Dear Experts,
I am a beginner in AMOS learning. I am confused about results reporting. Like when I run the model, its regression estimates are significant but when I apply to BCCI bootstrap with subsample 5000 at 95%, the standard error increased and results become non-significant. To reveal the hypotheses results, which results should be reported either simple (significant) or bootstrapped (non-significant)?
Need your guidance
Thank you
@Asif Ali If your model doesn't test the mediation then there is no need to use bootstrapping while running the model. However, from my personal experience, bootstrapping doesn't affect the results at all and the results remain intact. Don't know what's happening with your model when you use bootstrapping. It would be better had you attached the screenshot of regression coefficients table with and without bootstrapping.
Imran Anwar thanks for your response. Yes, I checked in both ways. When applying to bootstrap, standard error increased and some relationships become non-significant. Although, all other parameters are perfectly alright.
Zhenwei Dai
Thank you so much for your detailed response. I appreciate your findings and agreed your point of concern, but I did not find any abnormality in data because I already removed over 150 responses just to normalize the data.
Zhenwei Dai
I highly appreciate your guidance. The sample size is 577. I will try to solve this issue soon. Thank you so much
The significant increase in standard error may result in the t-value crossing the threshold value, in that case, check BCCI upper and lower limit. If these both provide evidence for the results to be insignificant, then check for missing data and normality
- Badges
- Science topic
- Similar topics
- Counseling Psychology
- Confusion