The following is default we use for calculating statistics. They can be modified in the Admin section as described in the link under (6) below.

**Type 1 error thresholds**:

When testing for the type 1 error we set the alpha (or significance) to 0.05.

**Type 2 error thresholds**:

When calculating the power of the experiment we use a type 2 error rate of 20% and enforce that by setting a sample size threshold that is calculated using the below settings as defaults.

- the alpha (or significance) = 0.05
- the statistical power or 1-beta = 0.8 where beta is 0.2

The effect size required is dependent on the sample size of your experiment.

Much more detail can be found in the Documentation:

(1) Metrics impact overview (here)

(2) Applying filters (here)

(3) Understanding impact (here)

(4) Attribution & exclusion (here)

(5) Statistical significance (here)

(6) Statistical settings (here)

## Comments

0 comments

Please sign in to leave a comment.