eTable 1.
(available online only). Average parameter estimates from regression calibration (RC) and inverse regression calibration (IRC) models.
| Calibration approach | Average parameter estimates | Model fit characteristics | ||||
|---|---|---|---|---|---|---|
| Model | Shape | Intercept | Slope | Square | −2LL | AIC |
| RC | ||||||
| Collapsed | L | −0.9813 | 1.2090 | N/A | −220.0 | −218.0 |
| Batch specific | L | −1.0521 | 1.2351 | N/A | −293.3 | −291.3 |
| Mixed models | L | −1.0428 | 1.2312 | N/A | −375.1 | −367.1 |
| Collapsed | CL | −0.7882 | 1.0574 | 0.0274 | −216.1 | −214.1 |
| Batch specific | CL | −1.0167 | 0.1667 | 0.0025 | −243.1 | −241.1 |
| Mixed models* | CL | ~ | ~ | ~ | ~ | ~ |
| IRC | ||||||
| Collapsed | L | 0.8709 | 0.8021 | −349.3 | −347.3 | |
| Batch specific | L | 0.8651 | 0.8029 | −382.6 | −380.6 | |
| Mixed models | L | 0.8657 | 0.8031 | −492.0 | −484.0 | |
| Collapsed | CL | 0.8847 | 0.7882 | 0.0029 | −341.9 | −339.9 |
| Batch specific | CL | 0.7675 | 0.8153 | −0.0076 | −341.5 | −339.5 |
| Mixed models* | CL | ~ | ~ | ~ | ~ | ~ |
*
Inestimable due to singularity of covariance matrix.
Abbreviations: L, linear; CL, curvilinear, -2LL, -2 log likelihood; AIC, Akaike Information Criterion.