Different people (and disciplines) define "experiment" differently, but one popular definition is that allocation into groups (at least one of the IVs) has a probability mechanism (e.g., random allocation). Were any of your IVs like this?

It appears from your question that the IVs are perceptions of the trainees. The categories in the variables of supervisor support, peer support, and subordinate support are not investigator assigned or naturally occurring distinctions, rather they are participants' self-reports. It is not an experimental design. However, you can look into the 'assessment reactivity' aspect of the participants having to answer the questions about your IVs immediately after training and try to elicit the 'question benefit effect', which would be experimental.

I think this is more of a survey design rather than an experimental design.

Thank you daniel sir. I have no random allocation sir. Moreover, i want to do CFA and structural model to reach hypothesis that made of cause and effect. It seems to me that it is explanatory. but if it is explanatory how can it be experimental sir?

Thank you chandu sir. Can you give me the procedure of experiment you are saying?

you are comparing post training transfer data with immediately after training data. Some books call this as pre-experimental design. For experimental design you have to divide the students into two group by randomization. One group will receive interventions and another group without intervention.

Dear Gem Sir, can i use CFA and structural model to see post training comparative study. if i use CFA and structural model, can it be called descriptive study?

Hi Md. Faisal-E-Alam , being experimental and wanting to make causal attributions are different things. An experiment helps with this: ``random assignment is the great \emph{ceteris paribus}--that is, other things being equal-of causal inference'' (Cook & Campbell, 1979, p.~5), but it is still difficult and requires assumptions. That said, there has been much work over previous decades (e.g., Rubin, Pearl) on causality in non-experimental research.

Thank you sir. That means, causal research and experimental research are different thing?

If by "causal research" you mean research with the aim of evaluating or measuring a causal effect, then this is a goal or purpose or aim of the research. It is not about how the study is conducted (though the purpose of the study should influence this). Experimental designs are one way to conduct a study, and a common choice when the aim is to test a causal hypothesis.

Yes sir, i have a hypothesis.But those hypothesises are not proved by experimental design. But they mentioned in hypothesis that IV has impact on DV. What can i do Wright sir?

Correct, they are not. From what I can tell your research is not designed to measure or test if the IV impacts the DV. If that is what you want to understand you should consider other designs and further research.

So, in that case, i have to take hypothesis that IV has association with DV. It will be Correlational research sir? For doing Correlational research, can hypothesis be taken as association between IV and DV sir?

I have seen lot of studies taken hypothesis IV has an impact on DV without experiment and random assignment, thus they are not properly make conclusion about the findings sir. Am i right sir?

Making causal attributions from any design requires assumptions. Just with non-experimental designs there as a few more. See https://www.tylervigen.com/spurious-correlations for examples where the author argues the causal attribution is suspect. But, if you see the books by Pearl, Rubin, and others you will see there has been a lot of work on reaching causal conclusions from non-experimental designs. What doesn't work well is throwing all observed variables into a regression and then treating the B value associated with any single DV as a measure of how much (or if) DV impacts IV. An old but good paper on this is: https://conservancy.umn.edu/handle/11299/184638

Dear Wright sir. After studying what you have given me, now i am thinking that i should make question in such a way that i can make association between IV and DV using hypothesis. if so is the case, can it be called Descriptive or Explantory research?

Can Hypothesis be used in such kind of association between IVs and DV?

Moreover, by CFA and strictural model, is it possible to draw conclusion about the relationshp?

Dear Mr. Faisal,

As per the information provided by you, I feel your research design may be considered as ex-post facto research design under quasi-experimental research design.

I have seen lots of study that make cause and effect prediction by Path analysis or structural model using SEM. Without experiment they are doing this. Is it right then how it is possible?

People can say anything is causal, they just need to make assumptions. I can say buying ice cream makes the weather hotter. The history of statistics is interesting for how it has dealt with causality. There is nothing magical that makes any of these statistical procedures able to turn correlation/covariance matrices into evidence for causes, but if you have a causal hypothesis that predicts data (or the correlation matrix) to have a certain pattern, you can see if the data can falsify this hypothesis.

I understood sir. I see many researches have causal hypothesis as mine. The interesting part is they are doing survey method to show causal relationship. After studying several books, i tries to understand they made assumptions but according to research strategy they should go for Exxperimental design. Using survey data is it possible to make causal assumption sir? As i studied survey is used in descriptive research not explanatory?

In social sciences, survey is a method for data collection. Survey does not determine the design of research. Experimental research design tries to overcome the experimental errors through randomisation and local control. Randomisation is a very common practice in social sciences. But problem comes in local control which is relatively tough to maintain. That's why quasi-experimental design is tried and survey method can be used for data collection. If you have problem in determining causal variables, then try multivariate analysis like PCA, EFA, Path Analysis etc. if basic assumptions are fulfilled.

Thank you Lahri sir. But, can you say, how i can run ex post facto design? As i studies there are more than one group to compare in ex post facto design. I have 200 respondents. Survey method is either longitudinal or cross sectional? i found cross sectional that is Restrospective in expost facto design.