Abstract: The present study investigated how people combine covariation information and pre-existing beliefs in causal reasoning. The results were as follows: (1)Subjects did not add up this two sources of information simply, nor did they use a threshold for covariation information. They considered beliefs first, then judged whether the covariation was consistent with their pre-existing beliefs. When these two sources of information collided, they would reconsider their pre-existing beliefs. (2) The effects of beliefs were larger than those of covariation when subjects changed their judgements. These changes were larger when the candidate cause was unbelievable.

Key words: causal judgment, pre-exiting beliefs, covariation