Statistics - Question

I'm doing AQA Statistics (B I think).

I can't get my head around whether you accept/reject H0/H1 in different situations.

Can someone explain how you decide your conclusion after doing a Hypothesis test (Binomial/Poisson/Normal/PMCC/Spearman's Rank/Sign/Wilcoxon-signed Rank/Man-Whittney U etc).

Thanks.

Reply 1

Drunk1000

Surely someone must know...

Reply 2

Squishy

Um, I did Edexcel Stats last year, but I'm not quite sure what you want to know. You reject the null hypothesis when there is a significant relationship found, and you accept it when there isn't.

Reply 3

Drunk1000

I mean when you've got you Critcial Value and your Test Statistic, you reject or accept Ho if one is greater than the other or something...

Reply 4

niksip

ok, am doing stats 2 and 3 on wed and this is what i think....

PMCC: If test statistic is greater than critical value, then you REJECT H0.

Spearman's Rank: Again, if test statistic is greater than critical value, then you REJECT H0.

Sign Test: If the relevant p value is greaer than 0.05 then you ACCEPT H0. If less than or equal to 0.05, reject H0.

Wilcoxon: If test statistic is less than the critical value, then you REJECT H0.

Mann Whitney: If test stat is less than the critical value, then you REJECT H0.

Hope this helps, best to check in the textbooks though just in case i have made a mistake,

Namrata x

Reply 5

Squishy

Oh right, for some tests, your test statistic has to be lower than the critical value to be significant, and for some, it has to be higher. For the Mann-Whitney, Wilcoxon and sign test you accept your null hypothesis if your test stat is higher than the critical value. For the t-test, Spearman's and Pearson's, you accept your null hypothesis if your test stat is lower than the critical value. I don't really remember much more than that, sorry.

Reply 6

Drunk1000

Thanks both of you, excellent.

One more thing...regarding the SIGN TEST.

This is the only one I don't fully understand. I get the first bit where you count the +'s/-'s and find the test statistic. I understand about the Significance level (e.g 5% = 0.05 and you look on the Binomial table with p=0.5). But how do decide whether you're looking for the first value HIGHER or LOWER?

thanks.

"This is the only one I don't fully understand. I get the first bit where you count the +'s/-'s and find the test statistic. I understand about the Significance level (e.g 5% = 0.05 and you look on the Binomial table with p=0.5). But how do decide whether you're looking for the first value HIGHER or LOWER?"

Anybody, before tomorrow?

Reply 9

niksip

ok, i think i understand your question,

say you've got a one tailed test at 5%and 20 +/- in total, 5 of them are + and 15 -. You find your test stat from the binomial tables and you get 0.0207. since this is smaller than 0.05, you reject H0. when looking at the binomial tables, you find the one which is smaller, in this case, 5. if you had 17+ and 3-, for example, you would look up 3. i would check this method in the textbook and answer some of the questions, jsut to make sure that you are getting the correct answer.

hope this helps, namrata x

Reply 10

Drunk1000

Thanks alot, you're an absolute legend!

Reply 11

niksip

hi, how did unit 3 go for you?

Reply 12

Drunk1000

Really well.

Unit 2 wasn't so good.

You?

Reply 13

niksip

pretty much the same, unit 3 was really good, but unit 2 was weird, didn't really like the time series question. the problem with unit 3 being nice is that the grade boundaries are going to be high!