Grubbs' test for outliers

URI: https://gptkb.org/entity/Grubbs'_test_for_outliers
GPTKB entity
Statements (26)
Predicate Object
gptkbp:instanceOf statistical analysis
gptkbp:alsoKnownAs gptkb:Grubbs'_test
maximum normalized residual test
gptkbp:alternativeTo one outlier exists in the data
gptkbp:appliesTo normally distributed data
gptkbp:assumes data are normally distributed
data are independent
gptkbp:canBeRepeated to detect multiple outliers
gptkbp:category outlier detection
parametric test
gptkbp:detects single outlier
gptkbp:developedBy gptkb:Frank_E._Grubbs
https://www.w3.org/2000/01/rdf-schema#label Grubbs' test for outliers
gptkbp:introducedIn 1950
gptkbp:limitation not robust to multiple outliers
sensitive to non-normality
gptkbp:nullHypothesis no outliers in the data
gptkbp:publishedIn gptkb:Annals_of_Mathematical_Statistics
gptkbp:relatedTo gptkb:Dixon's_Q_test
gptkb:Generalized_ESD_test
gptkbp:requires sample size n > 6
gptkbp:statistic G
gptkbp:testStatisticFormula G = max(|Yi - Ymean|) / s
gptkbp:usedFor detecting outliers in univariate data
gptkbp:bfsParent gptkb:Frank_E._Grubbs
gptkbp:bfsLayer 7