[help]interpretation needed please

#1
How would you interpret the regression result is good or not?>
How would you interpret coefficient and ?
Would you reject the hypothesis that the domestic output (x3) has the effect of -2 on y and what has the effect will become 3 on y??
And why you using t test but not using normal distribution test?



plase help me with it i make this solution from excel from data can you please help me solve it ??
 

noetsi

Fortran must die
#2
Good or bad is subjective:p When you look at a model a good place to start is the F test and the significance. If it is below .05 than by common usage it is statistically significant. Then you decide if substantively the adjusted R squared is high enough to make the model seem useful. This depends on what you are analyzing, does explaining 51 percent of the total variance in your dependent variable make your analysis useful or not?

The b's are the X variables (2 and 3 I assume). They tell you the change in Y for a one unit change in that X. Again you have to decide if that is meaningful. The p value tells you if it is statistically signficant - much like the F test for the model.

I do not understand question 3. The p values tell you if a variable has statistically significant effect, this effect is the slope which your ouput calls coefficients.

It is not a good idea to use regression p values, which are associated with t, if the distribution is not normal although to some extent the results are robust to nonnormality especially if the sample size is large.

We don't solve problems here, we make suggestions (particularly if it is homework related).
 
#3
Would you reject the hypothesis that the domestic output (x2) has the effect of 3 on y ?And why you using t test but not using normal distribution test?

this is the question three proper sorry my mistake
 

noetsi

Fortran must die
#6
The answer to 3 depends on the p value. At the 95% confidence level a p value from .05 down causes you to reject the null - which would mean in this case it does have an impact on Y (which is 3 in this case). So you have to look at the p value and determine if it is below .05 or not.
 

Dason

Ambassador to the humans
#7
I do not understand question 3. The p values tell you if a variable has statistically significant effect, this effect is the slope which your ouput calls coefficients.
This only gives a p-value for Ho: beta = 0. The question wants one for Ho: beta = -2

please solve this on page and paste a copy over here please... please thank you so much...
I've already warned you about this and noetsi even reminded you:
noetsi said:
We don't solve problems here, we make suggestions (particularly if it is homework related).
Honestly I'm going to post the homework help policy we have again. If you don't take the time to read every word and every link that is contained in here and take the advice to heart I will give you a temporary ban. You need to stop being demanding and asking for full solutions. You do realize that this is YOUR homework and you really should be the one doing all of it. We're here to help - but we're sure as hell not just going to write out a full solution for you so you copy it right onto your homework.

Read all of this and take it to heart or get banned - it's your choice:
-----
Hi! :welcome: We are glad that you posted here! This looks like a homework question though. Our homework help policy can be found here. We mainly just want to see what you have tried so far and that you have put some effort into the problem. I would also suggest checking out this thread for some guidelines on smart posting behavior that can help you get answers that are better much more quickly.
 

noetsi

Fortran must die
#10
This only gives a p-value for Ho: beta = 0. The question wants one for Ho: beta = -2
An interesting point that I was thinking about when answering the question. I know that formally the test is of the null beta=0. But I thought if you rejected the null you could then assume that the specific effect size (the slope) was true.

Is this not the case, that is is the only thing you can assume if you reject the null is that beta is not 0. That you can not assume it is the specific slope estimated (say -2)?
 

Dason

Ambassador to the humans
#12
An interesting point that I was thinking about when answering the question. I know that formally the test is of the null beta=0. But I thought if you rejected the null you could then assume that the specific effect size (the slope) was true.
"assume that the specific effect size was true" - what does that mean exactly?
Is this not the case, that is is the only thing you can assume if you reject the null is that beta is not 0. That you can not assume it is the specific slope estimated (say -2)?
Yeah the hypothesis test only tells you if you have evidence that the true slope is different from 0. You would need a confidence interval or another hypothesis test to make a claim about how the true slope compares to -2.
 
#13
no no i am sorry for my english
this is the right question
x2 t statistcai is greater thn 3 has the effect on y but t tabulated is < 3 should we reject or accept? hypothesis?
 
#15
The Thanks button is a very visible symbol of "thumbs up" in the lower left corner of every post. Besides the hand shape, there is a written Thanks. You can click it and officially thank every post that was useful or every post that had taken some time of the commenter. :)
 
#16
Would you reject the hypothesis that the domestic output (x2) has the effect of 3 on y ?And why you using t test but not using normal distribution test?

this is the question three proper sorry my mistake

answer this please ... :S adn tell me why we use t test and not using normal distribution it is because that when large sample size we use t test ?? right?
 
#18
Would you reject the hypothesis that the domestic output (x2) has the effect of 3 on y ?And why you using t test but not using normal distribution test?
Ah t and normal distributions are rather similar. The t distribution is used under the assumption of normality of error terms. Using t distribution allows more conservative (broader) confidence intervals. I think using the t distribution is advantageous when the sample is rather small (but the assumptions still hold). When the sample is large enough, using both distributions would provide rather similar results.

adn tell me why we use t test and not using normal distribution it is because that when large sample size we use t test ?? right?
No in a large-enough sample, both distributions would be OK (of course if the assumptions are met, in the first place). In smaller samples, the t distribution is better.

and should we accept H0 becauase calculated t value of x2 is greater thn tabulated t value right? please...
There is a critical value for the t statistics (the t value) which leads to a P value smaller than 0.05. This value is t = 1.96 for a 95% confidence limit. When the t statistic of the X2 variable is larger than the critical t value (in the case of X2, 3.2356 > 1.96), then the P value would become less than 0.05. When the P values gets less than 0.05, we reject the null hypothesis.
 

Dason

Ambassador to the humans
#19
Ah t and normal distributions are rather similar. The t distribution is used under the assumption of normality of error terms. Using t distribution allows more conservative (broader) confidence intervals. I think using the t distribution is advantageous when the sample is rather small (but the assumptions still hold).
You use a t-distribution when you have to estimate the standard deviation from the data itself. So in practice that means pretty much always.
 
#20
About your first question, is it good or not: Well you can judge the quality of the regression based on values such as the adjusted r-squared. However, as noetsi pointed out, the question could be asked in a better way: If the regression has a proper accuracy or not (rather than is it good or bad). Anyways, the adjusted r-squared is a key statistics in judging the regression. Based on its value, it seems that your regression is rather accurate (or good). Note that this can be refuted by more sophisticated discussion about the merit of the adjusted r-squared, in the first place. But be sure that this value is an important value that has its implications in judging a regression.