00:01
All right, so here we have an advertiser that we're trying to figure out if their groceries are lower than the competitors.
00:10
Okay, so in order to do this, we have some data for four different weeks, okay, and we have a number for the advertiser and the competitor for each week.
00:27
Okay, so what we're going to do is we're actually going to subtract each of this.
00:31
These, okay? and then we're going to test really the differences between the two.
00:38
Okay, so here's all my data, and the last one, 234 .2, and 261 .24.
00:50
And i'm going to subtract each of these in each of these cases, so i really am going to have four data points, which is the differences between each of these.
00:57
So this comes out to negative 1 .64, negative 14 .99, negative 23.
01:04
And negative 27 .04.
01:07
This is the data set that i'm going to do all my analysis on.
01:11
And what i'm really trying to do for the null alternative hypotheses, right, is the null hypothesis i'm trying to see is the mean difference really equal to zero because then they would really be the same, right? versus the alternative would be that the mean difference is less than zero because then that would mean that we really do make less than the competitors.
01:36
Now, it does look like it does just say, is there a significant difference? okay.
01:44
Which is weird because we probably would actually want to check and see if one is greater than the other.
01:54
Okay.
01:55
But this does say that we want to see if there's just a difference.
01:57
So we actually are going to use the not equal to sign for my alternative hypothesis.
02:02
Okay.
02:05
For the next part, we're going to get our test statistic, which is just x bar of the differences minus the mean from the null hypothesis divided by the standard deviation of the differences over the square root of n of the differences, so that's our four, right? so let me put this into my calculator.
02:30
Okay.
02:31
So when i do this, i get a test statistic of negative 2 .3.
02:36
97.
02:39
This is going to be a two -tailed test, meaning that the area here, right, for my alpha is split between the two tails...