Re: Hypothesis Testing

On May 25, 4:04 am, Richard Ulrich <Rich.Ulr...@xxxxxxxxxxx> wrote:
On Fri, 23 May 2008 22:01:34 -0700 (PDT), sagar <ariji...@xxxxxxxxx>
wrote:

Thanks everybody in this group for clearing my concepts. I have
another very minor problem.Suppose I've four  variablesY, X1, X2 and
X3 from 1992 to 2004. Now, I want to test the hypothesis of each
variable. Suppose I want to test that population X1=0, X2<0 & X3>0.

If you want to test whether the means are far from zero,
for X1, X2, and X3,  regression is not the way to do it.

Now, say I run the regression on Y,  independent variable, and
X1,X2,X3 as dependent variables using SPSS.

You are saying that you want to use Y to predict X1, X2 and X3.
I don't think that is what you want to do.

My problem is can I do above mentioned hypothesis test based on the
beta values I obtained as SPSS output.
I hope I presented the problem clearly.

Try again?  The regression coefficients have t-tests, which
show something about the partial effect of each one, when
controlling for the others.  

Go back and state one or more hypotheses in a better style.
Maybe you should be more concrete, that is, name the variables.

--
Rich Ulrich

http://www.pitt.edu/~wpilib/index.html

Let me reframe my Hypothesis testing question a bit. I've a model
where I want to test what are the factors that are affecting the
profitability of tea companies. Let my regression equation is
profit= beta0+beta1. price(tea)+beta2.cost+beta3price(coffee)+u.
Now, I've the hypothesis that if price of tea is increasing than it
will increase the profitability. Similarly if both cost of company &
price of coffee is decreasing than profit will increase. I run this
regression on SPSS & get the beta values & t-statistics. I know how to
calculate for hypothesis testing.
My question is can I simultaneously test 3 different hypothesis using
a single regression equation.
.

Loading