Abstract: An experiment was conducted to investigate human causal judgments in the concentrative presentation of information with the three different external representations and the paradigm of direct estimates of causal power, and to test the Power PC theory, Probabilistic Contrast Model, and pCI rule. The participants were 287 college students. It was found that judgments of causal power of single causal relationships had the four characteristics: (a) More judgments were in accord with the Power PC theory in preventive causes, but conversely most judgments were in accord with probabilistic contrast model in generative causes; (b) the representations of literal presentation, tables and graphs had no impact on the judgments of causal power in generative causes, but had the impact on that in preventive causes where more participants’ judgments were in accord with the Power PC theory in the representation of graphs, compared with the representations of literal presentation and tables; (c) no participant used pCI rule; and (d) there were obvious individual differences in causal judgments.

Key words: estimates of causal power, Power PC theory, Probabilistic Contrast Model, pCI rule concentrative presentation, literal presentation, tables, graphs