Work

Lipid concentrations in standard reference material (SRM) 1950: results from an interlaboratory comparison exercise for lipidomics

Year: 2017

Type: report

Abstract: Lipid consensus estimates and associated uncertainties were determined using three methods, the Vangel-Rukhin (VR), the DerSimonian-Laird (DSL), and the median of means (MEDM). The VR method (8, 9) is a weighted mean based on maximum likelihood. The derivation and mathematical details of the VR mean and its associated standard uncertainty are given in a previous publication This method incorporates both inter-laboratory and intra-laboratory variances in determining the weights. Maximum likelihood approaches have excellent statistical properties; however, the VR method is not robust against extreme outliers. In these cases, the VR weighting method gives the outlying laboratory value greater influence and results in unrepresentative estimates (i.e., the resulting value reflects neither the non-outlying values or the outlying value). Thus, to obtain a reasonable mean using this approach, the extreme outlying lipid values would need to be omitted from the estimation. The VR means were calculated with no outliers omitted.

DOI: 10.6028/nist.ir.8185

Authors: John A Bowden, Alan Heckert, Candice Z Ulmer, Christina M Jones

Institutions: National Institute of Standards and Technology, Information Technology Laboratory, Material Measurement Laboratory

Language: English

Topic: Scientific Measurement and Uncertainty Evaluation

Subfield: Statistics, Probability and Uncertainty

Field: Decision Sciences

Domain: Social Sciences

Open Access status: gold

Funders: Brigham and Women's Hospital, Karolinska Institutet, Baker Heart and Diabetes Institute, National Institute of Standards and Technology, Information Technology Laboratory, Material Measurement Laboratory, H2020 European Research Council

OpenAlex ID: W2741748676

Cites: 14

Cited by: 15

Related to: 10