Abstract
Researchers in the behavioral and social sciences use linear discriminant analysis (LDA) for predictions of group membership (classification) and for identifying the variables most relevant to group separation among a set of continuous correlated variables (description). In these and other disciplines, longitudinal data are often collected which provide additional temporal information. Linear classification methods for repeated measures data are more sensitive to actual group differences by taking the complex correlations between time points and variables into account, but are rarely discussed in the literature. Moreover, psychometric data rarely fulfill the multivariate normality assumption. The article introduces a dataset consisting of 160 one-minute affective video clips with normative values of valence and arousal. Each video was evaluated by 30 subjects, while each subject evaluated at least 20 videos. Compared to previous attempts to collect affective videos, the dataset has several advantages. Firstly, the high number of videos in different valence categories allows researchers to compile appropriate subsets for their studies. Secondly, the approximately equal and conventional duration of videos makes it possible to use them in psychophysiological studies applying EEG, fMRI, peripheral polygraphy, posturography, TMS, etc. Thirdly, the exclusion of sound or speech that might provoke culture-dependent interpretation makes the dataset useful in different cultures. The relationship between valence and arousal showed a typical quadratic pattern, with very negative and very positive videos receiving higher levels of arousal. Several negative videos received greater arousal scores than the most positive ones, reflecting negativity bias. The dataset encompasses more than 50 videos of different valence (negative, neutral, and positive ones). We believe that it will permit researchers to select corresponding subsamples of videos from different categories for their studies.