This study examined the comparability of scores on the National Nurses Aides Assessment Program (NNAAP) test across language and administration condition groups for calibration and validation samples that were randomly drawn from the same population. A sample of 20,568 candidate responses to 1 test form was used. This examination is given in English or Spanish, with or without audio equipment assistance. Results show that factor structure validities of the NNAAP are well supported. Statistically significant chi square (or difference of chi square) statistics occur because of the large sample sizes. For this reason, it is frequently appropriate to conclude that a SEM model fits the data even if "p" is significant. The overall pattern of NNAAP data indicates a reasonable fit even when the chi square test suggests rejection of factor models when sample sizes are large. The evidence of fit holds for the calibration and validation samples for language and administration condition groups. Data suggest that the test is fair across administration groups. (Contains 3 tables and 27 references.) (Author/SLD)