Hi,
I have been using CFA for a long time. I have been using both SPSS and AMOS. The main purpose of using CFA was to obtain more parsimonious dimensions related to variables in the model. My experience of using CFA is mixed. At times, I get desired loadings and at times, values of loadings in CFA remained low. May I know the reasons underlying reasons that influence values of loadings in CFA. I heard that taking a larger sample size may help in improving the values of loadings. Is it correct? If yes, what is the underlying mechanism that influences these values.
My experience of using CFA is mixed. At times, I get desired loadings and at times, values of loadings in CFA remained low. May I know the reasons underlying reasons that influence values of loadings in CFA. I heard that taking a larger sample size may help in improving the values of loadings. Is it correct? If yes, what is the underlying mechanism that influences these values?
My experience of using CFA is mixed. At times, I get desired loadings and at times, values of loadings in CFA remained low. May I know the reasons underlying reasons that influence values of loadings in CFA. I heard that taking a larger sample size may help in improving the values of loadings. Is it correct? If yes, what is the underlying mechanism that influences these values?
Please enlighten
Hi Aziz, You have given detail but still some more information is required. For instance, AMOS require at least 3 items for each construct to be retained (unlike PLS which works on a single item). If you got factors loading less than desire values-look into those items having low factor loading if each constructs has more than 3 items. Retain those 3 items having desired FL.
Yes sample size in some cases also influence your factor loading.
In addition, AMOS require normal data, if you afraid about factor loading, you can check data normality prior and exclude those respondents having abnormal distribution, it may positively influence your CFA.
Regards
Standardized loadings in CFA (AMOS) will be similar to those you will obtain if you use EFA in SPSS, maybe a little higher, as you are setting the small loadings in EFA to 0. That's all!
Hi Aziz, You have given detail but still some more information is required. For instance, AMOS require at least 3 items for each construct to be retained (unlike PLS which works on a single item). If you got factors loading less than desire values-look into those items having low factor loading if each constructs has more than 3 items. Retain those 3 items having desired FL.
Yes sample size in some cases also influence your factor loading.
In addition, AMOS require normal data, if you afraid about factor loading, you can check data normality prior and exclude those respondents having abnormal distribution, it may positively influence your CFA.
Regards
Hi Dear
I am sorry to late in answer because I was busy, I think that the large sample and variable as well will solve your statistical problem
The factors loadings are estimates of population values, and don't necessarily have to be high. Increasing the sample size will help with the precision of the estimates but not necessarily increase the magnitude of associations.
- Badges
- Science topic
- Similar topics
- Philosophy
- Philosophy Of Science
- Reasoning