First, I posted this as an answer to my original question. I am not sure if it will be viewed since it is considered an answer instead of a question so I am posting it as a new question. Included at the end of this post is the link to the original question.
I am including an example Strains Table with some questions that need clarification. These questions may seem basic so I apologize in advance.
Would the combination of (predictor variables) PV1+PV2+PV3 = a new PV of PV4?PV1 + PV2 + PV3 = PV4 (Cumulative Strain)
The presence of all three variables equate to cumulative strain (PV4) and the lack of one or more of the variables would not equate to cumulative strain (Other). If PV4 consists of a large % then it indicates a strong correlation with heinous crime.
Which analysis would be used to add all three PVs to determine the number and % of subjects (participants) that equate to PV4 or would that have to be tallied by hand - a YES/NO situation? After PV4 and Other are ascertained which analysis would be used to determine the strength of the PV4 and Other to heinous crime.
Or is cumulative strain a second dependent variable? The cumulative strain is dependent on the presence of the 3 PVs… If PV1+PV2+PV3 = YES or 1 – There is the presence of Cumulative Strain
If any of the PVs is no or 0 then PV1, PV2, PV3 = NO or 0 – Cumulative Strain is not present. Then would Cumulative strain become a predictor variable (Other would be the second predictor variable) since the prediction is that the presence of all three strains results in cumulative strain which correlates to heinous crime.
Hopefully, this is not terribly confusing!!
Link to original question: https://www.researchgate.net/post/Type_of_analysis_for_three_IVs_combine_and_correlate_to_DV
Thank you in advance for your help!
As I understand, your 'Strain Table Example2.jpg' is for a part of the population of 50 mass murderers. In order to be able to test the association between 'cumulative Strain' and 'heinous crime', you need to known the rate of 0/1 'cumulative Strain' in the general population. If you can get (a good guess of) this prevalence of 'cumulative Strain' in the general population (of non-murderers), you can apply a chi-square test.
I believe that you may have a very basic, fundamental problem with your design. Specifically, since you sample consists entirely on people who have committed a heinous crime, you have no variance in your dependent variable. It is not a variable -- it is a constant. Mr. Feys comment addresses this by providing a means of generating some contrast, and refocusing your analysis solely on comparing strain levels between two groups -- one known to be mass murderers and the other with a very low (effectively zero) rate of murder.
- Badges
- Science topic
- Similar topics
- Psychology