Skip to main content

Table 4 Likelihood ratio test between two nested cumulative ordinal logistic regression models for DIF examination

From: Validation of the food access survey tool to assess household food insecurity in rural Bangladesh

   Literacy (literate vs illiterate) Age (≥30 y vs <30 y)
Model 1a Model 2b −2D p-value Effect sizec Model 2b −2D p-value Effect sizec
Item 1 –5515.64 –5496.01 39.27 <0.001 0.18 % –5508.65 13.98 <0.001 0.05 %
Item 2 –3043.47 –3015.94 55.07 <0.001 0.60 % –3040.42 6.10 0.05 0.00 %
Item 3 –3359.23 –3348.06 22.34 <0.001 0.16 % –3348.15 22.16 <0.001 0.20 %
Item 4 –6012.05 –5991.26 41.56 <0.001 0.14 % –6009.88 4.32 0.12 0.00 %
Item 5 –3244.66 –3220.84 47.64 <0.001 0.33 % –3240.99 7.34 0.03 0.03 %
Item 6 –4281.38 –4256.09 50.59 <0.001 0.29 % –4268.56 25.64 <0.001 0.14 %
Item 7 –10278.28 –10184.37 187.82 <0.001 0.57 % –10245.95 64.66 <0.001 0.18 %
Item 8 –6046.82 –5989.04 115.54 <0.001 0.68 % –6035.00 23.64 <0.001 0.14 %
Item 9 –6296.15 –6262.50 67.29 <0.001 0.32 % –6286.59 19.11 <0.001 0.10 %
   Land ownership (Yes vs No) Household size (≥4 vs <4)
Model 1a Model 2b −2D p-value Effect sizec Model 2b −2D p-value Effect sizec
Item 1 –5515.64 –5216.83 597.62 <0.001 0.23 % –5502.33 26.62 <0.001 0.09 %
Item 2 –3043.47 –2904.33 278.28 <0.001 0.30 % –3031.35 24.24 <0.001 0.22 %
Item 3 –3359.23 –3218.90 280.65 <0.001 –0.22 % –3349.25 19.95 <0.001 0.14 %
Item 4 –6012.05 –5714.07 595.96 <0.001 0.01 % –5993.87 36.35 <0.001 0.11 %
Item 5 –3244.66 –3137.21 214.91 <0.001 –0.46 % –3235.91 17.50 <0.001 0.08 %
Item 6 –4281.38 –4103.63 355.50 <0.001 –0.21 % –4269.68 23.41 <0.001 0.10 %
Item 7 –10278.28 –9533.13 1490.29 <0.001 0.89 % –10206.68 143.19 <0.001 0.40 %
Item 8 –6046.82 –5735.46 622.72 <0.001 0.09 % –6033.31 27.00 <0.001 0.10 %
Item 9 –6296.15 –5888.44 815.41 <0.001 0.17 % –6282.91 26.48 <0.001 0.04 %
  1. aModel 1: The log odds of response in one category or below is modeled as a linear function of total food insecurity score only
  2. bModel 2: The log odds of response in one category or below is modeled as a linear function of total food insecurity score, group membership and interaction between total score and group membership
  3. cEffect size is defined as the difference of R-squared between model 2 and model 1.c. Effect size is defined as the difference of R-squared between model 2 and model 1