Statistical Techniques For Comparing Measurers And Methods Of Measurement: A Critical Review

SUMMARY 1. Clinical and experimental pharmacologists and physiologists often wish to compare two methods of measurement, or two measurers. 2. Biostatisticians insist that what should be sought is not agreement between methods or measurers, but disagreement or bias. 3. If measurements have been made...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Clinical and experimental pharmacology & physiology Jg. 29; H. 7; S. 527 - 536
1. Verfasser: Ludbrook, John
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Melbourne, Australia Blackwell Science Pty 01.07.2002
Schlagworte:
Bias
categorical variables
continuous variables
correlation
fixed bias
kappa statistic
least products regression analysis
limits of agreement
log-linear modelling
McNemar test
method of differences
proportional bias
Reproducibility of Results
Research Design - statistics & numerical data
Statistics as Topic - methods
ISSN:0305-1870, 1440-1681
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:SUMMARY 1. Clinical and experimental pharmacologists and physiologists often wish to compare two methods of measurement, or two measurers. 2. Biostatisticians insist that what should be sought is not agreement between methods or measurers, but disagreement or bias. 3. If measurements have been made on a continuous scale, the main choice is between the Altman–Bland method of differences and least products regression analysis. It is argued that although the former is relatively simple to execute, it does not distinguish adequately between fixed and proportional bias. Least products regression analysis, although more difficult to execute, does achieve this goal. There is almost universal agreement among biostatisticians that the Pearson product–moment correlation coefficient (r) is valueless as a test for bias. 4. If measurements have been made on a categorical scale, unordered or ordered, the most popular method of analysis is to use the kappa statistic. If the categories are unordered, the unweighted kappa statistic (K) is appropriate. If the categories are ordered, as they are in most rating scales in clinical, psychological and epidemiological research, the weighted kappa statistic (Kw) is preferable. But Kw corresponds to the intraclass correlation coefficient, which, like r for continuous variables, is incapable of detecting bias. Simple techniques for detecting bias in the case of ordered categorical variables are described and commended to investigators.
Bibliographie:istex:AB7D1C4DB8F9A2D65F633F35C5EE4A12F0169AD7
ArticleID:CEP3686
ark:/67375/WNG-VDSJ2ZDJ-G
ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
ObjectType-Review-3
content type line 23
ISSN:0305-1870
1440-1681
DOI:10.1046/j.1440-1681.2002.03686.x