T-TEST interpretation

[Grace Kells]

I’m testing the significance of a predictor through a t-test, after having made an univariate analysis and run a simple linear regression.
I’ve computed the t statistic as follows: since the number of observations is 100, and I’m now analyzing a model with one predictor only, my t test with 98 degrees of freedom at a significance level of alpha = 0.05 is 1.66.
The simple linear regression gave me these results:
Coefficient = 1.74
Standard error = 0.9
t = 1.94
P>|t| = 0.055
95% Confidence Interval = -0.39 + 3.5.

Looking ad the “t” column, I should conclude for the significance of the predictor. However, a p-value slightly bigger than 0.05 (as well as the confidence interval, limited by extremes of opposite signs) would make me look at this variable as a non significant one.
Could you help me interpret these results?

Thank you

[Carlo Lazzaro]

Grace:
-regress- reports a two-sided -ttest-.
Hence, as per you data:

. di (ttail(98, 1.94))*2
.05525431

Set aside some decimal digits, the output approaches P>|t| = 0.055.

[Grace Kells]

could you help me find the right command to compute on Stata the two sided T statistics that I should compare with 1.94?

[Carlo Lazzaro]

Grace:
see -ttest-.