Respiratory rate (RR) is one of the most sensitive markers of a patient condition and a core aspect of multiple clinical assessment tools. Doctors use a number of methods to assess RR, including formal measurement, and ‘spot’ assessments, although this is not recommended. This study aimed to assess the accuracy of the methods of RR measurement being used by doctors. A cross-sectional study assessing the accuracy (range, bias, and imprecision) of doctors’ ‘spot’ and ‘formal’ respiratory rate assessments, using videos of mock patients. 54 doctors in a London teaching hospital participated. Both methods showed high levels of inaccuracy, though formal methods were more accurate than ‘spot’ assessments. 52 and 19 % of doctors did not identify the respiratory rates shown as abnormal, using ‘spot’ and formal assessment methods respectively. We observed a trend towards decreasing accuracy of ‘spot’ assessments with increasing clinical experience (p = 0.0490). Current methods of RR assessment by doctors are inaccurate. This may be significantly delaying appropriate clinical care, or even misguiding treatment.