related or independent observations - help needed


New Member
I am working in a project using data gathered for business purposes and I have a problem related to the observations in my samples.

A company distributed questionnaires to its employees, and then aggregated the results at unit level. So my sample consists of unit level data. I have data for 5 years. I cannot get any information regarding the individual answers. I would like to test whether there are differences between the departments, with respect to some issues discussed in the questionnaires. Moreover, I would like to see how the opinion with respect to those issues evolved during the years.
So far so good. It would seem I have repeated observations.

But there is a problem: the organizational structure changed every year (so departments were combined, new ones created etc), therefore the data for the same department (for 2 consecutive years) might not include the answers of the same employees (especially because filling in the questionnaire was not mandatory). Moreover, there is a different number of units every year, so my samples don’t have the same number of observations from year to year. So the observations are not precisely repeated observations.

How should I consider the observations? Repeated or independent? Are there any statistical methods which are not dependent of the type of observations?

Thank you in advance for your help.



TS Contributor
The observations are definitely not repeated observations.

All you can really do here is plot the means or medians to look for general trends

Without knowing individual scores, coupled with all of the changes you mentioned, it is not really possible to do a "statistical" analysis.