Observed reliability and IRT reliability from the N(0, 1) distribution tended to be slightly lower compared with the IRT values from distributions two, three, and four.This may be because the true score variance in the two and three distributions ishigher than the variance (1.0) where item responses were simulated.