Hello,
Using EFA I developed a social scale with 4 factors each having 4 items. In a new round of data collection I collected three separate samples (150 each) to test if my new scale can show the differences in social ratings between three different product categories. My question (considering the results belong to three different types of products with various levels of social features): 1- Can I use the 450 responses combined for CFA? 2- Should I use them separately (i.e. three CFAs)? What if the model doesn't fit for one of them? Thank you
Dear Ehsan Mortazavi,
From my point of view: yes, you can use both samples together and conduct an invariance analysis between A-B, B-C and A-C. It will answer you if the instrument does not fit any sample. The point is: one thing is knowing whether the general structure of the instrument fits the sample. The second is related to the endorsement magnitude of the sample of each item and factor. Generally speaking, the instruments used to fit different samples, unless they are different from each other (as in the WEIRD discussion).
Here is some material that could help regarding invariance:
Hirschfeld, G., & Von Brachel, R. (2014). Multiple-Group confirmatory factor analysis in R – A tutorial in measurement invariance with continuous and ordinal. Practical Assessment, Research & Evaluation, 19(7), 1–11.
Fischer, R., & Karl, J. A. (2019). A Primer to (Cross-Cultural) Multi-Group Invariance Testing Possibilities in R. Frontiers in Psychology, 10. https://doi.org/10.3389/fpsyg.2019.01507
Lugtig, P., & Hox, J. (2012). Developmetrics A checklist for testing measurement invariance. European Journal of Developmental Psychology, 9(4), 486–492.
Some material regarding WEIRD:
Henrich, J., Heine, S. J., & Norenzayan, A. (2010). The weirdest people in the world? Behavioral and Brain Sciences, 33(2–3), 1–75. https://doi.org/10.1017/S0140525X0999152X
Muthukrishna, M., Bell, A. V, Henrich, J., Curtin, C. M., Gedranovich, A., McInerney, J., & Thue, B. (2020). Beyond Western, Educated, Industrial, Rich, and Democratic (WEIRD) Psychology: Measuring and Mapping Scales of Cultural and Psychological Distance. Psychological Science, 31(6), 678–701. https://doi.org/10.1177/0956797620916782
I hope it will help.
Hello
What was your goal for sampling in four different groups? If your goal is to determine the difference or effectiveness in four different situations; In this case, for each sample, while fitting the data with the model, you can examine the regression coefficients; Examine the variances and covariances between the variables and then interpret the results. But if your goal is to assess the validity of the structural model; Use multi-group structural modeling techniques to evaluate the validity of the model in different groups.
Note: your question title is not match by your question content.
EFA in title
CFA in content
EFA & CFA are different approaches of Factor analysis
L. Sibanda That is absolutely correct. I performed EFA with a sample size of 229 and reduced items to 4 items for 4 factors.
I collected another 450 samples from 3 different products (150 each) that I expect to have different values for the four factors. I wanted to know if can use the 450 responses combined in a single CFA or if I have to conduct 3 separate CFAs.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude