Edit: I think I've run afoul of an anti-triplebyte sentiment. I should clarify that I think this post did a good job building a very simple example of the stati

Edit: I think I've run afoul of an anti-triplebyte sentiment. I should clarify that I think this post did a good job building a very simple example of the statistical theory behind assessment,but I have no idea whether their product / approach is reasonable or not. Building a good assessment is much more than just statistics, and it sounds from other comments that there are serious concerns about the validity of their tool. Really enjoyed the build up from simple cases, to more complex models! If you're interested in the statistics behind estimating skill, and how well questions tell apart novices from experts, check out item response theory :) https://en.m.wikipedia.org/wiki/Item_response_theory https://en.m.wikipedia.org/wiki/Item_response_theory