I want to estimate individual’s meta-d’ (type-2 d’; Fleming, 2017) and preferably also d’ on decision tasks with multiple confidence levels. I would prefer using brms to achieve this. Do anyone know of a way of using brms for this purpose?
When searching articles and the internet in general I have only found Flemings own code to do it with JAGS: GitHub - metacoglab/HMeta-d: Hierarchical meta-d' model
If I understand Type II signal detection theory correctly, it can essentially be framed as a probit ordinal regression with accuracy as the predictor. In the ordinal regression framework, the intercept represents the confidence criterion, while the regression coefficient for accuracy corresponds to the Type II d′ value. The connection between signal detection theory and Gaussian probit regression on choice/accuracy data is discussed in this paper.: SDT and brms.
Thank you so much. The tutorial looks really helpful.