confusion about Ho/Ha hypothesis for simple/multiple regression

#1
The one thing I do not understand (and my book doesn't explain) is how do you determine if it's Ha > 0, Ha < 0 or Ha =/= 0? Anyone can shed some light on this problem?


And if you determine that Ha =/= 0, it becomes a two-tailed test, so you have to divide the alpha by 2 when finding the rejection zone, or also multiply the P-value by two, right?


I'll give an example from some questions that I have to answer for a homework:

1) Given the following statistics of a multiple regression model, can we conclude at the 1% significance level that X1 and y are linearly related?

2) Test at the 5% significance level to determine whether annual household income and annual family clothes expenditure are positively linearly related.

3) Test at the 1% significance level to determine whether the number of family members and annual family clothes expenditure are linearly related.

Thank you very much!
 
#2
Could anyone confirm if this is true?

If the question says positive linear, it's >

If the question says negative linear, it's <

If nothing is mentioned, it's =/=
 

terzi

TS Contributor
#3
Hi Parad0x,

When you run a regression model, you get parameter estimates, that help you understand the relationship between the regressors and the expected value of a dependent variable. Using those coefficients and the hypothesis testing related to them, you can answer the problem.