I'm doing a quantitative study entitled "Technology Transfer and Competitive Advantages in Oil and Gas Companies".

In my study, the unit of analysis is the organization and there are only 9 oil and gas companies in my country. The respondents will be randomly selected from the managerial level (stratified) based on the size of each company. The total number of employees in managerial level is 3,336 and the sample size is 959.

I'm going to use Structure Equation Modeling for the analysis.

My question is, taking into consideration the unit of analysis is the "organization", does the sample size "959" represent the population (9 companies), and is it enough? Is this the right methodology?

My friends told me that if the unit of analysis is organization and I have only 9 companies, so one person from each organization will represent the organization. That means I will have only 9 respondents, which is not enough for the study.


1 Answer 1


It sounds like your unit of analysis is the individual, not the organization. That means your sample size will be 959. Since you're using SEM, I assume that you're also planning on treating your data as hierarchical -- that is, people nested in companies. Then you can include individual level terms as well as company level terms in your model. Since you only have 9 companies, however, you can probably only include about 3-4 company level terms in your model.

Hope that helps.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.