Table 4.
Item | Item Calibration in logits (SE) | INFIT (MNSQ) | OUTFIT (MNSQ) | DIF Probability by Mantel-Haenszel statistics (p) | ||
---|---|---|---|---|---|---|
Gender | Agea | Languageb | ||||
#1 | −0.64 (0.20) | 1.14 | 1.25 | 0.4381 | 0.0074 | 0.7937 |
#2 | 0.48 (0.23) | 0.99 | 0.89 | 0.0005 | 0.9200 | 0.5870 |
#3 | 1.56 (0.30) | 1.08 | 0.93 | 0.5964 | 0.0037 | 0.6836 |
#4 | 0.63 (0.24) | 0.80 | 0.69 | 0.9713 | 0.2586 | 0.9458 |
#5 | 1.09 (0.26) | 0.87 | 0.78 | 0.7520 | 0.0476 | 0.9763 |
#6 | −0.51 (0.20) | 1.07 | 1.14 | 0.6056 | 0.8955 | 0.1197 |
#7 | −0.58 (0.20) | 0.95 | 0.87 | 0.0476 | 0.2484 | 0.0850 |
#8 | −0.85 (0.20) | 0.78 | 0.68 | 0.5995 | 0.9560 | 0.8635 |
#9 | 1.02 (0.26) | 0.90 | 0.72 | 0.5301 | 0.5218 | 0.1634 |
#10 | 0.32 (0.23) | 0.94 | 1.01 | 0.0631 | 0.9597 | 0.1467 |
#11 | 1.94 (0.33) | 1.09 | 0.95 | 0.5933 | 0.6896 | 0.3634 |
#12 | −1.61 (0.20) | 0.86 | 0.74 | 0.1651 | 0.0992 | 0.0293 |
#13 | −0.55 (0.20) | 0.89 | 0.97 | 0.4904 | 0.1512 | 0.6042 |
#14 | −1.69 (0.20) | 1.45 | 1.66 | 0.7045 | 0.6147 | 0.0397 |
#15 | −0.51 (0.20) | 0.92 | 0.82 | 0.6857 | 0.2089 | 0.0286 |
#16 | 2.44 (0.38) | 1.00 | 1.29 | 0.4703 | 0.0555 | 0.1191 |
#17 | −1.31 (0.20) | 1.13 | 1.12 | 0.4670 | 0.9162 | 0.9222 |
#18 | −1.23 (0.20) | 1.04 | 0.95 | 0.5116 | 0.2195 | 0.7969 |
Bold italics = MNSQ values outside range of 0.7 to 1.3. Italics = Bonferroni adjusted p < 0.0027 (denote concern of DIF)
INFIT information-weighted fit statistics, OUTFIT outlier-sensitive statistics, MNSQ mean square, DIF differential item functioning
aAge < 50 years vs. ≥50 years; bChinese vs. English