Table 3.
Interrater reliability measures of SOC 2010 coding of I&O data by different coding methods.
| Coding technique | n (missing) | Kappa (95% CI) |
| Investigator 1 x Investigator 2: 2 digit | 338 (21) | 0.84a (0.79-0.88) |
| Investigator 1 x Investigator 2: 4 digit | 337 (22) | 0.58 (0.52-0.63) |
| Investigator-agreed x NIOCCS high: 2 digit | 115 (244) | 0.84a (0.77-0.91) |
| Investigator-agreed x NIOCCS high: 2 digit (with write-in) | 129 (230) | 0.84a (0.78-0.91) |
| Investigator-agreed x NIOCCS high: 4 digit | 112 (247) | 0.70 (0.62-0.79) |
| Investigator-agreed x NIOCCS high: 4 digit (with write-in) | 129 (230) | 0.68 (0.60-0.76) |
| Investigator-agreed x NIOCCS med: 2 digit | 177 (182) | 0.71 (0.65-0.78) |
| Investigator-agreed x NIOCCS med: 2 digit (with write-in) | 207 (152) | 0.71 (0.64-0.77) |
| Investigator-agreed x NIOCCS med: 4 digit | 177 (182) | 0.60 (0.52-0.67) |
| Investigator-agreed x NIOCCS med: 4 digit (with write-in) | 207 (152) | 0.56 (0.49-0.63) |
a agreement = high