Computational protocol: Rating of personality disorder features in popular movie characters

Similar protocols

Protocol publication

[…] Inter-rater agreement was calculated through random effects analysis of variance. Intraclass correlations were calculated as the proportion of variance unique to each movie character relative to the total variance in a given scale. This measure of agreement is equivalent to kappa in interpretation []. A limitation to the ICC is that it is highly affected by variance, because if the total variance is small, then the unique variance of each rated target must necessarily be even smaller. This is similar to the way that the kappa statistic is limited by low base-rates when calculating agreement.Analysis of variance was used to assess the multivariate and univariate difference between characters, using the SPSS GLM multivariate ANOVA module. Both the movie character and the rater were entered as factors in the model, and the scales from each instrument were then entered as dependent variables in separate analysis. The interaction between the two was not entered (as that would have resulted in 32 cells with n = 1). Bonferroni adjustments were made for all p-values for ratings to adjust for family-wise type 1 error (with 28 tests of inter-rater reliability, 3 multivariate and 25 univariate, all p-values were multiplied by 28).Differences in the experienced difficulty of rating the characters were also analyzed using analysis of variance. Contrast analysis was reported for linear trend, and Bonferroni post hoc comparisons of the difficulty of the movies.Graphs and partial intraclass correlations were produced with STATISTICA for Windows, V. 6.0 [], and ANOVA was calculated on SPSS for Windows v. 11.5 []. […]

Pipeline specifications

Software tools SPSS, Statistica
Application Miscellaneous
Organisms Homo sapiens