HI..... I used a stratified random sampling technique and two strata were formed. should I analyse the data separately for each stratum, the entire dataset, or a hybrid approach (both)? please give me references if possible.
Thank you
Maybe my question is inaccurate. I am using smart pls to analyse data collected based on the Likert scale. My question Is it mandatory to compare the data for each stratum when analysing and reporting the result or can for the entire data? I know it depends on the research question and objective but I read that "Stratified random sampling is useful when it is needful to compare subgroups".
It depends of your research objectives, if the objective is to identify/ generalise for whole population I don't think there's any need for separate reporting, unless you have some unique findings across strata. On the other hand if the objective is to do a multi group analysis, the strata must be reported separately and present comparative analysis.
Lack of detail makes advising difficult here. One knows from your description that your dependent variable is Likert scaled and that you have two strata. What else one needs to know is what is/are your independent variable(s) or is your design correlational? Anyway, I would advise representation of your strata whatever they are in the analysis.
- Badges
- Science topic
- Similar topics
- Filing