# Thread: Please help... reall in a bind! (granger causality test)

1. ## Please help... reall in a bind! (granger causality test)

I have run a Granger causality test on two time series but im not sure based on the output whether or not to reject the null hypothesis. Below is the stata output. Any help would be greatly appreciated as im absolutely stumped!

H0: sse does not Granger-cause dax

F( 4, 458) = 4.15
Prob > F = 0.0026

chi2(4) = 16.92 (asymptotic)
Prob > chi2 = 0.0020 (asymptotic)

Source | SS df MS Number of obs = 467
-------------+------------------------------ F( 8, 458) = 4.04
Model | .007337069 8 .000917134 Prob > F = 0.0001
Residual | .10401977 458 .000227117 R-squared = 0.0659
-------------+------------------------------ Adj R-squared = 0.0496
Total | .111356839 466 .000238963 Root MSE = .01507

------------------------------------------------------------------------------
dax | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
dax |
L1. | -.0569729 .0466041 -1.22 0.222 -.1485573 .0346116
L2. | .0717223 .0461151 1.56 0.121 -.018901 .1623457
L3. | -.1502169 .0456892 -3.29 0.001 -.2400034 -.0604304
L4. | .0303349 .0457368 0.66 0.508 -.0595452 .1202149
sse |
L1. | .1538627 .0456992 3.37 0.001 .0640566 .2436688

2. I haven't taught Time Series, nor do I know what this granger thing is.
I do teach Design though
and it looks like your p-value is... Prob > F = 0.0026
That means your test stat is way out there in the tail of an F
and you should reject H_o and accept H_a, whatever they are.
I certianly need more info than this.
But your p-values do indicate, accept the alternative hypothesis.

-----------------------------------------------------------------------------------

Prob > chi2 = 0.0020
is another p-value saying accept H_a