Table 1

Crude odds ratios before and after removal of records with missing data: scenario 1 analyses

Full data (n = 662 farms)

Following the removal of records (n = 348 farms)


Farm-level predictors

ORa

s.e. ORb

p-value

ORa

s.e. ORb

p-value

% change


L vs. S bird counte

0.284

1.292

<0.0001

0.185

1.367

<0.0001

34.9c

L vs. S house counte

0.354

1.306

<0.0001

0.299

1.327

<0.0001

15.5

Indoors

0.298

1.289

<0.0001

0.301

1.307

<0.0001

1.0

Free-range

5.266

1.298

<0.0001

5.010

1.317

<0.0001

4.9

Housing other

0.829

1.510

0.650

0.598

1.590

0.268

27.9c

Partial housing

1.009

1.941

0.990

0.763

2.209

0.733

24.4

Integrated

0.271

1.222

<0.0001

0.258

1.309

<0.0001

4.8

East

1.905

1.298

0.014

2.545

1.392

0.005

25.1

Scotland

0.333

2.144

0.150

0.308

2.899

0.268

7.5

Wales

0.119

2.087

0.004

0.192

2.848

0.115

38.0d

West

0.418

1.287

0.001

0.401

1.429

0.010

4.1


aOR = odds ratio; bs.e. = standard error of the odds ratio; c >25% change in odds ratio but direction of association and significance is comparable; dsingle variable for which there is >25% change in odds ratio and no change in direction of association, but significance is altered; eL = large, S = small.

Nickbakhsh et al. BMC Veterinary Research 2011 7:66   doi:10.1186/1746-6148-7-66

Open Data