I am trying multiple panel models based on the identical dataset, which is large with 265,904 observations.
Let's say I am comparing Table 4's models (1), (2), and (3).
Table_4_1 <- plm(DV ~ X1 + X3 + X4 + X5 + X6, index="firm_id", model="within", data=Survivors_upto_2019)
Table_4_2 <- plm(DV ~ X2 + X3 + X4 + X5 + X6, index="firm_id", model="within", data=Survivors_upto_2019)
Table_4_3 <- plm(DV ~ X2 + X3 + X4 + X5 + X6 + X7 + X8 + X9, index="firm_id", model="within", data=Survivors_upto_2019)
summary(Table_4_1)
summary(Table_4_2)
summary(Table_4_3)
Table_4_1 and Table_4_2 are unnested, but Table_4_2 and Table_4_3 are nested. Therefore, I used the following commands,
jtest(Table_4_1,Table_4_2)
pFtest(Table_4_2, Table_4_3)
For J-test, I obtained the following results.
Estimate Std. Error t value Pr(>|t|)
M1 + fitted(M2) 0.95029 0.0034222 277.683 < 2.2e-16 ***
M2 + fitted(M1) 0.35485 0.0075335 47.103 < 2.2e-16 ***
For pFtest, I got an error.
Warning message:
In pf(stat, df1, df2, lower.tail = FALSE) : NaNs produced
My questions are:
- How do I interpret the results from J-statistic?
- For a panel nested comparison, how should I fix pFtest?