-0.15<p1-p2<-0.03

-0.07<p1-p2<0.03

-0.03<p1-p2<0.07

0.05<p1-p2<0.12

- Thread starter aenyiema
- Start date

-0.15<p1-p2<-0.03

-0.07<p1-p2<0.03

-0.03<p1-p2<0.07

0.05<p1-p2<0.12

This goes to the heart of what a confidence interval is.

You could think of a confidence interval as the error bounds on the estimated mean. That is, we are 95% confident that the real value is somewhere between the bounds. For example, below is the confidence interval around the estimate of the mean, x (black arrow ^). The red arrow (^) below is the true population mean, which 95% of the time will fall within the interval.

|------------x------------|

.................^.....^.........

When looking at a difference of means, if the confidence interval of the difference goes through zero, then (to 95% confidence given the current sample) you cannot tell whether one is different to the other. If, however, the entire interval is on one or other side of zero, then there is a detectable difference. This should give you enough to work out the rest.