Dear researchers,
I've been working on two different scales for two separate studies. I'm not getting the expected results from Exploratory Factor Analysis (EFA). Confirmatory Factor Analysis, on the other hand, provides me with results that confirm the construct's validity. Given this, do I need to perform EFA in the following cases?
1. A study of scale development based on Bandura's (1977) self-efficacy model. Bandura theorizes that efficacy expectations vary on several dimensions, including magnitude, generality, and strength. As a result, I accepted these dimensions as scale sub-factors and wrote some items about them.
2. A scale development study based on curriculum learning objectives. There are four units in this curriculum. Each unit has its own set of learning objectives. I wrote some items around these learning objectives. I wanted to learn students' perspectives on how well they met these learning objectives, so I accepted each unit as a dimension (sub-factor) of the scale in CFA...
Source: Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change. Psychological review, 84(2), 191.
In my opinion, to test the theoretical/hypothesized factor structure of a scale that you developed, you do not need exploratory factor analysis (EFA). EFA is mostly for exploring the number of factors. In your case, you would have clear a priori assumptions about the number of factors and the loading pattern (i.e., which item/variable is supposed to measure which factor). Therefore, confirmatory factor analysis is the method of choice to test the hypothesized factor structure.
Dear sir, in any social science study or more specifically, in study on human behavior it is very important to carry out EFA first in order to identify the 'latent variables'. This is because every published theory were empirically tested on their area or country or locality and every locality, country etc differs greatly with various demographic factors such as income level, perception about any phenomena. So it is always better to carry out EFA first, then after identifying Latent Variables, carry out a CFA so that framework or proposed model can be validated (convergent and discriminate) and also other fit indices. after that conduct hypothesis.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Regression