We thank Dr Reza Pakzad and Dr Saeid Safiri for showing interest and giving comments on our research.1 They suggested that “the results of confirmatory factor analysis (CFA) may be overfitted and the validity of the studied questionnaire can become questionable when principal components analysis (PCA) and CFA are applied in the same dataset.” Those comments might contribute to the improvement of the questionnaire in the methodological and statistical analysis.
In fact, PCA and CFA could be used in the same dataset under the condition that CFA is used to assess model fit of the original structure of the questionnaire. It is well-known that CFA is used to test whether the measure of a construct is consistent with the researcher's hypothesized measurement model.2 In our study,1 CFA was implemented on the original four-factor structure of the Menopause-Specific Quality-of-Life (MENQOL) questionnaire,3 and the results showed acceptable goodness of fit of the four-factor structure. Meanwhile, five-factor structure of the MENQOL questionnaire was extracted by PCA. Furthermore, a similar analytic method has been employed to evaluate the psychometric properties of questionnaires in several studies.4,5
We also checked the study on Diabetes Quality of Life-Brief Clinical Inventory (DQoL-BCI)6 commented on by Dr Safiri7 and found that the CFA was used to test the goodness of fit of the factor structure extracted by PCA, which could lead to the overfitted issue. However, our study is totally different from this study. As for the overfitted issue, an emphasis should be put on whether CFA was used to test the original model of the questionnaire or not, instead of whether PCA and CFA were used in the same dataset or not.