Calculates item nearest neighbor imputation approach discussed by Yeh et al. (2023)

Calculates item nearest neighbor imputation approach discussed by Yeh et al. (2023)

Usage

inn(dat, reg, random, verbose)

Arguments

dat

A single subject's MCQ data in long form

reg

Registry list from .instrument_registry() / .mcq_registry()

random

Boolean whether to insert a random draw (0 or 1) for NAs whose neighbours disagree or are all missing. When the observed neighbours agree (e.g. responses 1, NA, NA within a rank group) the NAs take that value even with random = TRUE, following Yeh et al. (2023). Items that stay missing are reported with a warning.

verbose

Boolean whether to print subject and question ids pertaining to missing data

Value

An imputed data set to be scored