This study investigated children's judgments of the epistemic status of justifications for causal claims. Twenty-six children (14 boys, 12 girls) between the ages of 8 and 10 were asked to help two story characters choose the "best reason" for believing a claim. The reasons included appeals to an authority, to a plausible causal mechanism, or to data. Authority and plausible mechanism justifications always supported the claim and were paired with data that either covaried in support of the claim or did not covary and so were ambiguous with respect to the claim. Most children appeared to have a loose ordering of the epistemic status of justifications with data being preferred, plausible mechanisms appealing and preferred to ambiguous data, and appeals to authority least preferable. The children's primary reason for preferring any justification was its credibility. The credibility of data, to these children, seemed to rest on its firsthand nature. We suggest that this preference for data is productive and that instructional attention can usefully be focused on the attributes of measurement and experimentation that make data credible. (Contains 4 figures and 1 table.)