Skip to main content

Table 1 Analysis by the Rasch model - overall fit statistics for J-EAT-10

From: Item analysis of the Eating Assessment Tool (EAT-10) by the Rasch model: a secondary analysis of cross-sectional survey data obtained among community-dwelling elders

Scale analysis % Extr. Fit residuals mean (SD) Item-trait interaction Reliability Unidi-mensionality
Item Person χ2 (df) P PSI / α Power of fit T-test %
Initial analyses
 1. Full sample (N = 1144) 42% −0.66 (4.16) − 0.30 (1.08) 485.48 (40) < 0.001 0.65 / 0.85 Good 2.27%
 2. Independent group (N = 594) 53% − 0.82 (2.13) −0.37 (0.86) 175.92 (30) < 0.001 0.32 / 0.72 Too low 0.72%
 3. Dependent group (N = 550) 30% −0.19 (3.53) −0.25 (1.17) 293.28 (40) < 0.001 0.74 / 0.87 Good 1.83%
Analysis after reducing to three scoring categories (full sample,N = 1144)
 4. Scoring structure: 00122a 65% −0.89 (3.21) −0.25 (0.77) 284.01 (40) < 0.001 0.38 / 0.76 Low 1.73%
 5. Scoring structure: 01112 42% −1.73 (4.72) −0.35 (0.90) 416.01 (40) < 0.001 0.69 / 0.83 Good 1.21%
 6. Scoring structure: 01122 42% −1.90 (4.34) −0.38 (1.01) 448.90 (40) < 0.001 0.66 / 0.83 Good 2.27%
 7. Scoring structure: 01222 43% −0.78 (4.26) −0.26 (1.11) 439.37 (40) < 0.001 0.65 / 0.89 Reasonable 0.76%
 8. Scoring structure: 00112 65% −0.56 (4.04) −0.24 (0.74) 225.24 (30) < 0.001 0.49 / 0.76 Low 1.98%
Optimal values <  15% 0 (<  1.00) 0 (<  1.00)   > 0.05 > 0.70 < 5%
  1. J-EAT-10 Japanese version of the Eating Assessment Tool, % Extr. percentage of extreme scores, χ2 Chi-square, SD standard deviation, df degrees of freedom, PSI Person Separation Index, α Cronbach’s alpha, CI confidence interval
  2. aScoring structure suggested by Cordier et al. [16]