Statistical Learning and Data Science

Bayesian Analysis of St ructural Equation Models Using Parameter Expansion 145

TABLE 12.2

Consensus value, associa ted uncertainty, and confidence intervals.

Method Cons. Val. As. Uncert. 95 % Conf. Int.

Reference 33 1.72 [29.55, 36.45]

Robust method 34.46 1.64 [31.18, 37.75]

Exper t labs 34 1.6 [30.8, 37.2]

Robustified CV 33.82 1.62 [30.57, 37.06]

Due to the small number of laboratories, estimates of SEM are of poor quality, with

relatively high standard deviations, so that this application cannot be used in the gen-

eral purpose of testing relationships between preanalytical step, analytical step, and quality

control.

The implementation of the methodology ...