This paper presents results from an experimental investigation of the accuracy and precision of various measurement procedures for determing sound‐power output of ’’large’’ machines in the free field over a reflecting plane out of doors. The purpose of the investigation was to place error bounds on several proposed nearfield measurement procedures, chiefly through the comparison of such properties with sound power levels determined from farfield measurements. The sources used in the study included 17 portable air compressors of various types and sizes powered by internal combustion engines. The data analysis centers on the comparison of sound power levels estimated from measured sound pressure levels on two measurement surfaces, one at 7‐m radius (farfield) and a second at 1 m from the surface of the machine (nearfield). Empirical estimates of precision and accuracy are derived for each of several proposed ISO procedures for determination of sound power level. The nearfield measurements were found to produce an overestimate of the farfield power level with the magnitude of the overestimate depending on the measurement surface shape. A measurement surface shape which ’’conformed’’ to the shape of the source was found to have the smallest associated overestimate. Microphone directivity was found to lead an underestimate of sound power level.