Effect of different distributional assumption of random parameters in the mixed logit model on willingness-to-pay values
Understanding consumers’ willingness-to-pay (WTP) value is vital for rational valuation of consumers’ benefit. Stated preference techniques such as Choice Experiment (CE) have been widely used to account for WTP value and one of the popular model specifications is the Mixed Logit (MXL) model. In the...
Saved in:
Main Authors: | , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Faculty of Economics and Management, Universiti Putra Malaysia
2019
|
Online Access: | http://psasir.upm.edu.my/id/eprint/76501/1/12%29%20Effect%20of%20Different%20Distributional%20Assumption.pdf http://psasir.upm.edu.my/id/eprint/76501/ http://www.ijem.upm.edu.my/vol13no2/12)%20Effect%20of%20Different%20Distributional%20Assumption.pdf |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | Understanding consumers’ willingness-to-pay (WTP) value is vital for rational valuation of consumers’ benefit. Stated preference techniques such as Choice Experiment (CE) have been widely used to account for WTP value and one of the popular model specifications is the Mixed Logit (MXL) model. In the MXL model, it is essential to assume the types of distribution of random parameters. The specification of MXL models with different distributional assumptions of random parameters has been explored by many researchers. Nevertheless, the effect of different distributional assumptions of random parameters on goodness-of-fit, the significance of the coefficients and WTP values has not been studied adequately particularly in the context of Malaysia. In the present work, the analysis is carried out in this regard based on visitors’ preferences for tourist facilities at Tasik Kenyir, Malaysia. A number of MXL models were attempted with different distributional assumptions of random parameters; normal, log-normal, triangular and uniform. The results suggested that, in all MXL models, the goodness-of-fit statistics, the significance of the attribute coefficients and WTP values were quite comparable, except for the log-normal distribution. The methodological implications concern the importance of developing several MXL models with different distributional assumptions as well as the recommendation for policy makers to improve the facilities at the lake. |
---|