Dear Researcher,
I am presenting the characteristics of a set of data, what I faced to struggling to analyse. So would you assist me by devising a solution and referring to the articles or books or lecture on YouTube?
Well, I would like to know the appropriate analysis tools for the following dataset:
1. Dichotomous Variable (Independent)
2. Dichotomous and categorical (Dependent Var.)
3.Follow up: highest 10 (every 4 months)
4. Participants were eliminated from the next visit who have completed or meet the set criterion.
5. No missing data were included.
Obj: Effectiveness of treatment. And Whom required less visit compare men and women (ratio- 3:1)
Above situation, what would be the best analytical process including a paramatric or non-parametric estimator that can be adopted to conclude?
Dear Mohammad Jobair Khan ,
Could you perhaps specify your actual main research question, and share actual descriptions of your variables, and a screenshot of your dataset - that would help.
Dear Elvis Munyaradzi Ganyaupfu ,
Thank you for your reply. Actually, it is difficult to do so at this moment. But if you have any specific question I am readily available to answer them.
THANK you again.
Regards,
Khan
Dear Mohammad Jobair Khan ,
Noted with thanks - please advise on the following:
(a). your main research question, and
(b). specific names of your dependent and independent variables (one IV?).
Dear Elvis Munyaradzi Ganyaupfu ,
Appreciate your reply.
1. To find out the treatment effectiveness among the patient suffered from BACKPAIN and who recovered quickly in between male and female.
2. IV:
A. Gender
B. Pain score (number of time taken before intervention )
C. Types of disease (2-4 or dichotomous)
D. History of pain (recent or old)
N.B. In each follow-up, the patient was eliminated if they do not have any symptoms. At the end of the 10 visits, the remaining patients were successfully recovered.
Thank you.
Regards,
Khan
Dear Mohammad Jobair Khan,
Firstly, let's appreciate that when we dichotomize a variable, it makes that variable a categorical one, regardless of it being an independent or a dependent variable.
Secondly, clarity in our objective(s) is paramount. It looks like you have two: firstly, the effectiveness of a treatment (dichotomy of effective or not, hence, categorical) and secondly, gender influence on frequency of visit. It looks as though that the second objective - frequency of visit - may be a numerical variable or a dichotomized one (becomes categorical); the threshold or point of dichotomy you will most probably decide based on subject matter knowledge; gender is a predictor/independent variable.
Overall, to start of exploring your data, you could use the basic chi square. The way I see it based on info above, survival analysis is the way to go. Here's where you can start: http://www.sthda.com/english/wiki/survival-analysis-basics
Dear Saiful Safuan Md Sani ,
Cordially appreciate your response. Will read the given resource.
Well, among the survival analysis which one supposed to be more effective?
Regards,
KHAN
Here's something that may help you decide: https://www.theanalysisfactor.com/the-six-types-of-survival-analysis-and-challenges-in-learning-them/
I think you will need to get Kaplan-Meier curves, log-rank test to compare those curves and then Cox proportional hazards regression to analyse which predictors are significant.
Dear Saiful Safuan Md Sani ,
Thank you for your reply and resources.
Regards,
KHAN
I suggest you take a look at discrete time survival models as these are more flexible than Cox models and they can be estimated by a logit model once the data has put into the person period form - they deal quite naturally with censoring and they are inherently categorical
Article It's About Time: Using Discrete-Time Survival Analysis to St...
and as part of a wider set of models:
https://www.bristol.ac.uk/media-library/sites/cmm/migrated/documents/discrete-time-eha-july2013-combined.pdf
Dear Prof. Kelvyn Jones ,
I extend my unconditional appreciation for your time to read and put the essential resources. References are extremely helpful to fulfill my expectations of the query.
Sincerely,
KHAN
- Badges
- Science topic
- Similar topics
- Organizational Psychology
- Training