After myocardial infarction, left ventricular volume and ejection fraction can be assessed by echocardiography, magnetic resonance imaging and radionuclide angiography to guide therapy and determine prognosis. Whether a measured parameter gives the same results irrespective of the method used and the observer who performs the analysis is only partly known. Intra-observer and inter-observer variability were determined for echo and magnetic resonance imaging. Left ventricular ejection fraction measured by these techniques was related to radionuclide angiograms performed in the same period. Intra-observer variability for both echo and MRI was low and in most instances below 5%. Inter-observer variability for the echo and MRI measurements were substantially higher than intra-observer variability. Comparison of the three imaging modalities revealed systematic differences. Therefore, in clinical studies, left ventricular volume and function parameters have to be measured with the same technique and by the same observer in qualified core laboratories.