Null Hypothesis

ShaShaBear

I've been asked in a question to state the null hypothesis. I'm having trouble stating it. It wants me to state it for the "mean difference in refraction values between eyes, μD". The mean difference, when I calculate it, is 0.263. The previous question related to this topic was "Suppose that it is desired to test the hypothesis that the mean refraction value in left eyes, μL , equals the mean refraction value in right eyes, μR".

So I understand that my null hypothesis is something I am meant to be trying to disprove, but also that it is assumed true unless overwhelming amounts of evidence suggest otherwise. I don't know if I am saying that the null hypothesis is zero, is not zero, is 0.263, is not 0.263.... Help??

Old Perry

I dont think you are looking for a specific value. when you state your null it should be something like μL = μR or the opposite of what ever your alternative hypothesis is. I think thats what you are looking for anyways although i may be corrected.

ShaShaBear

Old Perry wrote: »

I dont think you are looking for a specific value. when you state your null it should be something like μL = μR or the opposite of what ever your alternative hypothesis is. I think thats what you are looking for anyways although i may be corrected.

So you mean you think I should be stating essentially the first question using hypothesis terms? Like Ho = Mean uL= Mean uR ?
That does make a lot of sense when you look at later questions

Old Perry

yes, if it just says state the alt or null hypothesis then just stat it in statistical terms with the symbols you used above. sometimes its deceptively simple.

ShaShaBear

I'm learning over the past few weeks of desperately attempting to grasp this subject for repeats, that the entire of my Stats course is deceptively simple :rolleyes:

All I'm really struggling with now is Chi-Squared and Anova. Which sound like Jackie Chan movies, but I'm sure I'll know more tomorrow!

Old Perry

Tough draw, I had a choice of the two in my last exam and went with anova so if ya have q's about that post em here and I'll do my best to help. Best of luck with it .

ShaShaBear

The fail rate in my class was so high (38 out of 52) that he is letting us answer any parts of all four questions (normally we can only answer three). Hopefully if I manage to at least pick something up today, it will all add up to 40% on the day - at this stage that's all I'm hoping for!

Thanks for the offer though, it's great to be able to get help here!

usually a null hypothesis is that something is equal (or no difference between, bla bla) and the alternative is that it isn't

so in this case it's just mean 1 = mean 2

MathsManiac

The original post states that the question asked for the null hypothesis to be stated in terms of the difference between the two values. Accordingly, rather than stating it as μL = μR, I would say that the preferred answer is to state μL - μR = 0.

It becomes more explicit then that you're trying to determine whether this number (μL - μR) is significantly different from 0.