PDA

View Full Version : Hypothesis Tests Project



IdRatherSleep
04-26-2006, 08:37 PM
Hi, I need a little bit of help with my project. The work is to do a test statistic for the difference in means between age of cars owned by men and those owned by women. The other question is the difference between proportions in men athletes and women athletes(basing the proportion on how many are athletes out of their respective samples).

For the test statistic of difference of means I got -.7026, my question is that even though this is a t distribution should I multiply this number by 2? For degrees of freedom I got 40(I used the equation, no problem there), I am supposed to use 5% level of significance so then does this mean that my p-value(-.7026) is far bigger than alpha, meaning I should choose null hypothesis?
area: .2 .1 .05 .025 .01
d.f 40: .851 1.305 1.684 2.021 2.423
My other question is what does level of significance means? I think it means the probability that I could be wrong in choosing the null hypothesis, is this correct?

jsbeckton
04-26-2006, 10:50 PM
You do not "choose" the null hypothesis, you "do not reject" the null hypothesis.
Level of significance is the smallest level at which you "do not reject" null. The higher the level, the less likely you are to reject, which means lower Type I error. (rejecting Ho when true) It should have became clear once you got to the P-value.

As far as doubling P-value, are you doing a one or two tailed test? Do you know how to find the p-value for a t test? you have to find the value in the expanded t chart where your t and the v (n-1) line up. That is your p-value, double if two tailed.

IdRatherSleep
04-26-2006, 11:27 PM
It is a two tail test, so which number is the one that I have to double the -.7026? The t area and d.f are in my first post, I just didn't line them up. So do I have to multiply times two the number I got from the proportion test statistic? Should I turn it into a positive number?