Poor strength test results are sometimes not an indication of low concrete quality, but rather inferior testing quality. In a compression test, the strain distribution over the ends of the specimen is a critical factor for the test results. Non-uniform straining of a concrete specimen leads to locally different compressive stresses on the cross-section, and eventual premature breaking of the specimen. Its effect on a specimen can be quantified by comparing the compressive strength results of two specimens, one subjected to uniform strain and another to a specified strain gradient. This can be done with the help of a function that relates two parameters, the strain ratio and the test efficiency. Such a function depends on the concrete strength and cross-sectional shape of the specimen. In this study, theoretical relationships between the strain ratio and test efficiency are developed using a concrete stress-strain model. The results show that for the same strain ratio, the test efficiency is larger for normal strength concrete than for high strength concrete. Further, the effect of the strain gradient on the test result depends on the cross-sectional shape of the specimen. Implementation of the results is demonstrated with the aid of two examples.