T-stat on Regression analysis help

[nabran]

I have a question which is based on a linear-regression analysis simple model on Excel. I'm revising for my exams and have had absolutely no response from my university class tutors who are very busy researching for their own projects. I really need help on this. The question;

''Transcribe the values for alpha and beta in A80 and B80 respectively. Construct t-tests for the null hypothesis that alpha = 20000 and beta = 2000 putting your values in cells A81 and B81 respectively. Can you reject each hypothesis at the 5% significance level; indicating your answer by R (reject) or NR (not reject) in cells A82 and B82 respectively. ''

This part of the question follows on after calculating a linear trend (and therefore having the alpha and beta values for linear regression) and deseasoning the data.
I can smooth exponentially, I can trend data, I can de-trend, de-seasonalise and forecast.

However, what confuses me about that question. Is if I took the typical approach of taking a t-test question, it wouldn't be on a pre=computed value...
I,e, how can i calculate the mean and st.deviation of a single Alpha value? Furthermore, I guess I could use =TINV to get the critical t-value i guess...

Could anyone please clear up how to do such a calculation? What's more, it seems very simple by the way the question is set out, i.e. Alpha and beta in cells A80 and B80, told to put t-test values in A81 and B81, and Rejecting or Not Rejecting in A82 and B82. I'm missing something!

All help appreciated, willing to discuss too! Regards

[nabran]

I've been looking over it continuously for the last few hours.

I've noticed that in the initial regression calculation (i calculated alpha via =intercept and beta via =slope) however I noticed the question stated that instead use the Regression Command, I thought it odd at the time. I calculated it via regression command and it gave me the same intercept and slope values. However, I noticed that it also gives the standard error of both the intercept and slope. And i noticed whilst looking a lot online that both T-calculated = bx/sd.error of x

(i.e. x=0, so b0/st.e of the intercept)
I calculated these and of course they gave the SAME values as the t-stats given on the bottom line of the Regression Command (via data analysis)
So I have already got two computed t-stats... It does say a 95% interval, so using =TINV(0.95,68) (is 68 correct? I have 70 data pieces, n-2? online i've seen different approaces) and it's a very small number....so obviously i'd reject them both? This can't be right? Thanks if you cna help! :)