Imagine you want to test two subject lines on emails:
and you send them both to 1000 people. 1000 should go in the Sample column in both rows.
Then let's say A gets opened by 220 people and B gets opened by 194. Those numbers go in the Affected column. (See?)
You can also click Affected and change it to Opened, if that makes it clearer for you. (See?)
This tool calculates statistical significance for little experiments.
| Sample | ||
|---|---|---|
| Group A | ||
| Group B |
Just so you know...
89% or less could just be a fluke
90% is pretty good for a rough sense
95% is good enough for academic publishing
99% or higher is awesome
...but this is all assuming you set up the experiment correctly.
created by Aaron Swartz (me@aaronsw.com) with help from Ben Wikler