Working with Panel Data

[azizan danish]

Hello Everyone,

I'm a student conducting panel data analysis.
I'm a little confused as to whether I'm doing it correctly.

So, I've started with the panel data format, then model selection. I found the selected model to be the Fixed Effect Model.
Then, the resulting xtreg model...fe vce(cluster). Is this correct for interpreting the results, or is there missing processed data?

Thank you.

[Carlo Lazzaro]

Azizan:
welcome to this forum.
As per FAQ, please post what you typed and what Stata gave you back. Thanks.

[azizan danish]

like this?

[Carlo Lazzaro]

Azizan:
with N>T, your approach looks right to me.
That said, I would test what happens with default standard errors, as N=27 us not that high to go cluster (Cameron_Miller_JHR_2015_February.pdf).

[azizan danish]

Although the sample size is relatively small, the use of regional panel data makes heteroskedasticity and serial correlation likely. To address these issues, I applied robust standard errors clustered at the regional level to ensure reliable statistical inference without altering the estimated coefficients. Would you recommend any alternative approaches that might be more appropriate in this context

---

---

Danish
(Stata 17.0)

[Andrew Musau]

You may want to look at wild cluster bootstrap for the small number of clusters that Carlo points out in #4. See https://www.stata.com/features/overv...rap-inference/. For Stata 17, you can install boottest from SSC to implement this.

ssc describe boottest

[azizan danish]

Thank you for your help. I have tried it. Is this true?

[Andrew Musau]

Looks good!