Finite sample effects in group-based trajectory models

Tom Loughran, Daniel S. Nagin

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

Two desirable properties of maximum likelihood-based parameter estimates are that the estimates are asymptotically unbiased and asymptotically normally distributed. In this article, the authors test whether the asymptotic properties of maximum likelihood estimation are achieved in sample sizes typically used in applications of group-based trajectory modeling. Through empirical results generated by resampling of population data, they find that the maximum likelihood estimates obtained in group-based trajectory models still provide reasonably close estimates of their true population values and have approximately normal distributions, even when estimated with a sample size as small as n = 500. Furthermore, and more important for the users of these types of models, the authors find similarly good performance in the model's ability to estimate the transformed quantities of main interest: the group trajectories and mixing probabilities.

Original languageEnglish (US)
Pages (from-to)250-278
Number of pages29
JournalSociological Methods and Research
Volume35
Issue number2
DOIs
StatePublished - Nov 1 2006

Fingerprint

Group
ability
performance
Values

All Science Journal Classification (ASJC) codes

  • Social Sciences (miscellaneous)
  • Sociology and Political Science

Cite this

@article{4dc8a352aa954babb00572f2cc86849b,
title = "Finite sample effects in group-based trajectory models",
abstract = "Two desirable properties of maximum likelihood-based parameter estimates are that the estimates are asymptotically unbiased and asymptotically normally distributed. In this article, the authors test whether the asymptotic properties of maximum likelihood estimation are achieved in sample sizes typically used in applications of group-based trajectory modeling. Through empirical results generated by resampling of population data, they find that the maximum likelihood estimates obtained in group-based trajectory models still provide reasonably close estimates of their true population values and have approximately normal distributions, even when estimated with a sample size as small as n = 500. Furthermore, and more important for the users of these types of models, the authors find similarly good performance in the model's ability to estimate the transformed quantities of main interest: the group trajectories and mixing probabilities.",
author = "Tom Loughran and Nagin, {Daniel S.}",
year = "2006",
month = "11",
day = "1",
doi = "10.1177/0049124106292292",
language = "English (US)",
volume = "35",
pages = "250--278",
journal = "Sociological Methods and Research",
issn = "0049-1241",
publisher = "SAGE Publications Inc.",
number = "2",

}

Finite sample effects in group-based trajectory models. / Loughran, Tom; Nagin, Daniel S.

In: Sociological Methods and Research, Vol. 35, No. 2, 01.11.2006, p. 250-278.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Finite sample effects in group-based trajectory models

AU - Loughran, Tom

AU - Nagin, Daniel S.

PY - 2006/11/1

Y1 - 2006/11/1

N2 - Two desirable properties of maximum likelihood-based parameter estimates are that the estimates are asymptotically unbiased and asymptotically normally distributed. In this article, the authors test whether the asymptotic properties of maximum likelihood estimation are achieved in sample sizes typically used in applications of group-based trajectory modeling. Through empirical results generated by resampling of population data, they find that the maximum likelihood estimates obtained in group-based trajectory models still provide reasonably close estimates of their true population values and have approximately normal distributions, even when estimated with a sample size as small as n = 500. Furthermore, and more important for the users of these types of models, the authors find similarly good performance in the model's ability to estimate the transformed quantities of main interest: the group trajectories and mixing probabilities.

AB - Two desirable properties of maximum likelihood-based parameter estimates are that the estimates are asymptotically unbiased and asymptotically normally distributed. In this article, the authors test whether the asymptotic properties of maximum likelihood estimation are achieved in sample sizes typically used in applications of group-based trajectory modeling. Through empirical results generated by resampling of population data, they find that the maximum likelihood estimates obtained in group-based trajectory models still provide reasonably close estimates of their true population values and have approximately normal distributions, even when estimated with a sample size as small as n = 500. Furthermore, and more important for the users of these types of models, the authors find similarly good performance in the model's ability to estimate the transformed quantities of main interest: the group trajectories and mixing probabilities.

UR - http://www.scopus.com/inward/record.url?scp=33749823396&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33749823396&partnerID=8YFLogxK

U2 - 10.1177/0049124106292292

DO - 10.1177/0049124106292292

M3 - Article

AN - SCOPUS:33749823396

VL - 35

SP - 250

EP - 278

JO - Sociological Methods and Research

JF - Sociological Methods and Research

SN - 0049-1241

IS - 2

ER -