Revised 2-2-2016
By default, the system attempts to calculate three different norm values for each test score:
Overall - across all valid scores for this test.
Country - across all valid scores for this test where the test-taker was within same country as the current test-taker (based on IP Address).
Account - across all people within a given account who have taken a test.
Here's how each of these three percentiles is calculated:
1. Count the number of valid test events for this test for the current test version id. If there are 25 or more valid events, use this group to calculate the percentile. Otherwise:
2. Add progressively older test version ids to the pool until the total exceeds 25 valid events. If we find 25 or more, use this group to calculate the percentile.Otherwise:
3. Count the number of valid test events for this test for all test version ids. If there are 25 or more valid events, use this
group to calculate the percentile. Otherwise:
4. No percentile is reported.
All norms are calculated as a percentile - that is the percent of test events that have a score at or below the current test taker.
Since norms are calculated based on the number of valid test events AT THE TIME OF THE CALCULATION, values can change. For this reason the function "Recalculate Norms" is available in the detail test result section of the test.
