I've been running a split test for a while now and I'd like to work out whether or not my results are statistically significant. Firstly, I've recorded unique IPs in my database, with number of goals and total page loads. At the moment, my conversion rate is calculated by the total number of unique IPs in each group, divided by the total number of goals those IPs in that group have achieved. My first question is "is that the best way to determine conversion rate for this test"? Secondly, how do I work out whether the following data is significantly significant, or at least what the standard errors are: Group Sample Size Total Goals Conversion Rate 1 3565 100 2.81% 2 3479 80 2.3% 3 3502 74 2.11% Code (markup): Cheers