Multivariate exploratory data analysis: bias assessment.

The objective of this investigation was to evaluate if the bias of the subj ects varied during a stimulus classification task in relation to the presen tation format (suns and stars) and to the way the variables were assigned ( randomly, ordered according to their correlation in 360 degrees or using a biplot: factorial suns or stars). It was found that there was a significant interaction between the presentation format and the method of assigning va riables. In particular, it was found that the subjects made fewer mistakes classifying the factorial suns than the factorial stars. Likewise, they als o classified better the ordered suns than the ordered stars. There were no significant differences between the suns and stars when no information rega rding the correlation was included (random order) nor between the ordered s uns and factorial suns. Lastly, it was also found that the subjects took lo nger to complete a task with random assignment than one with the factorial representation.