00:01
Okay, so we've got a question about a hypothesis test to see whether adverts affect the sales on a website.
00:09
So putting up adverts for other companies puts off customers basically.
00:16
So we've got an experiment that's been taken up.
00:21
So we've recorded the mean sales when ads were so i'll call that x bar a is 26 .75.
00:32
The standard deviation when the ads were shown is 18 .21.
00:41
And the number of times an ad was shown was 8172.
00:46
The mean when the ad was not shown, which i'll call x bar n is 27 .21 sales.
00:54
So the deviation when the ad was not shown was 19 .2.
00:59
And the number of times the ad was not shown was 10 ,437.
01:05
So it's asking whether the basically to conduct a hypothesis to see whether the ad affects the sales or not.
01:11
And then they say that the ad, the advertiser claims that actually running the ads improves the sales.
01:21
So we've got two different null hypotheses.
01:22
The first, well, the null hypothesis is going to be the same.
01:25
So the first time the null hypothesis is that the mean sales with ads is the same as the mean sales without ads.
01:33
And the alternative hypothesis, it just says, does the showing the ad affect your sales? so the alternative hypothesis is just going to be that they're not equal.
01:44
You correctly computed a t statistic of minus 1 .669.
01:53
And your question is about the p value.
01:57
So in this case, when we're just answering the question, does the ad affect your sales? so our alternative hypothesis is that there's just a difference in the population means.
02:06
We're doing a two -tailed t -test.
02:11
So if our value of minus 1 .669 lies here, then the p value is going to be this area.
02:23
And if this is plus 1 .669, this area here.
02:30
So it is indeed going to be 2 times 1 minus norm .s .disc.
02:36
This is in excel, obviously.
02:37
Minus 1 .669.
02:41
True.
02:44
Sorry, i've just seen what your sort of the main thing that we're struggling with here...