This is how i've worded it within my 'Discussion' section (obviously with 'dropped' zero's on the nulls, that I can't write out properly here)...
''This research set out to investigate whether grandparenting was influential in child outcomes, and was able to predict adolescent adjustment difficulties, through a multiple hierarchical regression. Results indicated a small but very significant predictive effect of grandparenting variables (p =.001) when controlling for demographics, adversity, and SEM. This accounted for 1.5% ∆R² with the entire model predicting 12% R2 and significance through ANOVA (p
Generally in SEM we do have 11 or more variables.So we generally present each relation between two variables (hypothesis) in a tabular form.The first column would be the statement of the hypothesis.The second column will have the test and its result statistics and the third column will have the rejected or accepted status for the hypothesis.
In the discussion we generally talk of whether the model is fit or not.
Thank you Jaspreet, that has given me something to think about. My research is based within the context of a multiple hierarchical regression rather than SEM. I assume the presentation is different?
A table would be best for reporting a complex model. What was the model you tested? Do all 14 variables create the grandparenting construct? The format of the table(s) would depend on how you ran your model
Olivia Cody - Hi Olivia, and thankyou for your response. I have an APA format table that presents results for the hierarchical regression, already. I am writing the discussion at the moment. I have stated that I reject the null for the overall model and now need to reject most other nulls and accept two. All 3 models were significant, as follows;
Step 1 - Age (catagorical dummy; 0,1,2,3,4,5)/Sex (catagorical dummy; 0,1) (R squared = .016, p
Personally I find saying "Ho1 was rejected/failed to rejected" to be completely unnecessary outside of educating students. (Just a note since you had "accepted", statistically speaking you can't accept the null hypothesis, only fail to reject the null. The difference being you are not staying the null is true)
It would be easier on you (and the reader) to have something like "There was a significance difference between those who were 12yr compared to those who were 16yr where ...explain difference... (beta = #, p = #). All other contrasts between age groups were non-significant (p > .05)." [or what ever your alpha level is set to]
If you are have required to have the Ho1 was failed to rejected/rejected, then the last sentence could be "All other contrast between age groups were non-significant (p > .05), thus we failed to reject these null hypotheses."
Dose that answer your question?
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Techniques
- Structural Equation Modeling