On Reverse Shrinkage Effects and Shrinkage Overshoot

Pascal Jordan

doi:10.1007/s11336-022-09872-8

On Reverse Shrinkage Effects and Shrinkage Overshoot

Standard

On Reverse Shrinkage Effects and Shrinkage Overshoot. / Jordan, Pascal.

in: PSYCHOMETRIKA, Jahrgang 88, Nr. 1, 03.2023, S. 274-301.

Publikationen: SCORING: Beitrag in Fachzeitschrift/Zeitung › SCORING: Zeitschriftenaufsatz › Forschung › Begutachtung

Harvard

Jordan, P 2023, 'On Reverse Shrinkage Effects and Shrinkage Overshoot', PSYCHOMETRIKA, Jg. 88, Nr. 1, S. 274-301. https://doi.org/10.1007/s11336-022-09872-8

APA

Jordan, P. (2023). On Reverse Shrinkage Effects and Shrinkage Overshoot. PSYCHOMETRIKA, 88(1), 274-301. https://doi.org/10.1007/s11336-022-09872-8

Vancouver

Jordan P. On Reverse Shrinkage Effects and Shrinkage Overshoot. PSYCHOMETRIKA. 2023 Mär;88(1):274-301. https://doi.org/10.1007/s11336-022-09872-8

Bibtex

@article{711c5d9968c749049f6fcf3b1338a717,

title = "On Reverse Shrinkage Effects and Shrinkage Overshoot",

abstract = "Given a squared Euclidean norm penalty, we examine some less well-known properties of shrinkage estimates. In particular, we highlight that it is possible for some components of the shrinkage estimator to be placed further away from the prior mean than the original estimate. An analysis of this effect is provided within three different modeling settings-encompassing linear, logistic, and ordinal regression models. Additional simulations show that the outlined effect is not a mathematical artefact, but likely to occur in practice. As a byproduct, they also highlight the possibilities of sign reversals ({"}overshoots{"}) for shrinkage estimates. We point out practical consequences and challenges, which might arise from the observed effects with special emphasis on psychometrics.",

author = "Pascal Jordan",

note = "{\textcopyright} 2022. The Author(s).",

year = "2023",

month = mar,

doi = "10.1007/s11336-022-09872-8",

language = "English",

volume = "88",

pages = "274--301",

journal = "PSYCHOMETRIKA",

issn = "0033-3123",

publisher = "Springer New York",

number = "1",

}

RIS

TY - JOUR

T1 - On Reverse Shrinkage Effects and Shrinkage Overshoot

AU - Jordan, Pascal

PY - 2023/3

Y1 - 2023/3

N2 - Given a squared Euclidean norm penalty, we examine some less well-known properties of shrinkage estimates. In particular, we highlight that it is possible for some components of the shrinkage estimator to be placed further away from the prior mean than the original estimate. An analysis of this effect is provided within three different modeling settings-encompassing linear, logistic, and ordinal regression models. Additional simulations show that the outlined effect is not a mathematical artefact, but likely to occur in practice. As a byproduct, they also highlight the possibilities of sign reversals ("overshoots") for shrinkage estimates. We point out practical consequences and challenges, which might arise from the observed effects with special emphasis on psychometrics.

AB - Given a squared Euclidean norm penalty, we examine some less well-known properties of shrinkage estimates. In particular, we highlight that it is possible for some components of the shrinkage estimator to be placed further away from the prior mean than the original estimate. An analysis of this effect is provided within three different modeling settings-encompassing linear, logistic, and ordinal regression models. Additional simulations show that the outlined effect is not a mathematical artefact, but likely to occur in practice. As a byproduct, they also highlight the possibilities of sign reversals ("overshoots") for shrinkage estimates. We point out practical consequences and challenges, which might arise from the observed effects with special emphasis on psychometrics.

U2 - 10.1007/s11336-022-09872-8

DO - 10.1007/s11336-022-09872-8

M3 - SCORING: Journal article

C2 - 35665449

VL - 88

SP - 274

EP - 301

JO - PSYCHOMETRIKA

JF - PSYCHOMETRIKA

SN - 0033-3123

IS - 1

ER -