I have a survey of the participants, some leave (one item), some get their professions acknowledged, some infos about earnings to evaluate their success.(You also know infos before and after.)
In my opinion you can't establish causal effects, because you don't know a group of similar people who haven't participated. So, I think only descriptive analysis like logit probit works, or?
Hello Klaus,
Your question is not fully clear. but let me try to help:
first, you're right, in survey one cannot establish/confirm a causal relationship, mostly because the data on exposure and outcome are collected at the same time, so you do not know with certitude what came first (with possible reverse causality). In survey designs, you mostly look for "associations"; i.e. "how many times more those experiencing X, compared to those experiencing Y are likely to have outcome Z"? You measure this with prevalence ratios, but also with odds ratios if you need to do multivariable analysis
Hope this helps, otherwise do not hesitate to come back
Best
Rodolfo
You need to do research in 2 parts as follows:
In this way please have a look at my articles as below:
- "Explanation of the“Strategic Assessment Model” and verification of its impacts on the “Managerial Research Scientific Process” (Mahmoudi Maymand & Samaeizadeh, 2017),
- "Explanation of the " Strategic Marketing Management " Model and verification of its impacts on " Increasing the Organizational Profitability" (Mahmoudi Maymand & Samaeizadeh, 2017).
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology