The lay abstract featured today (for Decompositions of Information Criteria in Joint Model Under Probit Link for Response by Fang Liu and Lei Cao) is from Stat with the full article now available to read here.

How to Cite

Liu, F. and Cao, L. (2024), Decompositions of Information Criteria in Joint Model Under Probit Link for Response. Stat, 13: e70010. https://doi.org/10.1002/sta4.70010

In computerized tests, we look at how people answer questions and how long it takes them. We want to find a model that fits this data well. Some new ways to check if our model is good have been invented. These ways look at how well one part of the data fits when we know the other part. They use something called DIC and LPML, which are like scores for the model. But these methods are usually used with a “logistic” model, and it’s harder to use them with a “probit” model, which is another way to look at data. This paper shows how we can use these methods with the probit model to make our tests better. In simple words, we’re finding new tools to make sure our test questions work well together.