in the manuscript). Also, we trusted the organizers of the World Values Survey to provide high quality data (in fact, all other variables we have reported are clearly more reliable). To get a better picture of the reliability and validity of the BFI-10, we extended the analysis of Ludeke and Larsen (2017) across countries and found that the BFI-10 is also not highly valid. For example, previous research has found that openness is negatively correlated with age and agreeableness positively (Srivastava et al., 2003, JPSP). However, in the WVS data, the correlations were across countries unsystematically.

In summary, we recommend to ignore our results regarding the Big-5 as a precaution. However, the other seven variables included in our analysis are reliable and can be interpreted as described in the article. We apologies for any inconvenience caused.

Competing interests declared: I am one of the authors