Table 3.
Psychometric properties of the Yale Food Addiction Scale for children in item levels
| Item # | Analyses from classical test theory | Analyses from item response theory | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Factor loadinga | Item total correlation | Test–retest reliabilityb | Infit MnSq | Outfit MnSq | Difficulty | Discrimination | DIF contrast across genderc,d | DIF contrast across weight statusc,e | |
| #1 | 0.41 | 0.49 | 0.73 | 1.03 | 0.88 | − 0.11 | 1.17 | − 0.17 | 0.31 |
| #2 | 0.53 | 0.50 | 0.78 | 0.88 | 0.73 | − 0.16 | 1.14 | 0.21 | 0.49 |
| #3 | 0.51 | 0.46 | 0.74 | 0.95 | 0.96 | − 0.75 | 0.96 | 0.01 | 0.33 |
| #4 | 0.44 | 0.42 | 0.83 | 1.12 | 1.04 | − 0.30 | 0.89 | 0.24 | 0.12 |
| #5 | 0.43 | 0.45 | 0.84 | 0.94 | 0.85 | − 1.52 | 0.87 | 0.43 | 0.24 |
| #6 | 0.53 | 0.52 | 0.88 | 0.96 | 0.59 | 0.57 | 1.00 | − 0.23 | 0.33 |
| #7 | 0.63 | 0.46 | 0.92 | 1.08 | 1.29 | − 0.63 | 0.81 | 0.27 | − 0.30 |
| #8 | 0.69 | 0.60 | 0.76 | 0.89 | 0.74 | − 0.30 | 1.08 | 0.19 | − 0.22 |
| #9 | 0.70 | 0.64 | 0.80 | 0.75 | 0.58 | 0.93 | 1.13 | 0.44 | 0.21 |
| #10 | 0.51 | 0.54 | 0.86 | 0.96 | 0.68 | − 0.35 | 1.03 | 0.16 | 0.45 |
| #11 | 0.55 | 0.54 | 0.94 | 0.96 | 0.82 | − 0.12 | 1.01 | 0.43 | 0.49 |
| #12 | 0.69 | 0.58 | 0.88 | 0.85 | 0.52 | 0.82 | 1.09 | − 0.47 | 0.23 |
| #13 | 0.70 | 0.57 | 0.72 | 0.92 | 0.70 | 0.15 | 1.05 | − 0.27 | 0.41 |
| #14 | 0.68 | 0.59 | 0.70 | 0.91 | 0.64 | − 0.08 | 1.17 | − 0.39 | − 0.22 |
| #15 | 0.66 | 0.60 | 0.76 | 0.89 | 0.58 | − 0.15 | 1.08 | 0.34 | − 0.43 |
| #16 | 0.64 | 0.57 | 0.72 | 0.88 | 0.65 | 0.84 | 1.03 | 0.09 | − 0.14 |
| #17 | 0.49 | 0.41 | 0.82 | 1.10 | 1.55 | 1.35 | 1.17 | 0.41 | − 0.42 |
| #18 | 0.46 | 0.40 | 0.75 | 1.09 | 1.02 | 1.39 | 1.19 | 0.01 | − 0.34 |
| #21 | 0.40 | 0.40 | 0.73 | 1.26 | 1.49 | − 2.90 | 1.15 | 0.43 | − 0.39 |
| #22 | 0.41 | 0.42 | 0.79 | 1.16 | 1.47 | − 3.02 | 1.34 | 0.21 | − 0.46 |
| #23 | 0.44 | 0.48 | 0.82 | 1.20 | 1.17 | 0.19 | 1.31 | − 0.24 | − 0.17 |
| #25 | 0.57 | 0.41 | 0.77 | 1.24 | 1.33 | − 1.01 | 0.35 | − 0.41 | 0.38 |
MnSq mean square error, DIF differential item functioning
aBased on confirmatory factor analysis
bUsing intraclass correlation coefficient
cDIF contrast > 0.5 indicates substantial DIF
dDIF contrast across gender = difficulty for females − difficulty for males
eDIF contrast across weight status = difficulty for participants with obesity − difficulty for participants with overweight