Hi dear all!

As quite à lot of us know conducting a EFA (exploratory factor analysis) permits to identify a latent structure in a data set (often Likert-scale data obtained by a paper-pencil or web-based questionnaire). This exploration mostly leads to a CFA (confirmatory factor analysis) to validate the questionnaire that was used to gather the data set. Most references, authors, etc. assume that CFA permits to ensure that the latent structure of the observation (data set) fits to a theoretical, conceptual structure hypothesized by the researcher ; conceptual or theoretical structure that was used to construct the data gathering instrument (often a Likert-scale questionnaire). The validation process mostly begins with a AFE and ends with a CFA. My question is the following: is it possible, if one has strong theoretical arguments to validate the hypothesised latent structure of the questionnaire, to bypass the EFA stage in the validation process? Does anyone have references (articles, books, etc.) to support (or reject) the EFA bypass? Thanks for your responses, advises, etc.

Phil

More Philippe Wanlin's questions See All
Similar questions and discussions