Understanding Random Frequencies

The TestNotice platform, specifically the Calendar Generation tools, are based on the National Association of Drug Court Professionals (NADCP) Best Practice Standards. As these standards are sometimes unfamiliar to an agency performing random drug screening we have created this article to explain how it works and the reasoning behind it.

In the context of court-mandated testing, typically an offender is sentenced to a frequency expressed in straightforward terms such as "Randomly Tested Twice a Week". A typical interpretation of this order would imply that, in a seven day period from Monday to Monday, two random days should be selected for the offender to submit to a random screening. 

As pointed out by the NADCP, this is a flawed approach. Mathematically speaking, it will not take an offender who desires to know long to realize that in this approach being randomly tested on Monday and then again on Tuesday means they will not be tested again until the following Monday. This presents an opportunity to the offender to violate the intent of the court order and works against the goals of the court order itself.

Thus, the NADCP recommends a different interpretation to such an order, which is that it be considered a percentage chance daily the offender will be tested. For example, "Randomly Tested Twice a week" means 2 of every 7 days (or a 28% chance on any given day) they will be tested. TestNotice uses this percentage chance / frequency approach when generating a calendar in our Calendar Generation tool. 

Continuing with our example of twice a week testing, the side effect of such an approach is that, in any given month, it is possible the offender might be tested more or less than the 28% chance it represents. This is an example of the so-called "Gambler's Fallacy", which is defined as:

"Reasoning that, in a situation that is pure random chance, the outcome can be affected by previous outcomes."

A more simplistic example of this concept is the idea that because you flipped a coin and it landed on heads 5 times in a row the odds of it landing on tails the 6th time are higher, which is untrue.

Thus, when generating a calendar using the Calendar Generation tool it is entirely possible within a given 30 day period of a single month a user with a 28% chance may only have been called 20% of the time. Conversely, it was just as probable they would be called 36% of the time. Overall however, if you were to examine that same user over the course of 6 months the percentage would much more closely match the expectation of 28%.

The result of such an approach is a truly random, unpredictable testing schedule that complies with the intent of the court's wishes.


Please sign in to leave a comment.
Powered by Zendesk