Does Contrast or Comparison Help More? The Role of Learning Mode and Category Type
Schuyler deVos ’17 and Janet Andrews (Cognitive Science)
Recent research has brought up the question of which methods are the most effective in helping people learn categories. Certain advantages have been suggested for observational learning methods using within-category comparison as well as for traditional classification learning using between-category contrast. Understanding the methods through which people best learn categories may help us determine the best way to teach them. We used previously created stimuli resembling micro-organisms and presented participants with sets of three stimuli at once on each trial in order to train them in identifying members of each of three categories. Our study used four modes of learning the categories, crossing learning method – observational and classification with feedback – and type of item triple – between-category (one item from each category) and within-category (all three items from one category) – in order to investigate the role of these factors. After the training phase, we tested participants by presenting them with stimuli singly in order to gauge how well they had learned each category. The experiment was conducted online using Amazon’s Mechanical Turk and the results revealed a significant effect of triple type, with between-category learners doing better on the test portion of the experiment than within-category learners regardless of learning method. It is possible that the within-category advantage reported previously for observational learning depended on the fact that the categories being learned were relational and we will probably test that directly in a follow-up study. We would also like to explore why a bimodal pattern occurred in all groups, with classification performance on the test either excellent or near chance.