2.50
Hdl Handle:
http://hdl.handle.net/10755/159295
Type:
Presentation
Title:
How to Make Decisions in Centering Predictors for Multilevel Data Analysis
Abstract:
How to Make Decisions in Centering Predictors for Multilevel Data Analysis
Conference Sponsor:Midwest Nursing Research Society
Conference Year:2004
Author:Wu, Yow-Wu, PhD
P.I. Institution Name:University of Buffalo/SUNY
Title:Associate Professor
Contact Address:SON, 920 Kimball Tower, Buffalo, NY, 14214, USA
Co-Authors:Powhatan Wooldridge
In multilevel modeling, centering issues must be addressed for each independent variable at each level of analysis. There are three decisions that one can make in modeling independent variables at the first level of analysis, not centering (using raw scores), group mean centering, and grand mean centering. Each of these approaches implies a different interpretation of the intercept of the resulting regression equation. At the second or higher level, similar choices must be made, with the added complication that the approaches used at the lower levels may affect the meaning of the approaches used at the higher levels. While most of these issues have been discussed in the statistical literature, the implications of which centering approach is used for each variable and level have not been adequately discussed in the nursing literature. In addition, many complex substantive issues of conceptualization and interpretation for specific kinds of variables remain largely unexplored. The purpose of this presentation is to illustrate the effects of various choices in centering at each level of a two level analysis, and to discuss which centering choices are optimal from a conceptual or interpretive point of view. We will model the first level by using one predictor in three different formats -- raw data, centering by group mean and centering by grand mean. We will then model the second level data in three different ways: (1) without a predictor, (2) with a contextual predictor that is the group mean of the first level predictor without grand mean centering at the second level and (3) with a contextual predictor that is the group mean of the first level predictor with grand mean centering at the second level. This study will provide researchers using multilevel data analysis with a better understanding of how properly to model and interpret their findings.
Repository Posting Date:
26-Oct-2011
Date of Publication:
17-Oct-2011
Sponsors:
Midwest Nursing Research Society

Full metadata record

DC FieldValue Language
dc.typePresentationen_GB
dc.titleHow to Make Decisions in Centering Predictors for Multilevel Data Analysisen_GB
dc.identifier.urihttp://hdl.handle.net/10755/159295-
dc.description.abstract<table><tr><td colspan="2" class="item-title">How to Make Decisions in Centering Predictors for Multilevel Data Analysis </td></tr><tr class="item-sponsor"><td class="label">Conference Sponsor:</td><td class="value">Midwest Nursing Research Society</td></tr><tr class="item-year"><td class="label">Conference Year:</td><td class="value">2004</td></tr><tr class="item-author"><td class="label">Author:</td><td class="value">Wu, Yow-Wu, PhD</td></tr><tr class="item-institute"><td class="label">P.I. Institution Name:</td><td class="value">University of Buffalo/SUNY</td></tr><tr class="item-author-title"><td class="label">Title:</td><td class="value">Associate Professor</td></tr><tr class="item-address"><td class="label">Contact Address:</td><td class="value">SON, 920 Kimball Tower, Buffalo, NY, 14214, USA</td></tr><tr class="item-co-authors"><td class="label">Co-Authors:</td><td class="value">Powhatan Wooldridge</td></tr><tr><td colspan="2" class="item-abstract">In multilevel modeling, centering issues must be addressed for each independent variable at each level of analysis. There are three decisions that one can make in modeling independent variables at the first level of analysis, not centering (using raw scores), group mean centering, and grand mean centering. Each of these approaches implies a different interpretation of the intercept of the resulting regression equation. At the second or higher level, similar choices must be made, with the added complication that the approaches used at the lower levels may affect the meaning of the approaches used at the higher levels. While most of these issues have been discussed in the statistical literature, the implications of which centering approach is used for each variable and level have not been adequately discussed in the nursing literature. In addition, many complex substantive issues of conceptualization and interpretation for specific kinds of variables remain largely unexplored. The purpose of this presentation is to illustrate the effects of various choices in centering at each level of a two level analysis, and to discuss which centering choices are optimal from a conceptual or interpretive point of view. We will model the first level by using one predictor in three different formats -- raw data, centering by group mean and centering by grand mean. We will then model the second level data in three different ways: (1) without a predictor, (2) with a contextual predictor that is the group mean of the first level predictor without grand mean centering at the second level and (3) with a contextual predictor that is the group mean of the first level predictor with grand mean centering at the second level. This study will provide researchers using multilevel data analysis with a better understanding of how properly to model and interpret their findings.</td></tr></table>en_GB
dc.date.available2011-10-26T21:53:02Z-
dc.date.issued2011-10-17en_GB
dc.date.accessioned2011-10-26T21:53:02Z-
dc.description.sponsorshipMidwest Nursing Research Societyen_GB
All Items in this repository are protected by copyright, with all rights reserved, unless otherwise indicated.