Bayesian imputation of non-chosen attribute values in revealed preference surveys

Simon Washington, Srinath Ravulaparthy, John M. Rose, David Hensher, Ram Pendyala

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

Obtaining attribute values of non-chosen alternatives in a revealed preference context is challenging because non-chosen alternative attributes are unobserved by choosers, chooser perceptions of attribute values may not reflect reality, existing methods for imputing these values suffer from shortcomings, and obtaining non-chosen attribute values is resource intensive. This paper presents a unique Bayesian (multiple) Imputation Multinomial Logit model that imputes unobserved travel times and distances of non-chosen travel modes based on random draws from the conditional posterior distribution of missing values. The calibrated Bayesian (multiple) Imputation Multinomial Logit model imputes non-chosen time and distance values that convincingly replicate observed choice behavior. Although network skims were used for calibration, more realistic data such as supplemental geographically referenced surveys or stated preference data may be preferred. The model is ideally suited for imputing variation in intrazonal non-chosen mode attributes and for assessing the marginal impacts of travel policies, programs, or prices within traffic analysis zones.

Original languageEnglish (US)
Pages (from-to)48-65
Number of pages18
JournalJournal of Advanced Transportation
Volume48
Issue number1
DOIs
StatePublished - 2014

Fingerprint

Travel time
Calibration
Imputation
Revealed preference
Multiple imputation
Multinomial logit model
Resources
Choice behavior
Posterior distribution
Missing values
Stated preference

Keywords

  • Bayesian methods
  • choice models
  • imputation
  • missing data analysis
  • multinomial logit
  • synthesized data
  • unobserved choice attributes

ASJC Scopus subject areas

  • Strategy and Management
  • Economics and Econometrics
  • Mechanical Engineering
  • Computer Science Applications
  • Automotive Engineering

Cite this

Bayesian imputation of non-chosen attribute values in revealed preference surveys. / Washington, Simon; Ravulaparthy, Srinath; Rose, John M.; Hensher, David; Pendyala, Ram.

In: Journal of Advanced Transportation, Vol. 48, No. 1, 2014, p. 48-65.

Research output: Contribution to journalArticle

Washington, Simon ; Ravulaparthy, Srinath ; Rose, John M. ; Hensher, David ; Pendyala, Ram. / Bayesian imputation of non-chosen attribute values in revealed preference surveys. In: Journal of Advanced Transportation. 2014 ; Vol. 48, No. 1. pp. 48-65.
@article{73aca01c2ea14c468483b03c33af2a66,
title = "Bayesian imputation of non-chosen attribute values in revealed preference surveys",
abstract = "Obtaining attribute values of non-chosen alternatives in a revealed preference context is challenging because non-chosen alternative attributes are unobserved by choosers, chooser perceptions of attribute values may not reflect reality, existing methods for imputing these values suffer from shortcomings, and obtaining non-chosen attribute values is resource intensive. This paper presents a unique Bayesian (multiple) Imputation Multinomial Logit model that imputes unobserved travel times and distances of non-chosen travel modes based on random draws from the conditional posterior distribution of missing values. The calibrated Bayesian (multiple) Imputation Multinomial Logit model imputes non-chosen time and distance values that convincingly replicate observed choice behavior. Although network skims were used for calibration, more realistic data such as supplemental geographically referenced surveys or stated preference data may be preferred. The model is ideally suited for imputing variation in intrazonal non-chosen mode attributes and for assessing the marginal impacts of travel policies, programs, or prices within traffic analysis zones.",
keywords = "Bayesian methods, choice models, imputation, missing data analysis, multinomial logit, synthesized data, unobserved choice attributes",
author = "Simon Washington and Srinath Ravulaparthy and Rose, {John M.} and David Hensher and Ram Pendyala",
year = "2014",
doi = "10.1002/atr.201",
language = "English (US)",
volume = "48",
pages = "48--65",
journal = "Journal of Advanced Transportation",
issn = "0197-6729",
publisher = "John Wiley and Sons Ltd",
number = "1",

}

TY - JOUR

T1 - Bayesian imputation of non-chosen attribute values in revealed preference surveys

AU - Washington, Simon

AU - Ravulaparthy, Srinath

AU - Rose, John M.

AU - Hensher, David

AU - Pendyala, Ram

PY - 2014

Y1 - 2014

N2 - Obtaining attribute values of non-chosen alternatives in a revealed preference context is challenging because non-chosen alternative attributes are unobserved by choosers, chooser perceptions of attribute values may not reflect reality, existing methods for imputing these values suffer from shortcomings, and obtaining non-chosen attribute values is resource intensive. This paper presents a unique Bayesian (multiple) Imputation Multinomial Logit model that imputes unobserved travel times and distances of non-chosen travel modes based on random draws from the conditional posterior distribution of missing values. The calibrated Bayesian (multiple) Imputation Multinomial Logit model imputes non-chosen time and distance values that convincingly replicate observed choice behavior. Although network skims were used for calibration, more realistic data such as supplemental geographically referenced surveys or stated preference data may be preferred. The model is ideally suited for imputing variation in intrazonal non-chosen mode attributes and for assessing the marginal impacts of travel policies, programs, or prices within traffic analysis zones.

AB - Obtaining attribute values of non-chosen alternatives in a revealed preference context is challenging because non-chosen alternative attributes are unobserved by choosers, chooser perceptions of attribute values may not reflect reality, existing methods for imputing these values suffer from shortcomings, and obtaining non-chosen attribute values is resource intensive. This paper presents a unique Bayesian (multiple) Imputation Multinomial Logit model that imputes unobserved travel times and distances of non-chosen travel modes based on random draws from the conditional posterior distribution of missing values. The calibrated Bayesian (multiple) Imputation Multinomial Logit model imputes non-chosen time and distance values that convincingly replicate observed choice behavior. Although network skims were used for calibration, more realistic data such as supplemental geographically referenced surveys or stated preference data may be preferred. The model is ideally suited for imputing variation in intrazonal non-chosen mode attributes and for assessing the marginal impacts of travel policies, programs, or prices within traffic analysis zones.

KW - Bayesian methods

KW - choice models

KW - imputation

KW - missing data analysis

KW - multinomial logit

KW - synthesized data

KW - unobserved choice attributes

UR - http://www.scopus.com/inward/record.url?scp=84891866318&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84891866318&partnerID=8YFLogxK

U2 - 10.1002/atr.201

DO - 10.1002/atr.201

M3 - Article

VL - 48

SP - 48

EP - 65

JO - Journal of Advanced Transportation

JF - Journal of Advanced Transportation

SN - 0197-6729

IS - 1

ER -