To examine the effect of clinical history on the electrocardiogram (ECG) interpretation skills of physicians with different levels of expertise, we randomly allocated to an ECG test package 30 final-year medical students, 15 second-year internal medicine residents, and 15 university cardiologists at university-affiliated teaching hospitals. All participants interpreted the same set of 10 ECGs. Each ECG was accompanied by a brief clinical history suggestive of the correct ECG diagnosis, or the most plausible alternative diagnosis, or no history. Provision of a correct history improved accuracy by 4% to 12% compared with no history, depending on level of training. Conversely, a misleading history compared with no history reduced accuracy by 5% for cardiologists, 25% for residents, and 19% for students. Clinical history also affected the participants' frequencies of listing ECG features consistent with the correct diagnosis and features consistent with the alternative diagnosis (all p values < .05). For physicians at all levels of expertise, clinical history has an influence on ECG diagnostic accuracy, both improving accuracy when the history suggests the correct diagnosis, and reducing accuracy when the history suggests an alternative diagnosis.