help please..hypothesis testing

Harpy

probably over complicating this question for myself but i cant do it at all i don't know what formulas to use:
the question is 25 voters and they say 19 of them will vote for, 6 against

and have to construct a hypotheiss test with the initial hypothesis that the for side will win with 95% confidence level..



ok after writing this out i just had a thought when it says initial hypothesis does that mean i have my null hypothesis as h0=19???

Thomas_S_Hunterson

I think your null hypothesis is that the number of votes for is equal to 17.5 votes or equal to 0.5 as a proportion. Your alternative hypothesis is that the number of votes is greater than this figure.

Then get your t values.

Time Magazine

I think...

H0: yes voters <= 0.5(population)
H1: (by default) yes voters > 0.5(population)

You run your t-tests and reject the null, i.e. conclude that more than half vote yes aka the motion is passed.

Harpy

thanks so i tried it and this is what i ended up with...
ok so i got the standard error of the proportion by
using the formula
the sqr rt of P(1-P)/n
and put in .5 for P and got .1

and i used p-P/SE(P) to get my test statistic so i said .76-.5/.1 =2.6
.76 being the sample proportion.
i then looked it up as a one tailed test(am i right in saying it's one tailed?)
and the value i got from the t tables was 1.711

so the test statistic of 2.6 fell in the reject area....
is that right do yee think

MathsManiac

If your null hypothesis is that the population is in favour of the proposition, and if your sample has a majority in favour, you can't possibly reject the null hypotheseis. It is a one-tailed test and you're at the opposite side from the rejection region.

That is, the observed value in the sample is obviously consistent with the null hypothesis.

Are you sure you've presented your problem correctly?

And are you sure you were asked to frame an "initial hypothesis" or was it a "null hypothesis"?

Harpy

thanks for the help so far...
this is the qusetion exactly how its written
A sample of 25 voters have been polled and 19 report they will vote FOR, 6 AGAINST on an up-coming referendum.
(a) Conduct a formal hypothesis test with the initial hypothesis that the FOR side will win, with a 95% confidence level

im going to have another go at it.. in the last one i've put .5 as my null hypothesis..
what does it mean when it says the initial hypothesis??does that mean i use 19 or the proportion of 19/25 as my null hypothesis.??.
aw maths and me don't get along!!

Time Magazine

Harpy wrote: »

im going to have another go at it.. in the last one i've put .5 as my null hypothesis..

0.5 what? Greater than, less than? I specified what I consider the appropriate hypotheses in my last post.

For hypothesis testing, you're hoping to reject the null.* So frame the hypothesis such that if you reject the null, you get the result you want. In this case, the result you want is that more than half of the population will vote Yes.

*This is not quite true. You are, of course, not hoping for anything other than discovering the truth. However rejecting the null is the most powerful endorsement you can provide. Thus you'd "rather" reject the null that more than half will vote No than failing to reject that more than half will vote Yes.

MathsManiac

Just to clarify: I wasn't saying that the hypothesis suggested by Ibid was inappropriate to the situation. It just seems to me that it may not fulfil the requirement of the (poorly phrased?) question you were asked. "Initial hypothesis" is not a term that's normally used in these contexts.

Harpy

Ibid wrote: »

0.5 what? Greater than, less than? I specified what I consider the appropriate hypotheses in my last post.

For hypothesis testing, you're hoping to reject the null.* So frame the hypothesis such that if you reject the null, you get the result you want. In this case, the result you want is that more than half of the population will vote Yes.

*This is not quite true. You are, of course, not hoping for anything other than discovering the truth. However rejecting the null is the most powerful endorsement you can provide. Thus you'd "rather" reject the null that more than half will vote No than failing to reject that more than half will vote Yes.

yea sorry i just didnt put in the=> in the last post just was saying the figure i used sorry if i confused..

i did it again today but i decided to use 0.52 as my ho because my thinking was that they'd need 13 votes or more for, the FOR side to win...
i ended up getting fail to reject the null hypothesis so i was happy with that... aw and still 3 more questions to do this is gonna be fun:(

Thanks for the help!!