Information presented by one source often has a continued influence on judgments even after the information was retracted by another source (Continued Influence Effect; CIE). According to the model-updating account, the CIE occurs because the original information is integrated into a situation model, but this model is not updated after retraction. Hey et al. (under review) found that the credibility of the sources of original information and retraction antagonistically affected the CIE. The model-updating account predicts such source-credibility effects on the CIE only if source credibility is available during situation-model integration at encoding, but not if it becomes available only after integration is completed. To test this, we had 432 participants read a fictitious news report in which original information and retraction were presented by sources that had either high or low credibility. Source credibility was either disclosed before presentation of original information and retraction (before groups) or thereafter (after groups). As predicted by the model-updating account, the timing of source-credibility disclosure moderated the source-credibility effects on the CIE: In the before groups, higher credibility of the source presenting the original information led to a larger CIE, and higher credibility of the source presenting the retraction led to a weaker CIE. In the after groups, in which source credibility presumably could not affect the situation model, there were no source-credibility effects on the CIE. Thus, source-credibility effects on the CIE are contingent on situation-model integration, which supports the model-updating account.