Skip to main content

Table 5 Item-fit indexes applying the one-dimensional approach to the various short versions

From: Establishing the HLS-Q12 short version of the European Health Literacy Survey Questionnaire: latent trait analyses applying Rasch modelling and confirmatory factor analysis

 

HLS-Q12b

HL-SF12 [26]

HLS-EU-Q16 [11]

Item

Loc.

Chi prob.

Infit

t

Com.

Loc

Chi prob.

DIF

ordered

Infit

t

Com.

Loc

Chi prob.

DIF

Ordered

Infit

t

Com.

1

                   

2

−0.097

0.997

1.01

0.2

0.269

− 0.041

0.699

  

0.99

−0.1

0.278

0.133

0.606

  

1.03

0.5

0.276

3

                   

4

            

−0.236

0.703

  

1.00

0.0

0.338

5

            

−0.453

0.313

Edu, h > l

 

0.97

−0.7

0.387

6

     

0.137

0.348

  

1.05

1.0

0.183

       

7

0.053

0.705

1.03

0.7

0.239

              

8

            

−0.865

0.087

 

No

0.94

−1.1

0.376

9

                   

10

0.892

0.761

0.97

−0.6

0.292

0.964

0.848

  

1.00

−0.1

0.266

       

11

            

1.084

0.001*

Age, y < o

Edu, h < l

 

1.17

3.4a

0.139

12

                   

13

            

−0.082

0.332

Gen m > f

 

0.97

−0.6

0.319

14

−0.625

0.275

1.00

0.0

0.260

              

15

     

−0.842

0.025

Gen, m > f

Age, y > o

No

1.06

1.2

0.207

       

16

            

−1.305

0.044

  

0.91

−1.6

0.443

17

                   

18

0.420

0.357

0.98

−0.5

0.363

0.490

0.006

  

0.93

−1.6

0.416

0.687

0.633

  

1.00

0.1

0.299

19

                   

20

                   

21

            

−0.793

0.002*

Edu, h > l

No

0.95

−0.9

0.347

22

                   

23

−1.158

0.065

0.98

−0.3

0.281

−0.997

0.431

  

0.96

−0.7

0.291

−0.908

0.385

  

0.96

−0.7

0.319

24

                   

25

                   

26

     

0.526

0.129

Age, y < o

 

1.07

1.5

0.187

       

27

                   

28

1.068

0.265

1.04

0.9

0.176

       

1.360

0.039

  

1.07

1.4

0.232

29

                   

30

0.422

0.284

1.03

0.6

0.230

0.502

0.845

Edu, h < l

 

1.04

0.9

0.217

       

31

            

1.167

0.166

  

0.98

−0.4

0.349

32

−1.135

0.618

0.98

−0.5

0.337

              

33

     

−0.298

0.652

  

0.97

−0.5

0.321

−0.110

0.820

  

1.03

0.6

0.240

34

                   

35

                   

36

                   

37

            

0.242

0.070

Age, y > o

 

1.06

1.2

0.220

38

0.702

0.484

1.06

1.3

0.196

              

39

     

0.372

0.831

  

1.00

0.0

0.293

0.531

0.305

  

0.96

−0.7

0.366

40

                   

41

                   

42

                   

43

−0.728

0.375

0.95

−0.9

0.426

−0.621

0.236

  

0.94

−1.1

0.385

−0.452

0.450

  

1.00

0.0

0.322

44

0.186

0.105

1.04

0.8

0.248

              

45

     

−0.194

0.072

  

1.07

1.4

0.202

       

46

                   

47

                   
  1. Note. This table shows the item-fit indexes when applying the one-dimensional approach to the various short versions. The HLS-EU-Q16 could not be deemed sufficiently unidimensional. The results are displayed for comparison purposes
  2. Chi square probability (Chi prob), differential item functioning (DIF), item location estimate (Loc) and unordered response categories (ordered) were obtained from Rasch modelling using RUMM2030 software. ConQuest 4 software was used for the Infit measures
  3. aA t–value > 1.96 and an infit-value > 1 indicate a poor fit with the Rasch model due to there being more variation in the data than expected given the model (item is under-discriminating)
  4. *Item with a significant misfit (p-value below Bonferroni-adjusted 5%)
  5. For items displaying uniform DIF, the relevant dichotomized person factor levels are indicated. For example, “Age, y > o” refers to the uniform DIF for the person factor age in favour of the y = younger respondents (47 years or younger) factor level as compared to o = older respondents (48 years or older) factor level. The “Highest completed education level” (Edu) factor has the l = low (primary and secondary school) and h = high (university or university college) levels, and the factor “Gender” (Gen) has the f = females and m = males levels
  6. Ordered: “No” refers to items with unordered response categories
  7. None of the items on the HLS-Q12 displayed DIF or unordered response categories
  8. Com: communalities obtained from CFA using the LISREL software
  9. bHLS-Q12 developed through the present study
Back to article page

BMC Health Services Research

ISSN: 1472-6963

Contact us