TDA's rmod on binary item response data.
Arguments
- x
matrix of 0/1 responses, persons in rows.
- ...
passed to tda_run.
Value
centered item parameters in $items (joint ML is
identified only up to a constant, and its item estimates carry
the known upward bias of about k/(k-1) for k items), the
pattern parameters in $patterns, and the maximised log
likelihood in $logLik.
Examples
set.seed(1)
ab <- rnorm(80)
x <- sapply(c(-1, 1), function(d) rbinom(80, 1, plogis(ab - d)))
tda_rmod(x)$items # easier item first, harder second
#> item1 item2
#> 2.590267 -2.590267