Figure 2.

Recovery of standard errors (SEs) of item discrimination and difficulty parameters.

Note. MAD = mean absolute difference; ML = maximum likelihood; MLF = maximum likelihood with standard error approximation using the first-order derivative; MLR = maximum likelihood with robust standard errors; WLS = weighted least squares; WLSMV = weighted least squares with mean- and variance-adjusted chi-square test; ULSMV = unweighted least squares with mean- and variance-adjusted chi-square test. On the x-axis, 1 = ML-Logit, 2 = MLF-Logit, 3 = MLR-Logit, 4 = ML-Probit, 5 = MLF-Probit, 6 = MLR-Probit, 7 = WLS-Probit-Theta, 8 = WLSMV-Probit-Theta, and 9 = ULSMV-Probit-Theta.