A Taiwan test standard was established in 1989 using outdoor daily efficiency test methods. This test standard has been implemented for 12 years with satisfactory results. However, it was also found from field applications that the pattern of solar irradiation would affect the result of the performance test. In the present study, we used a distribution factor R-i defined as the ratio of the total irradiation in the morning to that in the afternoon to characterize this effect. R-i reflects the asymmetry of solar irradiation distribution in the morning and afternoon. A field study was carried out. The data collected from the daily efficiency tests were screened using the criterion of 0.5 less than or equal to R-i less than or equal to 1.6, in addition to the conditions, defined in the Taiwan test standard. Two commercial products separately located in latitude 23degrees N and 25degrees N were tested. Data scattering occurs without using the R-i criterion. If we adopt test data using R-i the results turn out to have a much better data correlation coefficient, from 0.915 to 0.969. The system characteristic efficiency eta(s)* changes significantly, from 0.479 to 0.514. There is a regulation that the commercial product should have a value of eta(s)* exceeding 0.5 in order to obtain a subsidy from the government in Taiwan. The performance test using the old standard is shown to result in a significant error, suggesting modification of the former test standard. (C) 2003 Elsevier Science Ltd. All rights reserved.