T-test compared to alpha = 0.05, please help!?

[DylanLe]

Variable Coefficient Std. Error t-Statistic Prob.

C -1.594359 0.466764 -3.415773 0.0014
LOG(K) 0.723402 0.033362 21.68329 0.0000
LOG(L) 0.625569 0.133864 4.673153 0.0000

R-squared 0.995429 Mean dependent var 3.957870
Adjusted R-squared 0.995206 S.D. dependent var 0.394490
S.E. of regression 0.027315 Akaike info criterion -4.297042
Sum squared resid 0.030590 Schwarz criterion -4.175393
Log likelihood 97.53492 Hannan-Quinn criter. -4.251928
F-statistic 4464.062 Durbin-Watson stat 0.631066
Prob(F-statistic) 0.000000
--------------------------------------…
Hello, above is my work from Eviews 7.
I'm required to answer if variable K has impact on out Y. Given equation: ln(Y)=B0+B1ln(K)+ B2ln(L)+U.
Based on eviews and pluging B1 and B2 into equation I have LOG(Y) = -1.59435907224 + 0.723401787597*LOG(K) + 0.625568755672*LOG(L)
It means that K and L do affect on Y ( K stands for Capital, L is Labour, Y is GDP).

By doing t-test, as on the eviews shows t-Statistic of Log(K) =21.68329.

Can we compare this number 21.68329 which is greater than 0.05, so we reject the null or we have to work out t-critical value to compare to t-Statistic ofLog(K)? If yes can someone please tell me how to work out t-critical value on evews7 please

I make Ho: mean K = mean Y ( no relationship between variable)
H1: mearn K # mean Y ( relationship exists)

Am I correct if I write null and alt-null hypothesis as above? Because as I know we use Mean(x)=Mean(Y) (mearning, there's no difference between means or B0=0) to test or compare the mean of 2 set of variable, so its understandable to state hypothesis as above. For example: Compare the height of 20 male to 20 female in a class.

However, I'm required to test if X affects on Y, can I use the same way for the hypothesis?

Regards