Nurturing maths awesomeness
The performance increase given is the average percentage reduction in answering times for the 4 basic arithmetic operations (+ - × ÷).
The methodolody was to draw a straight line of best fit through a graph of each individual's times for number facts within a constraint (ie adding to 10 was a separate graph to adding to 20; the 10 times table was a separate graph to the 7 times table). Users who rejected cookies were not included. We then averaged the amount of time spent on 0Maths per day (just under 10 minutes) and interpolated / extrapolated to 1 week and looked at the change compared to the y intercept.
Our methodology assumes that the relationship between time spent on the platform and improvement in answering times is approximately linear which of course it's not, but assuming linearity over a week is not too large a stretch. To draw a curve through relatively small amounts of noisy data would be more error prone.
It is an extraordinary figure and probably warrants a little discussion. In the time period concerned (to December 2024), we were mainly targetting users with maths anxiety. The fact that we don't punish mistakes allows bright maths anxious children to relax and perform their best, once they get used to the fact that mistakes are OK.
There was a clearer trend with time data than with error rates which were low enough that the noise overwhelmed the signal. This makes sense as pupils tend to be set online tasks to practice number facts and build automaticity rather than to try the operation for the first time. Mistakes are relatively rare; slow answers aren't.
© Copyright Influenca Ltd 2020 - 2025