00:01
So we want to test if the exposure to their campaign changed people's water usage.
00:06
So null hypothesis, if we call the difference the score before the campaign minus the score after, then the null hypothesis is going to be that the mean of these differences is zero and the alternative hypothesis is going to be that it's non -zero, i .e.
00:18
There's some difference caused by it.
00:20
We're doing it at the 5 % level of significance.
00:24
So, oh sorry, that was part a, stating the null alternative hypotheses.
00:29
The design requirements and assumptions, we need the data to be independent.
00:41
The paired sets of data must be from the same subject for each pair of data.
00:54
And we assume that the population of the differences of the d's are normally distributed.
01:04
That's capital n for normal distribution.
01:06
Part c says state the decision rule.
01:11
So we'll reject if our test statistic is greater than the positive critical value or less than the negative critical value.
01:25
And the positive critical value is going to be that bounding the top 0 .025 % because we've got a two -tailed test.
01:31
So the top 2 .5%, sorry.
01:37
And at degrees of freedom equal to n minus one, which is seven.
01:42
And that turns out to be 2 .365...