# Schema | Field Name | Description | Data Type | Format/ Example | Null value allowed? | | --- | --- | --- | --- | --- | | experiment_id | Anonymised ID for the A/B test | string | “036afc” | No | | variant_id | The ID of the treatment group. The summary statistics for the corresponding control group is included in Columns 5-7, and hence there are no dedicated rows for the control groups in this dataset. <br /><br /> Note: The variants are not necessarily numbered consecutively. | integer | 2 | No | | metric_id | The ID of the organisational metric. (See notes below on list of metric) | integer | (1 \| 2 \| 3 \| 4) | No | | time_since_start | Number of days since the start of the experiment. | double | 12.5 | No | | count_c | Number of samples in the control group. (The number is stored as a double, but they are clearly integers) | double | 123456.0 | No | | mean_c | The sample mean of responses across the control group. | double | 4.361 | Yes | | variance_c | The sample variance of responses across the control group. | double | 72.354 | Yes | | count_t | Number of samples in the treatment group. | double | 123572.0 | Yes | | mean_t | The sample mean of responses across the treatment group. | double | 4.345 | Yes | | variance_t | The sample variance of responses across the treatment group. | double | 73.591 | Yes | ## List of metrics The dataset features four metrics, numbered 1, 2, 3, and 4. Metric 1 accepts binary responses, metrics 2 and 3 accept count-based responses, and metric 4 accepts non-negative real number responses. The responses for metrics 2, 3, and 4 demonstrate various degrees of right skewness.