We quantify the information loss incurred by categorizing an unobserved continuous variable () into an ordered categorical scale (). The continuous variable is conceptualized as true score (Ï„) (which varies across individuals) plus random error ((epsilon)), with both components assumed to be...