Using residualized change versus difference scores for longitudinal research

Laura Castro-Schilo, Kevin Grimm

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

Researchers interested in studying change over time are often faced with an analytical conundrum: Whether a residualized change model versus a difference score model should be used to assess the effect of a key predictor on change that took place between two occasions. In this article, the authors pose a motivating example in which a researcher wants to investigate the effect of cohabitation on pre- to post-marriage change in relationship satisfaction. Key features of this example include the likely self-selection of dyads with lower relationship satisfaction to cohabit and the impossibility of using experimentation procedures to attain equivalent groups (i.e., cohabitants vs. not cohabitants). The authors use this example of a nonrandomized study to compare the residualized change and difference score models analytically and empirically. The authors describe the assumptions of the models to explain Lord’s paradox; that is, the fact that these models can lead to different inferences about the effect under investigation. They also provide recommendations for modeling data from nonrandomized studies using a latent change score framework.

Original languageEnglish (US)
Pages (from-to)32-58
Number of pages27
JournalJournal of Social and Personal Relationships
Volume35
Issue number1
DOIs
StatePublished - Jan 1 2018

Fingerprint

Research Personnel
Marriage
Research
cohabitation
dyad
Data structures
marriage
Group

Keywords

  • ANCOVA
  • difference scores
  • latent change scores
  • latent difference scores
  • longitudinal analysis
  • Lord’s paradox
  • residualized change

ASJC Scopus subject areas

  • Social Psychology
  • Communication
  • Developmental and Educational Psychology
  • Sociology and Political Science

Cite this

Using residualized change versus difference scores for longitudinal research. / Castro-Schilo, Laura; Grimm, Kevin.

In: Journal of Social and Personal Relationships, Vol. 35, No. 1, 01.01.2018, p. 32-58.

Research output: Contribution to journalArticle

@article{d1711cc6128c45179e3c12c0dcc2b8bb,
title = "Using residualized change versus difference scores for longitudinal research",
abstract = "Researchers interested in studying change over time are often faced with an analytical conundrum: Whether a residualized change model versus a difference score model should be used to assess the effect of a key predictor on change that took place between two occasions. In this article, the authors pose a motivating example in which a researcher wants to investigate the effect of cohabitation on pre- to post-marriage change in relationship satisfaction. Key features of this example include the likely self-selection of dyads with lower relationship satisfaction to cohabit and the impossibility of using experimentation procedures to attain equivalent groups (i.e., cohabitants vs. not cohabitants). The authors use this example of a nonrandomized study to compare the residualized change and difference score models analytically and empirically. The authors describe the assumptions of the models to explain Lord’s paradox; that is, the fact that these models can lead to different inferences about the effect under investigation. They also provide recommendations for modeling data from nonrandomized studies using a latent change score framework.",
keywords = "ANCOVA, difference scores, latent change scores, latent difference scores, longitudinal analysis, Lord’s paradox, residualized change",
author = "Laura Castro-Schilo and Kevin Grimm",
year = "2018",
month = "1",
day = "1",
doi = "10.1177/0265407517718387",
language = "English (US)",
volume = "35",
pages = "32--58",
journal = "Journal of Social and Personal Relationships",
issn = "0265-4075",
publisher = "SAGE Publications Ltd",
number = "1",

}

TY - JOUR

T1 - Using residualized change versus difference scores for longitudinal research

AU - Castro-Schilo, Laura

AU - Grimm, Kevin

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Researchers interested in studying change over time are often faced with an analytical conundrum: Whether a residualized change model versus a difference score model should be used to assess the effect of a key predictor on change that took place between two occasions. In this article, the authors pose a motivating example in which a researcher wants to investigate the effect of cohabitation on pre- to post-marriage change in relationship satisfaction. Key features of this example include the likely self-selection of dyads with lower relationship satisfaction to cohabit and the impossibility of using experimentation procedures to attain equivalent groups (i.e., cohabitants vs. not cohabitants). The authors use this example of a nonrandomized study to compare the residualized change and difference score models analytically and empirically. The authors describe the assumptions of the models to explain Lord’s paradox; that is, the fact that these models can lead to different inferences about the effect under investigation. They also provide recommendations for modeling data from nonrandomized studies using a latent change score framework.

AB - Researchers interested in studying change over time are often faced with an analytical conundrum: Whether a residualized change model versus a difference score model should be used to assess the effect of a key predictor on change that took place between two occasions. In this article, the authors pose a motivating example in which a researcher wants to investigate the effect of cohabitation on pre- to post-marriage change in relationship satisfaction. Key features of this example include the likely self-selection of dyads with lower relationship satisfaction to cohabit and the impossibility of using experimentation procedures to attain equivalent groups (i.e., cohabitants vs. not cohabitants). The authors use this example of a nonrandomized study to compare the residualized change and difference score models analytically and empirically. The authors describe the assumptions of the models to explain Lord’s paradox; that is, the fact that these models can lead to different inferences about the effect under investigation. They also provide recommendations for modeling data from nonrandomized studies using a latent change score framework.

KW - ANCOVA

KW - difference scores

KW - latent change scores

KW - latent difference scores

KW - longitudinal analysis

KW - Lord’s paradox

KW - residualized change

UR - http://www.scopus.com/inward/record.url?scp=85038850723&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85038850723&partnerID=8YFLogxK

U2 - 10.1177/0265407517718387

DO - 10.1177/0265407517718387

M3 - Article

AN - SCOPUS:85038850723

VL - 35

SP - 32

EP - 58

JO - Journal of Social and Personal Relationships

JF - Journal of Social and Personal Relationships

SN - 0265-4075

IS - 1

ER -