Establishing causal order in longitudinal studies combining binary and continuous dependent variables

Gerhard Kling*, Charles Harvey, Mairi Maclean

*Corresponding author for this work

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Longitudinal studies with a mix of binary outcomes and continuous variables are common in organizational research. Selecting the dependent variable is often difficult due to conflicting theories and contradictory empirical studies. In addition, organizational researchers are confronted with methodological challenges posed by latent variables relating to observed binary outcomes and within-subject correlation. We draw on Dueker?s (2005) qualitative vector autoregression (QVAR) and Lunn et al.?s (2014) multivariate probit model to develop a solution to these problems in the form of a qualitative short panel vector autoregression (QSP-VAR). The QSP-VAR combines binary and continuous variables into a single vector of dependent variables, making every variable endogenous a priori. The QSP-VAR identifies causal order, reveals within-subject correlation and accounts for latent variables. Using a Bayesian approach, the QSP-VAR provides reliable inference for short time dimension longitudinal research. This is demonstrated through analysis of the durability of elite corporate agents, social networks and firm performance in France. We provide our OpenBUGS code to enable implementation of the QSP-VAR by other researchers.
Original languageEnglish
Pages (from-to)770-799
Number of pages30
JournalOrganizational Research Methods
Volume20
Issue number4
Early online date30 Nov 2015
DOIs
Publication statusPublished - 1 Oct 2017

Fingerprint

Longitudinal study
Vector autoregression
Latent variables
Elites
Durability
Inference
Empirical study
Organizational research
Probit model
Social networks
Firm performance
Multivariate probit
Endogenous variables
Longitudinal research
Bayesian approach
France

Keywords

  • Bayesian statistics
  • binary dependent variables
  • causality
  • longitudinal research
  • vector autoregression

Cite this

Establishing causal order in longitudinal studies combining binary and continuous dependent variables. / Kling, Gerhard; Harvey, Charles; Maclean, Mairi.

In: Organizational Research Methods, Vol. 20, No. 4, 01.10.2017, p. 770-799.

Research output: Contribution to journalArticle

@article{5e75f880fe7540619db0d69750edb780,
title = "Establishing causal order in longitudinal studies combining binary and continuous dependent variables",
abstract = "Longitudinal studies with a mix of binary outcomes and continuous variables are common in organizational research. Selecting the dependent variable is often difficult due to conflicting theories and contradictory empirical studies. In addition, organizational researchers are confronted with methodological challenges posed by latent variables relating to observed binary outcomes and within-subject correlation. We draw on Dueker?s (2005) qualitative vector autoregression (QVAR) and Lunn et al.?s (2014) multivariate probit model to develop a solution to these problems in the form of a qualitative short panel vector autoregression (QSP-VAR). The QSP-VAR combines binary and continuous variables into a single vector of dependent variables, making every variable endogenous a priori. The QSP-VAR identifies causal order, reveals within-subject correlation and accounts for latent variables. Using a Bayesian approach, the QSP-VAR provides reliable inference for short time dimension longitudinal research. This is demonstrated through analysis of the durability of elite corporate agents, social networks and firm performance in France. We provide our OpenBUGS code to enable implementation of the QSP-VAR by other researchers.",
keywords = "Bayesian statistics, binary dependent variables, causality, longitudinal research, vector autoregression",
author = "Gerhard Kling and Charles Harvey and Mairi Maclean",
note = "Acknowledgments We wish to thank the Editor, James LeBreton, the Associate Editor, Brian Boyd, and two anonymous referees for their most helpful comments. Without their efforts, our qualitative short panel vector autoregression (QSP-VAR) would not have reached its full potential. We had the benefit of advice on our OpenBUGS code from Dave Lunn, one of the developers of WinBUGS/OpenBUGS, for which we are grateful. Sadly, when we sought advice from Michael Dueker, who devised the QVAR and whose work inspired our own, we learned that he had passed away on January 29, 2015. Dueker’s research has had a significant impact in economics, econometrics, and related disciplines. We dedicate this paper to him in recognition of his many pioneering achievements. Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.",
year = "2017",
month = "10",
day = "1",
doi = "10.1177/1094428115618760",
language = "English",
volume = "20",
pages = "770--799",
journal = "Organizational Research Methods",
issn = "1552-7425",
publisher = "Sage Publications",
number = "4",

}

TY - JOUR

T1 - Establishing causal order in longitudinal studies combining binary and continuous dependent variables

AU - Kling, Gerhard

AU - Harvey, Charles

AU - Maclean, Mairi

N1 - Acknowledgments We wish to thank the Editor, James LeBreton, the Associate Editor, Brian Boyd, and two anonymous referees for their most helpful comments. Without their efforts, our qualitative short panel vector autoregression (QSP-VAR) would not have reached its full potential. We had the benefit of advice on our OpenBUGS code from Dave Lunn, one of the developers of WinBUGS/OpenBUGS, for which we are grateful. Sadly, when we sought advice from Michael Dueker, who devised the QVAR and whose work inspired our own, we learned that he had passed away on January 29, 2015. Dueker’s research has had a significant impact in economics, econometrics, and related disciplines. We dedicate this paper to him in recognition of his many pioneering achievements. Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

PY - 2017/10/1

Y1 - 2017/10/1

N2 - Longitudinal studies with a mix of binary outcomes and continuous variables are common in organizational research. Selecting the dependent variable is often difficult due to conflicting theories and contradictory empirical studies. In addition, organizational researchers are confronted with methodological challenges posed by latent variables relating to observed binary outcomes and within-subject correlation. We draw on Dueker?s (2005) qualitative vector autoregression (QVAR) and Lunn et al.?s (2014) multivariate probit model to develop a solution to these problems in the form of a qualitative short panel vector autoregression (QSP-VAR). The QSP-VAR combines binary and continuous variables into a single vector of dependent variables, making every variable endogenous a priori. The QSP-VAR identifies causal order, reveals within-subject correlation and accounts for latent variables. Using a Bayesian approach, the QSP-VAR provides reliable inference for short time dimension longitudinal research. This is demonstrated through analysis of the durability of elite corporate agents, social networks and firm performance in France. We provide our OpenBUGS code to enable implementation of the QSP-VAR by other researchers.

AB - Longitudinal studies with a mix of binary outcomes and continuous variables are common in organizational research. Selecting the dependent variable is often difficult due to conflicting theories and contradictory empirical studies. In addition, organizational researchers are confronted with methodological challenges posed by latent variables relating to observed binary outcomes and within-subject correlation. We draw on Dueker?s (2005) qualitative vector autoregression (QVAR) and Lunn et al.?s (2014) multivariate probit model to develop a solution to these problems in the form of a qualitative short panel vector autoregression (QSP-VAR). The QSP-VAR combines binary and continuous variables into a single vector of dependent variables, making every variable endogenous a priori. The QSP-VAR identifies causal order, reveals within-subject correlation and accounts for latent variables. Using a Bayesian approach, the QSP-VAR provides reliable inference for short time dimension longitudinal research. This is demonstrated through analysis of the durability of elite corporate agents, social networks and firm performance in France. We provide our OpenBUGS code to enable implementation of the QSP-VAR by other researchers.

KW - Bayesian statistics

KW - binary dependent variables

KW - causality

KW - longitudinal research

KW - vector autoregression

U2 - 10.1177/1094428115618760

DO - 10.1177/1094428115618760

M3 - Article

VL - 20

SP - 770

EP - 799

JO - Organizational Research Methods

JF - Organizational Research Methods

SN - 1552-7425

IS - 4

ER -