01604573

Component

All researchers expect model parameter estimates should be unbiased, regardless of the stated preference designs used in the discrete choice model. In fact, the different designs could yield somewhat different parameter values. The question raised is how the stated preference designs could bias the parameter estimates, and the answer is not clear. Although some researchers inferred the reason is that the choice scenarios contain dominant alternatives, but this conjecture does not be verified. In this paper, the authors will analyze how the dominant alternatives in design could bias the parameter estimates. Frist the utility balance is used to quantitatively describe the dominant alternatives in the different designs. Meanwhile, the "nested logit trick" model is used to test whether the bias is induced by the changes of error variances resulting from dominant alternatives. Finally, the authors found that the main reason for the bias is that some choice situations consist of dominant alternatives in design. If there are more choice situations consisting of dominant alternatives, the bias of the parameter estimates is more significant. The utility balance is very important for reducing the parameter estimate bias.

01550057

15-1410

English

13p

2015

Transportation Research Board 94th Annual Meeting

Digital/other

References; Tables

Case studies; Choice models; Data collection; Public transit; Stated preferences; Surveys

Planning and Forecasting; Public Transportation

Transportation Research Board Annual Meeting 2015 Paper #15-1410

TRIS, TRB, ATRI

Jul 5 2016 3:31PM