A Random Parameter Model with Onsite Sampling for Recreation Site Choice

An Application to Southern California Shoreline Sportfishing

Koichi Kuriyama, James Hilger, William Hanemann

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Estimation of consistent parameter estimates for recreational demand models faces challenges arising from the choice-based nature of the data collected primarily for resource management purposes. As an alternative to randomized respondent-based sampling, choice-based onsite sampling can provide information on actual choices made by a subset of the population where participation has a low incidence. While the literature has shown that under specific restrictions the estimation of choice models from onsite sampling data yields unbiased fixed parameter estimates for the conditional logit model, this result does not carry over to estimation of the random parameter logit model. We propose an estimator for the unbiased estimation of the random parameter model using choice-based data; our estimator uses weights based on information about the level of sampling effort. An empirical application of the standard and weighted discrete choice RUM models to onsite sample data on recreational fishing illustrates the advantages of the proposed estimator. The estimation results indicate the compensating variation associated with an decrease, or increase, of 50 % in expected catch rates for a recreational shoreline sportfishing trip to a man-made structure in southern California is -$2.80 or $3.54 per trip, respectively.

Original languageEnglish (US)
Pages (from-to)481-497
Number of pages17
JournalEnvironmental and Resource Economics
Volume56
Issue number4
DOIs
StatePublished - 2013

Fingerprint

shoreline
Sampling
sampling
recreation
parameter
Recreation
Random parameters
resource management
fishing
Estimator
Choice-based sampling

Keywords

  • Onsite sampling
  • Random parameter logit
  • Random utility models
  • Recreation demand
  • Recreational fishing

ASJC Scopus subject areas

  • Management, Monitoring, Policy and Law
  • Aerospace Engineering

Cite this

@article{33e58863d4fb44f486e666c9e8ce5485,
title = "A Random Parameter Model with Onsite Sampling for Recreation Site Choice: An Application to Southern California Shoreline Sportfishing",
abstract = "Estimation of consistent parameter estimates for recreational demand models faces challenges arising from the choice-based nature of the data collected primarily for resource management purposes. As an alternative to randomized respondent-based sampling, choice-based onsite sampling can provide information on actual choices made by a subset of the population where participation has a low incidence. While the literature has shown that under specific restrictions the estimation of choice models from onsite sampling data yields unbiased fixed parameter estimates for the conditional logit model, this result does not carry over to estimation of the random parameter logit model. We propose an estimator for the unbiased estimation of the random parameter model using choice-based data; our estimator uses weights based on information about the level of sampling effort. An empirical application of the standard and weighted discrete choice RUM models to onsite sample data on recreational fishing illustrates the advantages of the proposed estimator. The estimation results indicate the compensating variation associated with an decrease, or increase, of 50 {\%} in expected catch rates for a recreational shoreline sportfishing trip to a man-made structure in southern California is -$2.80 or $3.54 per trip, respectively.",
keywords = "Onsite sampling, Random parameter logit, Random utility models, Recreation demand, Recreational fishing",
author = "Koichi Kuriyama and James Hilger and William Hanemann",
year = "2013",
doi = "10.1007/s10640-013-9640-4",
language = "English (US)",
volume = "56",
pages = "481--497",
journal = "Environmental and Resource Economics",
issn = "0924-6460",
publisher = "Springer Netherlands",
number = "4",

}

TY - JOUR

T1 - A Random Parameter Model with Onsite Sampling for Recreation Site Choice

T2 - An Application to Southern California Shoreline Sportfishing

AU - Kuriyama, Koichi

AU - Hilger, James

AU - Hanemann, William

PY - 2013

Y1 - 2013

N2 - Estimation of consistent parameter estimates for recreational demand models faces challenges arising from the choice-based nature of the data collected primarily for resource management purposes. As an alternative to randomized respondent-based sampling, choice-based onsite sampling can provide information on actual choices made by a subset of the population where participation has a low incidence. While the literature has shown that under specific restrictions the estimation of choice models from onsite sampling data yields unbiased fixed parameter estimates for the conditional logit model, this result does not carry over to estimation of the random parameter logit model. We propose an estimator for the unbiased estimation of the random parameter model using choice-based data; our estimator uses weights based on information about the level of sampling effort. An empirical application of the standard and weighted discrete choice RUM models to onsite sample data on recreational fishing illustrates the advantages of the proposed estimator. The estimation results indicate the compensating variation associated with an decrease, or increase, of 50 % in expected catch rates for a recreational shoreline sportfishing trip to a man-made structure in southern California is -$2.80 or $3.54 per trip, respectively.

AB - Estimation of consistent parameter estimates for recreational demand models faces challenges arising from the choice-based nature of the data collected primarily for resource management purposes. As an alternative to randomized respondent-based sampling, choice-based onsite sampling can provide information on actual choices made by a subset of the population where participation has a low incidence. While the literature has shown that under specific restrictions the estimation of choice models from onsite sampling data yields unbiased fixed parameter estimates for the conditional logit model, this result does not carry over to estimation of the random parameter logit model. We propose an estimator for the unbiased estimation of the random parameter model using choice-based data; our estimator uses weights based on information about the level of sampling effort. An empirical application of the standard and weighted discrete choice RUM models to onsite sample data on recreational fishing illustrates the advantages of the proposed estimator. The estimation results indicate the compensating variation associated with an decrease, or increase, of 50 % in expected catch rates for a recreational shoreline sportfishing trip to a man-made structure in southern California is -$2.80 or $3.54 per trip, respectively.

KW - Onsite sampling

KW - Random parameter logit

KW - Random utility models

KW - Recreation demand

KW - Recreational fishing

UR - http://www.scopus.com/inward/record.url?scp=84888822682&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84888822682&partnerID=8YFLogxK

U2 - 10.1007/s10640-013-9640-4

DO - 10.1007/s10640-013-9640-4

M3 - Article

VL - 56

SP - 481

EP - 497

JO - Environmental and Resource Economics

JF - Environmental and Resource Economics

SN - 0924-6460

IS - 4

ER -