The Effect of Different Forms of Centering in Hierarchical Linear Models

Ita G G Kreft, Jan de Leeuw, Leona S. Aiken

Research output: Contribution to journalArticle

484 Citations (Scopus)

Abstract

Multilevel models are becoming increasingly used in applied educational social and economic research for the analysis of hierarchically nested data. In these random coefficient regression models the parameters are allowed to differ over the groups in which the observations are nested. For computational ease in deriving parameter estimates, predictors are often centered around the mean. In nested or grouped data, the option of centering around the grand mean is extended with an option to center within groups or contexts. Both are statistically sound ways to improve parameter estimation. In this article we study the effects of these two different ways of centering, in comparison to the use of raw scores, on the parameter estimates in random coefficient models. The conclusion is that centering around the group mean amounts to fitting a different model from that obtained by centering around the grand mean or by using raw scores. The choice between the two options for centering can only be made on a theoretical basis. Based on this study, we conclude that centering rules valid for simple models, such as the fixed coefficients regression model, are no longer applicable to more complicated models, such as the random coefficient model. We think researchers should be made aware of the consequences of the choice of particular centering options.

Original languageEnglish (US)
Pages (from-to)1-21
Number of pages21
JournalMultivariate Behavioral Research
Volume30
Issue number1
DOIs
StatePublished - Jan 1 1995

Fingerprint

Hierarchical Linear Models
Random Coefficient Models
Linear Models
Economics
Research Personnel
Research
Regression Model
Grouped Data
Multilevel Models
Estimate
Parameter Estimation
Predictors
Model
Valid
Form
Coefficient

ASJC Scopus subject areas

  • Arts and Humanities (miscellaneous)
  • Statistics and Probability
  • Experimental and Cognitive Psychology

Cite this

The Effect of Different Forms of Centering in Hierarchical Linear Models. / Kreft, Ita G G; de Leeuw, Jan; Aiken, Leona S.

In: Multivariate Behavioral Research, Vol. 30, No. 1, 01.01.1995, p. 1-21.

Research output: Contribution to journalArticle

Kreft, Ita G G ; de Leeuw, Jan ; Aiken, Leona S. / The Effect of Different Forms of Centering in Hierarchical Linear Models. In: Multivariate Behavioral Research. 1995 ; Vol. 30, No. 1. pp. 1-21.
@article{dd9cc54426a64fbf806b134ef100c145,
title = "The Effect of Different Forms of Centering in Hierarchical Linear Models",
abstract = "Multilevel models are becoming increasingly used in applied educational social and economic research for the analysis of hierarchically nested data. In these random coefficient regression models the parameters are allowed to differ over the groups in which the observations are nested. For computational ease in deriving parameter estimates, predictors are often centered around the mean. In nested or grouped data, the option of centering around the grand mean is extended with an option to center within groups or contexts. Both are statistically sound ways to improve parameter estimation. In this article we study the effects of these two different ways of centering, in comparison to the use of raw scores, on the parameter estimates in random coefficient models. The conclusion is that centering around the group mean amounts to fitting a different model from that obtained by centering around the grand mean or by using raw scores. The choice between the two options for centering can only be made on a theoretical basis. Based on this study, we conclude that centering rules valid for simple models, such as the fixed coefficients regression model, are no longer applicable to more complicated models, such as the random coefficient model. We think researchers should be made aware of the consequences of the choice of particular centering options.",
author = "Kreft, {Ita G G} and {de Leeuw}, Jan and Aiken, {Leona S.}",
year = "1995",
month = "1",
day = "1",
doi = "10.1207/s15327906mbr3001_1",
language = "English (US)",
volume = "30",
pages = "1--21",
journal = "Multivariate Behavioral Research",
issn = "0027-3171",
publisher = "Psychology Press Ltd",
number = "1",

}

TY - JOUR

T1 - The Effect of Different Forms of Centering in Hierarchical Linear Models

AU - Kreft, Ita G G

AU - de Leeuw, Jan

AU - Aiken, Leona S.

PY - 1995/1/1

Y1 - 1995/1/1

N2 - Multilevel models are becoming increasingly used in applied educational social and economic research for the analysis of hierarchically nested data. In these random coefficient regression models the parameters are allowed to differ over the groups in which the observations are nested. For computational ease in deriving parameter estimates, predictors are often centered around the mean. In nested or grouped data, the option of centering around the grand mean is extended with an option to center within groups or contexts. Both are statistically sound ways to improve parameter estimation. In this article we study the effects of these two different ways of centering, in comparison to the use of raw scores, on the parameter estimates in random coefficient models. The conclusion is that centering around the group mean amounts to fitting a different model from that obtained by centering around the grand mean or by using raw scores. The choice between the two options for centering can only be made on a theoretical basis. Based on this study, we conclude that centering rules valid for simple models, such as the fixed coefficients regression model, are no longer applicable to more complicated models, such as the random coefficient model. We think researchers should be made aware of the consequences of the choice of particular centering options.

AB - Multilevel models are becoming increasingly used in applied educational social and economic research for the analysis of hierarchically nested data. In these random coefficient regression models the parameters are allowed to differ over the groups in which the observations are nested. For computational ease in deriving parameter estimates, predictors are often centered around the mean. In nested or grouped data, the option of centering around the grand mean is extended with an option to center within groups or contexts. Both are statistically sound ways to improve parameter estimation. In this article we study the effects of these two different ways of centering, in comparison to the use of raw scores, on the parameter estimates in random coefficient models. The conclusion is that centering around the group mean amounts to fitting a different model from that obtained by centering around the grand mean or by using raw scores. The choice between the two options for centering can only be made on a theoretical basis. Based on this study, we conclude that centering rules valid for simple models, such as the fixed coefficients regression model, are no longer applicable to more complicated models, such as the random coefficient model. We think researchers should be made aware of the consequences of the choice of particular centering options.

UR - http://www.scopus.com/inward/record.url?scp=84861194072&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84861194072&partnerID=8YFLogxK

U2 - 10.1207/s15327906mbr3001_1

DO - 10.1207/s15327906mbr3001_1

M3 - Article

VL - 30

SP - 1

EP - 21

JO - Multivariate Behavioral Research

JF - Multivariate Behavioral Research

SN - 0027-3171

IS - 1

ER -