Accurate Inference from Chemical Measurement Data Under the Rocke-Lorenzato Model
Year of Publication
Joint Statistical Meeting
Coverage Probability; Higher order asymptotic; Log-likelihood ratio statistic; Rocke and Lorenzato Model
Rocke and Lorenzato (1995) proposed a two-component model for measurement error for calibration analysis in analytical chemistry. Their analysis is likelihood based, and the coverage probabilities of the resulting confidence intervals are not always satisfactory, unless the sample sizes are large. In our research, higher order asymptotic procedures are used to obtain more accurate inferences when the sample sizes are small. An algorithm will be provided to implement our methodology. Numerical results and illustrative examples will be provided.