For readers who what a copy of questionaire, it can be obtained by clicking here. 

The analysis was treated judge by character as the unit. Using the MIXED option in SPSS, both judge and character are treated as random factors. The variance due to character over the total variance (character, judge, residual plus error) was the measure of consensus. Analyses used an average of each Big Five trait.

We checked to see how much our judges agreed with one another. They all substantially agreed with each other, expect. The one who did not agree was my wife, Marina Julian. We exluded her data for the analyses presented in the paper.