Our algorithm knows best: Adhering to Blank Slate’s schedule keeps cognitive performance high

Clients sometimes ask how we know that our algorithm is doing its job, and we love this question. It’s good to question authority, especially when that authority is a mystical AI algorithm.

First, a quick recap on how our algorithm works. This custom-built machine learning algorithm reads the performance data from our users all day every day. It keeps track of each user’s cognitive performance and, among other things, determines the ideal time to remind them to review their Blank Slate content before it fades from their memory. We double-check its work using a variety of statistical tools, but we also like to look at clients’ data to confirm that things are going as expected.

In a recent peek into our algorithm’s accuracy, we looked to see whether people are more accurate on their Blank Slate questions when they adhere to the app’s reminders to use it. If the algorithm is doing well, people should be more accurate when they adhere to their reminders and less accurate when they don’t.

To determine this, we ran a correlational analysis between two variables:

1. Users’ knowledge levels
2. The average number of days that passed between when each user received a Blank Slate reminder and when they actually used the app.

These data come from 165 law enforcement officers who use Blank Slate to stay sharp on policing protocols. Check out their results!

Curious about those values in the upper right corner? Pearson’s r is the “correlation coefficient.” It ranges from -1 to 1, and it tells you the strength and direction of the relationship between two variables. In our case, we found a strong negative correlation of -.30. This means that as one variable increases (days passed), the other variable decreases (knowledge levels). The p value represents the probability that these results are due to chance. In our case, there is a 1 in 10,000 chance that our results are a fluke. These are strong results!

Law enforcement officers who answered their Blank Slate questions more promptly following a notification demonstrated high knowledge levels than those who waited. Conversely, officers who waited several days or weeks to answer their Blank Slate questions performed more poorly.

All that is to say, our algorithm is quite good at determining when the optimal moment is to help folks enhance their cognitive performance. When people wait too long to answer their Blank Slate questions, they miss out on the recall window that the algorithm determined for them individually. Then they forget critical information, which can affect job performance.

Prospective clients: If you currently conduct training sessions once or twice a year, we hope that a quick glance at our graph will convince you that this frequency is inadequate. People often forget information rapidly. To minimize forgetting and enhance cognitive performance, Blank Slate leverages cutting-edge AI to maintain peak knowledge year-round for just a few minutes of users’ time each day.

Current clients: If you don’t already, consider mandating Blank Slate usage on your team. If your team members are motivated to adhere to their Blank Slate reminders, their knowledge levels will stay high.

__

Amy Smith, PhD

Chief Science Officer, Blank Slate

__

Want to improve cognitive performance at your organization? Get in touch!