My sample size is too small to conduct a factor analysis. I created my own psychometric scale for measuring Effectiveness of Support, which contains 38 items (4 sub-scales)
Im really struggling to analyse my data from such a small sample size and was wondering how i could go around this or if there are any alternatives?
Ooof, ok, that is really small. I actually doubt that you can anything reasonable with such an N (not even calculating a correlation coefficient. The reason is simply that the unreliability of the sample estimate is so large that everytime you would draw another 28, you would get completely different results.
There is only one solution, namely to collect additional data....
Best,
Holger
Hello Chantelle,
you did not tell how large your sample is :)
Best regards,
Holger
Ooof, ok, that is really small. I actually doubt that you can anything reasonable with such an N (not even calculating a correlation coefficient. The reason is simply that the unreliability of the sample estimate is so large that everytime you would draw another 28, you would get completely different results.
There is only one solution, namely to collect additional data....
Best,
Holger
I am really struggling at the moment probably to do with the strict criteria and short time frame to collect data! But thank you for responding :)
Chantelle Fraser, I think that the comments from Holger Steinmetz are on the ball. Maybe, however, you could look carefully at interitem correlations to obtain some idea about the extent to which your items are associated with each other.
Dear Imran Anwar ,
please read also the posts in a thread. The sample size was discussed.
https://www.researchgate.net/post/Netiquette_proposal_for_questions_on_structural_equation_modeling
Best,
--Holger
I completely agree with @Holger. Your sample is too low and as soon as more data is added, the results would be completely different.
Dear @Holger As soon as I went through the post's thread I deleted my comment. 🙂
Dear Chantelle Fraser you can choose SEM-PLS for any number of sample. The most effective statistical tools for data analysis.
Chantelle Fraser your sample size is too small even for SEM. I don't why Mahadi has suggested you so. There is a sample requirement even for SEM in SmartPLS.
You will need SmartPLS to run PLS SEM nonetheless!
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size