Research Article
Reliability of Family Proxy Data for Studies of Malignant Mesothelioma: Results from the ATSDR Pilot Surveillance
Table 3
McNemar/Durkalski statistics for paired responses, by type of relationship and type of question.
| Question set | Proxy type | Paired responses |
McNemar/Durkalski Test | Agreed | Disagreed |
Missing/unknown* | # | % | # | % | # | % | Chi-square | value |
| | Spouse | 229 | 89.8 | 24 | 9.4 | 2 | 0.8 | 1.48 | 0.22 | Respondent exposure | Child | 76 | 84.4 | 14 | 15.6 | 0 | 0.0 | 1.87 | 0.17 | (15 questions) | Sibling | 64 | 85.3 | 9 | 12.0 | 2 | 2.7 | 0 | 1 | | Other/parent | 62 | 82.7 | 13 | 17.3 | 0 | 0.0 | 0.14 | 0.71 | | Total | 431 | 87.1 | 60 | 12.1 | 4 | 0.8 | 0.15 | 0.70 |
| | Spouse | 384 | 94.1 | 24 | 5.9 | 0 | 0.0 | 1.14 | 0.29 | Respondent medical | Child | 119 | 82.6 | 7 | 4.9 | 18 | 12.5 | 2 | 0.16 | (24 questions) | Sibling | 98 | 81.7 | 10 | 8.3 | 12 | 10.0 | 1.92 | 0.17 | | Other/parent | 103 | 85.8 | 17 | 14.2 | 0 | 0.0 | 3 | 0.08 | | Total | 704 | 88.9 | 58 | 7.3 | 30 | 3.8 | 2.53 | 0.11 |
| | Spouse | 213 | 89.5 | 25 | 10.5 | 0 | 0.0 | 1.37 | 0.24 | Family exposure | Child | 71 | 84.5 | 13 | 15.5 | 0 | 0.0 | 0.02 | 0.88 | (14 questions) | Sibling | 58 | 82.9 | 12 | 17.1 | 0 | 0.0 | 3 | 0.08 | | Other/parent | 67 | 95.7 | 3 | 4.3 | 0 | 0.0 | 1 | 0.32 | | Total | 409 | 88.5 | 53 | 11.5 | 0 | 0.00 | 1.30 | 0.25 |
| | Spouse | 199 | 90.0 | 22 | 10.0 | 0 | 0.0 | 0.02 | 0.89 | Family medical | Child | 64 | 82.1 | 8 | 10.3 | 6 | 7.7 | 1.29 | 0.24 | (13 questions) | Sibling | 47 | 72.3 | 6 | 9.2 | 12 | 18.5 | 0.67 | 0.41 | | Other/parent | 56 | 86.2 | 9 | 13.8 | 0 | 0.0 | 3 | 0.08 | | Total | 366 | 85.3 | 45 | 10.5 | 18 | 4.2 | 0.95 | 0.33 |
| | Spouse | 102 | 85.7 | 16 | 13.4 | 1 | 0.8 | 5.91 | 0.02 | Smoking | Child | 38 | 90.5 | 4 | 9.5 | 0 | 0.0 | 1.8 | 0.18 | (7 questions) | Sibling | 30 | 85.7 | 5 | 14.3 | 0 | 0.0 | 0.67 | 0.41 | | Other/parent | 30 | 85.7 | 5 | 14.3 | 0 | 0.0 | 3.57 | 0.06 | | Total | 200 | 86.6 | 30 | 13.0 | 1 | 0.4 | 5.57 | 0.02 |
|
|
Missing or unknown answers were excluded from the McNemar/Durkalski statistic calculation.
|