Questions about F test & t test?

[Andy Lee]

Hello guys! My topic is about researching on impact of capital structure to profitbility. I use regression with FEM & REM. After running regression and using Hausman accreditation, my result is below and I can use FEM. However, my teacher told me that I needed to run F test & t test and after that I can discuss. I have read about t test and got it. But about F test, I still do not understand about command or explaination of this test. However, whether I must run F test & t test or not? Please help me! Thanks.

[Carlo Lazzaro]

Andy:
probably your teacher refers to the F-test appearing as a footnote of -xtreg,fe. outcome table (the F-test appearing above the table is usually returned by any linear regression model): if it reaches statistical significance, -xtreg, fe- outperforms pooled OLS (POLS).
However, I fail to get the role of -ttest- in deciding between -xtreg,fe-, -xtreg,re- and POLS.

[Marcos Almeida]

Hello Andy,

Welcome to the Stata Forum / Statalist.

To start, I gather you should kindly ask your teacher what he/she means by F test and t test in this precise context.

Additionally, it seems your query is related to an assignment. There are recommendations on the issue of assignment in the FAQ. Please take a look at that.

That said, generally speaking, with regards to the F test: besides the test you get under analysis of variance as well as the initial part of the OLS regression, in the output you shared there are two sorts of F-test, the "omnibus" F test and the one to test whether a panel data is better than OLS.

Hope that helps.

P.S. Crossed with Carlo's advice.

[Andy Lee]

Thank you @Carlo and @Marcos!!