Skip to main content

Table 4 Differences Between Estimations from Sparse and Dense Data, with Hausman Test Statistics (q)

From: Improving knowledge about disability transitions by adding retrospective information to panel surveys. Population Health Metrics

 

Men

Women

 

Sparsea

Dense

q

Sparse

Dense

Q

ADL

      

Eating

3001.4

3140.2

0c

2910.1

3109.3

0c

I/O Bedb

3344.4

3842.6

553.4*

3519.1

4182.7

282.3*

Mobility

3441.1

4032.7

3098.1*

3771.7

4723.4

592.7*

Dressing

3244.2

3600.6

480.4*

3366.5

3885.4

0c

Bathing

3492.4

4088.9

0c

3947.2

4948.9

248.0*

Toileting

3298.6

3613.7

98.7*

3501.0

4109.2

362.1*

  1. aListed are -log likelihood values for estimates from sparse and dense data for each ADL.
  2. bI/O Bed = Getting in or out of bed.
  3. cNegative q values are treated as zeros [23,24].
  4. *p < 0.0001