From Data to Causes III: Bayesian Priors for General Cross-Lagged Panel Models (GCLM)
This article describes some potential uses of Bayesian estimation for time-series and panel data models by incorporating information from prior probabilities (i.e., priors) in addition to observed data. Drawing on econometrics and other literatures we illustrate the use of informative “shrinkage” or...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2021-02-01
|
| Series: | Frontiers in Psychology |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fpsyg.2021.612251/full |
| _version_ | 1819051028817903616 |
|---|---|
| author | Michael J. Zyphur Ellen L. Hamaker Louis Tay Manuel Voelkle Kristopher J. Preacher Zhen Zhang Zhen Zhang Paul D. Allison Dean C. Pierides Peter Koval Edward F. Diener Edward F. Diener |
| author_facet | Michael J. Zyphur Ellen L. Hamaker Louis Tay Manuel Voelkle Kristopher J. Preacher Zhen Zhang Zhen Zhang Paul D. Allison Dean C. Pierides Peter Koval Edward F. Diener Edward F. Diener |
| author_sort | Michael J. Zyphur |
| collection | DOAJ |
| description | This article describes some potential uses of Bayesian estimation for time-series and panel data models by incorporating information from prior probabilities (i.e., priors) in addition to observed data. Drawing on econometrics and other literatures we illustrate the use of informative “shrinkage” or “small variance” priors (including so-called “Minnesota priors”) while extending prior work on the general cross-lagged panel model (GCLM). Using a panel dataset of national income and subjective well-being (SWB) we describe three key benefits of these priors. First, they shrink parameter estimates toward zero or toward each other for time-varying parameters, which lends additional support for an income → SWB effect that is not supported with maximum likelihood (ML). This is useful because, second, these priors increase model parsimony and the stability of estimates (keeping them within more reasonable bounds) and thus improve out-of-sample predictions and interpretability, which means estimated effect should also be more trustworthy than under ML. Third, these priors allow estimating otherwise under-identified models under ML, allowing higher-order lagged effects and time-varying parameters that are otherwise impossible to estimate using observed data alone. In conclusion we note some of the responsibilities that come with the use of priors which, departing from typical commentaries on their scientific applications, we describe as involving reflection on how best to apply modeling tools to address matters of worldly concern. |
| first_indexed | 2024-12-21T11:57:26Z |
| format | Article |
| id | doaj.art-67bc4764643845c48f6bd0ca615ae093 |
| institution | Directory Open Access Journal |
| issn | 1664-1078 |
| language | English |
| last_indexed | 2024-12-21T11:57:26Z |
| publishDate | 2021-02-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Psychology |
| spelling | doaj.art-67bc4764643845c48f6bd0ca615ae0932022-12-21T19:04:54ZengFrontiers Media S.A.Frontiers in Psychology1664-10782021-02-011210.3389/fpsyg.2021.612251612251From Data to Causes III: Bayesian Priors for General Cross-Lagged Panel Models (GCLM)Michael J. Zyphur0Ellen L. Hamaker1Louis Tay2Manuel Voelkle3Kristopher J. Preacher4Zhen Zhang5Zhen Zhang6Paul D. Allison7Dean C. Pierides8Peter Koval9Edward F. Diener10Edward F. Diener11Department of Management and Marketing, The University of Melbourne, Parkville, VIC, AustraliaDepartment of Methodology and Statistics, Utrecht University, Utrecht, NetherlandsDepartment of Psychological Sciences, Purdue University, West Lafayette, IN, United StatesDepartment of Psychology, Humboldt University of Berlin, Berlin, GermanyDepartment of Psychology and Human Development, Humboldt University of Berlin, Berlin, GermanyCox School of Business, Southern Methodist University, Dallas, TX, United StatesW.P. Carey School of Business, Arizona State University, Tempe, AZ, United StatesDepartment of Sociology, University of Pennsylvania, Philadelphia, PA, United StatesStirling Management School, University of Stirling, Stirling, United Kingdom0Melbourne School of Psychological Sciences, The University of Melbourne, Parkville, VIC, Australia1Department of Psychology, The University of Utah, Salt Lake City, UT, United States2Department of Psychology, University of Virginia, Charlottesville, VA, United StatesThis article describes some potential uses of Bayesian estimation for time-series and panel data models by incorporating information from prior probabilities (i.e., priors) in addition to observed data. Drawing on econometrics and other literatures we illustrate the use of informative “shrinkage” or “small variance” priors (including so-called “Minnesota priors”) while extending prior work on the general cross-lagged panel model (GCLM). Using a panel dataset of national income and subjective well-being (SWB) we describe three key benefits of these priors. First, they shrink parameter estimates toward zero or toward each other for time-varying parameters, which lends additional support for an income → SWB effect that is not supported with maximum likelihood (ML). This is useful because, second, these priors increase model parsimony and the stability of estimates (keeping them within more reasonable bounds) and thus improve out-of-sample predictions and interpretability, which means estimated effect should also be more trustworthy than under ML. Third, these priors allow estimating otherwise under-identified models under ML, allowing higher-order lagged effects and time-varying parameters that are otherwise impossible to estimate using observed data alone. In conclusion we note some of the responsibilities that come with the use of priors which, departing from typical commentaries on their scientific applications, we describe as involving reflection on how best to apply modeling tools to address matters of worldly concern.https://www.frontiersin.org/articles/10.3389/fpsyg.2021.612251/fullpanel data modelGranger causality (VAR)Bayesianshrinkage estimationsmall-variance priors |
| spellingShingle | Michael J. Zyphur Ellen L. Hamaker Louis Tay Manuel Voelkle Kristopher J. Preacher Zhen Zhang Zhen Zhang Paul D. Allison Dean C. Pierides Peter Koval Edward F. Diener Edward F. Diener From Data to Causes III: Bayesian Priors for General Cross-Lagged Panel Models (GCLM) Frontiers in Psychology panel data model Granger causality (VAR) Bayesian shrinkage estimation small-variance priors |
| title | From Data to Causes III: Bayesian Priors for General Cross-Lagged Panel Models (GCLM) |
| title_full | From Data to Causes III: Bayesian Priors for General Cross-Lagged Panel Models (GCLM) |
| title_fullStr | From Data to Causes III: Bayesian Priors for General Cross-Lagged Panel Models (GCLM) |
| title_full_unstemmed | From Data to Causes III: Bayesian Priors for General Cross-Lagged Panel Models (GCLM) |
| title_short | From Data to Causes III: Bayesian Priors for General Cross-Lagged Panel Models (GCLM) |
| title_sort | from data to causes iii bayesian priors for general cross lagged panel models gclm |
| topic | panel data model Granger causality (VAR) Bayesian shrinkage estimation small-variance priors |
| url | https://www.frontiersin.org/articles/10.3389/fpsyg.2021.612251/full |
| work_keys_str_mv | AT michaeljzyphur fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT ellenlhamaker fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT louistay fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT manuelvoelkle fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT kristopherjpreacher fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT zhenzhang fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT zhenzhang fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT pauldallison fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT deancpierides fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT peterkoval fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT edwardfdiener fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm AT edwardfdiener fromdatatocausesiiibayesianpriorsforgeneralcrosslaggedpanelmodelsgclm |