The Impact of Missing Risk Factor Data on Semiparametric Group-Based Trajectory Models

James V. Ray, Christopher J. Sullivan, Thomas Loughran, Shayne E. Jones

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Purpose: To investigate how missing data (Missing Completely at Random [MCAR] vs. Missing Not at Random [MNAR]) on risk factors can impact trajectory solutions (i.e., latent class probabilities) and coefficient estimates capturing the relationship between covariates and trajectory group solutions using a semiparametric group-based trajectory modeling (GBTM) approach. Methods: To address this issue, we conducted a systematic investigation using Monte Carlo simulation. Data were generated from a population with known growth parameters and risk factors. Observations for risk factors were then systematically deleted in a way that reflects key missing data assumptions (MCAR and MNAR). Models were then estimated to test the sensitivity of the estimates to each missing data scenario. Results: Two key findings emerged: (1) trajectory solutions were largely unaffected by missing data on risk factors; and, (2) there was some degree of bias in estimating relationships between risk factors and trajectory group membership when data were missing on those risk factors. Conclusions: GBTM may be useful for testing etiological explanations of long-term patterns of offending. Missing data on risk factors poses a threat to this approach, however.

Original languageEnglish (US)
Pages (from-to)276-296
Number of pages21
JournalJournal of Developmental and Life-Course Criminology
Volume4
Issue number3
DOIs
StatePublished - Sep 1 2018

Fingerprint

Group
group membership
Intercellular Signaling Peptides and Proteins
threat
scenario
simulation
trend
Population

All Science Journal Classification (ASJC) codes

  • Applied Psychology
  • Law
  • Life-span and Life-course Studies

Cite this

@article{c106bfc601844c00a6d4f3e4642dd9e9,
title = "The Impact of Missing Risk Factor Data on Semiparametric Group-Based Trajectory Models",
abstract = "Purpose: To investigate how missing data (Missing Completely at Random [MCAR] vs. Missing Not at Random [MNAR]) on risk factors can impact trajectory solutions (i.e., latent class probabilities) and coefficient estimates capturing the relationship between covariates and trajectory group solutions using a semiparametric group-based trajectory modeling (GBTM) approach. Methods: To address this issue, we conducted a systematic investigation using Monte Carlo simulation. Data were generated from a population with known growth parameters and risk factors. Observations for risk factors were then systematically deleted in a way that reflects key missing data assumptions (MCAR and MNAR). Models were then estimated to test the sensitivity of the estimates to each missing data scenario. Results: Two key findings emerged: (1) trajectory solutions were largely unaffected by missing data on risk factors; and, (2) there was some degree of bias in estimating relationships between risk factors and trajectory group membership when data were missing on those risk factors. Conclusions: GBTM may be useful for testing etiological explanations of long-term patterns of offending. Missing data on risk factors poses a threat to this approach, however.",
author = "Ray, {James V.} and Sullivan, {Christopher J.} and Thomas Loughran and Jones, {Shayne E.}",
year = "2018",
month = "9",
day = "1",
doi = "10.1007/s40865-018-0085-x",
language = "English (US)",
volume = "4",
pages = "276--296",
journal = "Journal of Developmental and Life-Course Criminology",
issn = "2199-4641",
publisher = "Springer International Publishing AG",
number = "3",

}

The Impact of Missing Risk Factor Data on Semiparametric Group-Based Trajectory Models. / Ray, James V.; Sullivan, Christopher J.; Loughran, Thomas; Jones, Shayne E.

In: Journal of Developmental and Life-Course Criminology, Vol. 4, No. 3, 01.09.2018, p. 276-296.

Research output: Contribution to journalArticle

TY - JOUR

T1 - The Impact of Missing Risk Factor Data on Semiparametric Group-Based Trajectory Models

AU - Ray, James V.

AU - Sullivan, Christopher J.

AU - Loughran, Thomas

AU - Jones, Shayne E.

PY - 2018/9/1

Y1 - 2018/9/1

N2 - Purpose: To investigate how missing data (Missing Completely at Random [MCAR] vs. Missing Not at Random [MNAR]) on risk factors can impact trajectory solutions (i.e., latent class probabilities) and coefficient estimates capturing the relationship between covariates and trajectory group solutions using a semiparametric group-based trajectory modeling (GBTM) approach. Methods: To address this issue, we conducted a systematic investigation using Monte Carlo simulation. Data were generated from a population with known growth parameters and risk factors. Observations for risk factors were then systematically deleted in a way that reflects key missing data assumptions (MCAR and MNAR). Models were then estimated to test the sensitivity of the estimates to each missing data scenario. Results: Two key findings emerged: (1) trajectory solutions were largely unaffected by missing data on risk factors; and, (2) there was some degree of bias in estimating relationships between risk factors and trajectory group membership when data were missing on those risk factors. Conclusions: GBTM may be useful for testing etiological explanations of long-term patterns of offending. Missing data on risk factors poses a threat to this approach, however.

AB - Purpose: To investigate how missing data (Missing Completely at Random [MCAR] vs. Missing Not at Random [MNAR]) on risk factors can impact trajectory solutions (i.e., latent class probabilities) and coefficient estimates capturing the relationship between covariates and trajectory group solutions using a semiparametric group-based trajectory modeling (GBTM) approach. Methods: To address this issue, we conducted a systematic investigation using Monte Carlo simulation. Data were generated from a population with known growth parameters and risk factors. Observations for risk factors were then systematically deleted in a way that reflects key missing data assumptions (MCAR and MNAR). Models were then estimated to test the sensitivity of the estimates to each missing data scenario. Results: Two key findings emerged: (1) trajectory solutions were largely unaffected by missing data on risk factors; and, (2) there was some degree of bias in estimating relationships between risk factors and trajectory group membership when data were missing on those risk factors. Conclusions: GBTM may be useful for testing etiological explanations of long-term patterns of offending. Missing data on risk factors poses a threat to this approach, however.

UR - http://www.scopus.com/inward/record.url?scp=85068067869&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068067869&partnerID=8YFLogxK

U2 - 10.1007/s40865-018-0085-x

DO - 10.1007/s40865-018-0085-x

M3 - Article

VL - 4

SP - 276

EP - 296

JO - Journal of Developmental and Life-Course Criminology

JF - Journal of Developmental and Life-Course Criminology

SN - 2199-4641

IS - 3

ER -