Data Validity Exploratory Factor Analysis

[@ansumanabhisek]

Thank u..sir..very well explained

[@ansumanabhisek]

But 1 doubt..can we do factor analysis during pilot testing for 15-20 sample size?

[@dr.yasirhayatmughalstatsfo7381]

For factor analysis there are few rules of thumb given by Field (2013). for sampling adequacy KMO is suggested. Range of KMO is 0-1. and 0.5 is acceptable above 0.5 is considere medicore and 0.7-0.9 is excellent. for further reading about sample size i suggest Field (2013) chapter 17 Exploratory factor analysis page no 646-647 is all about sample size and its adequcy i.e. KMO third edition

[@ansumanabhisek]

@@dr.yasirhayatmughalstatsfo7381 thank u sir..let me read it..

[@jernaf7340]

Hello sir! Is there a way that we can email you for some inquiries if it is ok?

[@princessniq7820]

Is CFA all about the data? I mean if you have 5-likert scale and you get more 5 in each question. Is it considered a nice data?

[@dr.yasirhayatmughalstatsfo7381]

Can you explain your question again?

[@princessniq7820]

@@dr.yasirhayatmughalstatsfo7381 sir, my data has terrible RMSEA..that's what it says. What was wrong with it? Do I have to conduct or administer another batch of respondents? Thanks.

[@dr.yasirhayatmughalstatsfo7381]

@@princessniq7820 try unweighted least squares (ULS) and diagonally weighted least squares (DWLS) based on polychoric correlation matrices, it will give you least RMSEA

[@princessniq7820]

@@dr.yasirhayatmughalstatsfo7381 thank you so much.. such a great help. more power and godbless. Greetings from Philippines.

[@jernaf7340]

Our school statistitian told us our data for cfa is poor and not a good fit as well. We cannot figure out what is wrong since he doesn't explain us clearly. Is there any way or website to test our data?

[@dr.yasirhayatmughalstatsfo7381]

Yes all data not fit for cfa, what likert scale you have used? And how much os your sample size and whether you run cfa in Amos or Pls?

[@jernaf7340]

@@dr.yasirhayatmughalstatsfo7381 sir we have 4-likert scale with 400 sample size,since our statement indicator is 39. Our statistitian told us he used SPSS.

[@dr.yasirhayatmughalstatsfo7381]

@@jernaf7340 first of all 5 likert or for better result especially SEM 7 likert is recommended bcoz 7 likert helps in explaining variance. I suggest you use 7 likert scale and recollect data it will be much better. If u use spss then 5ikert scale also works. But better use 7 likert scale

[@jernaf7340]

@@dr.yasirhayatmughalstatsfo7381 thank you sir we will try to do that.

[@jernaf7340]

Our school statistitian told us our data for cfa is poor and not a good fit as well. We cannot figure out what is wrong since he doesn't explain us clearly. Is there any way or website to test our data?