So say I have 4 variables. Independent Variables A and B, and dependent variables C and D. Also a categorical variable, say, age. In correlations, I see a significant relationship between A and D and also B and D. Logically I proceed to apply regression on these three values without including C. I see a strong impact on A on D and age on D but not that of B on D (R change is not significant in hierarchical regression for B).
To summarise:
Significant correlations: A-D, B-D, Age-D
Insignifant correlations: A-C, B-C
Significant effects: A on D, age on D
Insignificant effect: B on D
Now, do I include the IV B and DV C in the title because I used them in the study? Or do I exclude them because they did not produce significant result and I omitted C in the subsequent analysis on the basis of weak correlation?
(Note: I haven't specified my variables just to make the question simpler.)
Are any of these titles appropriate or is there another suggestion you would make
Title 1: effect of A, B and age on C and D
Title 2: C and D in relation to A, B, and age
Title 3: Effect of A and age on D
Usually you would not include them, but comment if you think there may still be a biologically (if you are a biologist) significant result, even though it is not statistically significant. For example, it may be that the p value is > 0.05 simply because of a small sample size.
Put aside the title for the moment, and concentrate on your _findings_ first.
What have you really found, and why is this so?
You get an anomaly in the influence of B on D in your two analyses.
You _must_ explain this. You have not changed the underlying data set.
So, is this an artifact caused by you introducing a categorical test???
In general, we report the positive findings in the title. Negative findings are uninteresting unless they are unexpected.
If B really is an important influence, you have to report it. If not, you relegate it to a footnote, to help future researchers from going down a dead-end street.
In the title one could add all the variables.This is because the study has made an effort to study all the variables in the model.Whethere you found it correlated or not is juts result.The title should show what you have studied.
- Badges
- Science topic
- Similar topics
- Education
- Higher Education
- Academic Writing