I am using joint null test if all averaging effect of these variables (dkt_4 dkt_3 dkt_2 dkt_1) is 0
Code:
reghdfe y dkt_4 dkt_3 dkt_2 dkt_1
test dkt_4=0
test dkt_3=0, accum
test dkt_2=0, accum
test dkt_1=0, accum
The result I got is as below
Code:
. test dkt_4=0

 ( 1)  dkt_4 = 0

       F(  1, 25334) =    2.95
            Prob > F =    0.0860

. test dkt_3=0, accum

 ( 1)  dkt_4 = 0
 ( 2)  dkt_3 = 0

       F(  2, 25334) =    6.09
            Prob > F =    0.0023

. test dkt_2=0, accum

 ( 1)  dkt_4 = 0
 ( 2)  dkt_3 = 0
 ( 3)  dkt_2 = 0

       F(  3, 25334) =    5.82
            Prob > F =    0.0006

. test dkt_1=0, accum

 ( 1)  dkt_4 = 0
 ( 2)  dkt_3 = 0
 ( 3)  dkt_2 = 0
 ( 4)  dkt_1 = 0

       F(  4, 25334) =    4.48
            Prob > F =    0.0013
From this result, I am wondering that why Prob>F is higher when accounting for the last hypothesis (bolded values (0.0013 and 0.0006)
And normally, I usually look at p-value or t-stat to see if the hypothesis is significant, I am wondering how to see if the result is significant based on the results here?