People's causal judgments show systematic biases, including over-weighting of confirming information, and favoring prior beliefs over data. We investigated the effects of self-construal on data-weighting in causal judgments. We primed participants with interdependent or independent self-construals (interdependent people define themselves through relationships; independent through individual traits). On 56 trials, each containing complete frequency information (i.e., frequency of presence/absence of the cause and presence/absence of the effect), participants judged the ability of a cause to produce an effect. We observed a main effect of prime: Interdependent participants' judgments covaried more with frequency information, suggesting they make greater use of data. Furthermore, on non-contingent trials, interdependents weighted all frequency data equally (independents did not). However, being interdependent did not perfectly ameliorate biases in the weighting of data: on contingent trials, both groups weighted the data unequally, favoring confirming, cause-present information. Thus, independents only sometimes demonstrated extra sensitivity to background information.