Skip to main content

Table 4 Items with significant uniform and non-uniform differential item functioning

From: Psychometric validation techniques applied to the IND-VFQ-33 visual function questionnaire: the Hyderabad ocular morbidity in the elderly study (HOMES)

 

Subgroup

Discrimination

B2 threshold

B3 threshold

B4 threshold

B5 threshold

P-value (BH) a

P-value (DIF) b

Factor 1: Daily Activities

Uniform Differential Item Functioning

  23: Enjoy social functions less

No depression

2.4

1.5 c

2.0

2.3

0.008

< 0.001

Depression

2.4

1.3

1.7

2.0

  

  25: Become a burden on others

No depression

1.8

2.0

2.8

2.9

0.004

< 0.001

Depression

1.8

1.5

2.0

2.2

  

  26: Frightened to lose remaining vision

No depression

1.6

1.2

2.1

2.4

< 0.001

< 0.001

Depression

1.6

0.8

1.5

1.9

  

  27: Do you have reduced vision

No depression

2.0

0.1

1.7

2.2

0.006

< 0.001

Depression

2.0

0.0

1.5

2.2

  

  33: Do you have blurred vision

No depression

2.0

0.4

1.7

2.4

0.006

< 0.001

Depression

2.0

0.4

1.5

2.2

  

Non-Uniform Differential Item Functioning

  24: Ashamed that you can’t see

No depression

2.6

1.9

2.4

2.7

0.01

< 0.001

Depression

1.6

1.5

2.3

2.9

  

Factor 2: Bright Lights

Uniform Differential Item Functioning

  31: Vehicle light makes you close eyes

Male

1.8

−0.7

0.8

1.3

0.003

=0.001

Female

1.8

−0.5

1.0

1.4

  
  1. Key: (a) Benjamini-Hochberg adjusted p-value significance threshold; (b) unadjusted raw p-value (results of the likelihood ratio test (uniform or non-uniform DIF vs base model)); (c) beta-threshold representing the individuals’ visual difficulty theta level at which it becomes more likely for them to choose that option on the 4- or 5-point Likert scale, i.e. at visual difficulty theta level 1.5, it becomes more likely for an individual without depression to choose option 2 rather than option 1 on the 4-point scale