Fixed or Random Effect Model in Cross -Sectional data

[Ashish Bandhu]

Dear All,

I have a query regarding the fixed or random effect model in cross-sectional data (One observation for each individual). I want to know if I can run fixed, random or both model in such data. Any resources will be very helpful.

Thanks

Ashish

[Carlo Lazzaro]

Ashish:
the answer is no, as you have cross-sectional (and not panel) data.
Go -regress-.

[Ashish Bandhu]

Ashish:
the answer is no, as you have cross-sectional (and not panel) data.
Go -regress-.

Thanks, what about multilevel cross- sectional data where individuals are grouped within different villages. Confusion arises from here https://stats.stackexchange.com/ques...sectional-data.

Thanks in advance

[Rich Goldstein]

it is certainly possible to have multi-level cross-sectional data (and thus to have FE or RE models): students nested within classrooms or patients nested within hospitals (and many other examples); this may not be "panel" data in the sense that economists use that term but it is certainly clustered data and thus amenable to multi-level models

[Carlo Lazzaro]

Ashish:
while Rich is obviously correct, your original post meant something different.

[Rich Goldstein]

Carlo Lazzaro I posted because I thought there was so little info in #1, that what I had to say might be relevant - and I still don't see that #1 means something different - maybe you can clarify for me, please?

[Carlo Lazzaro]

Rich:
I meant that the first post mentioned cross-sectional data without further details.
Therefore, as you say, the information were not detailed enough (for me at least) to figure out that the original poster (OP) meant " multi-level cross-sectional data", as you envisaged.
That's why I considered your guidance relevant but, at the same time, I am still under the impression that the OP should have been clear(er) from the very start.

[Rich Goldstein]

Carlo - thanks for the clarification - I agree - rich