A widely-used correlation to obtain efficiency is considered, and the hypothesis of dividing a photovoltaic module into a number of sub-regions and calculating efficiency of each region employing the temperature of that part to determine the efficiency of the module more precisely is proposed and examined. Experimental data recorded during a year are employed to check the accuracy of the proposed hypothesis and the original method (suggestion of the 61215 standard, where one temperature is considered for the whole module). Comparison of the both hourly and monthly profiles of performance criteria shows that the proposed hypothesis is much more accurate than the original method. It is found that in July, as the sample month for which the hourly profiles are investigated, the average values of the error in prediction of efficiency, produced power, and generated energy are 5.70, 7.72, and 6.07% for the original approach whereas using the proposed hypothesis reduces them to only 1.45, 1.75, and 1.35%, respectively. Moreover, the values of annual average error in prediction of the three aforementioned performance criteria are improved from 4.14, 5.45, and 4.55% in the original method to 1.05, 1.25, and 1.09% for the proposed hypothesis, which is a huge achievement. (C) 2020 Elsevier Ltd. All rights reserved.