Auditing standards recommend separate assessments of the likelihood and magnitude of risks (hereafter, LM decomposition). Prior research shows that decomposition can focus individuals on the components of a judgment and make them more sensitive to information. An experiment with 101 experienced auditors shows that LM decomposition leads auditors to be less concerned about high-risk fraud schemes relative to auditors who make holistic risk assessments. Our analyses also show that, relative to those making holistic risk assessments, the correlation between auditors’ likelihood judgments and their overall fraud risk judgments and the coherence of their fraud risk judgments are higher for auditors who perform an LM decomposition. Two follow-up experiments with students replicate these findings for higher-risk events, and (unlike the auditor experiment) we also find that LM decomposition results in lower risk judgments for lower-risk issues. We also find that LM decomposition mitigates the influence of affective responses on high-risk judgments.