One Tailed Hypothesis Test of Proportions

magnifico99

New member
Joined
Apr 29, 2019
Messages
1
Hello!

I want to know whether the difference in the two proportions of the two samples (S1 and S2) is greater than 10% at a 5% significance level. Data and my calculations are in the attached picture.

There are two Hypotheses I want to check. Please refer to “Hypothesis 1” and “Hypothesis 2” in the attached picture (for some reasons I cannot upload the MS excel file here, would've been so much more better!)

Hypothesis 1:
Ho: S1-S2 <= 0.1
Ha: S1-S2 > 0.1

As you can see in the calculations, the p-value is greater than 0.05, therefore we do not reject the null, and conclude that the difference in the two proportions is less than 10%.

Hypothesis 2:
Ho: S1-S2 >= 0.1 (
Notice that this hypothesis is the opposite of Hypothesis 1.)
Ha: S1-S2 < 0.1

Even here, the p-value is greater than 0.05, therefore we do not reject the null, and conclude that the difference in the two proportions is greater than 10%.


So, my doubt really is this: How can the two Null Hypothesis both be true when they are the opposite of each other? I mean, the difference in proportions cannot be less than 10% AND greater than 10% at the same time. Right?

Of course, I am missing something in my calculations/ knowledge here. Please help me out!
 

Attachments

  • One Tailed Hypothesis Test of Proportions.PNG
    One Tailed Hypothesis Test of Proportions.PNG
    51 KB · Views: 5
Last edited:
I think your null hypothesis in both cases should be S1-S2=0. It is the statement of status quo.
 
Even as stated (which some authors do), the two null hypotheses can both be true, right? If the difference is both <= 0 and >= 0, that means it is zero.

They are not opposites. They overlap.
 
Top