A discrete choice model incorporating thresholds for perception in attribute values

Cantillo, Víctor; Heydecker, Benjamin; Ortúzar, Juan de Dios

doi:10.1016/j.trb.2005.11.002

Cited by 82 publications

(52 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Transportation (2018) 45:1-21 15 The results show that the VTTC increases consistently with the level of Dt, in line with existing literature (e.g. Mackie et al 2003;Cantillo et al 2006). However, most of the differences are not statistically significant, and only the VTTC when Dt = 3 is significantly lower than the rest at the 90 % confidence level.…”

Section: Exercise 2-data Split By Dtsupporting

confidence: 87%

Understanding valuation of travel time changes: are preferences different under different stated choice design settings?

2016

View full text Add to dashboard Cite

Stated choice (SC) experiments are the most popular method to estimate the value of travel time changes (VTTC) of a population. In the simplest VTTC experiment, the SC design variables are time changes and cost changes. The levels of these variables create a particular setting from which preferences are inferred. This paper tries to answer the question ''do preferences vary with SC settings?''. For this, we investigate the role of the variables used in the SC experiment on the estimation of the set of VTTC (i.e. mean and covariates). Ideally, one would like to observe the same individuals completing different SC experiments. Since that option is not available, an alternative approach is to use a large dataset of responses, and split it according to different levels of the variable of interest. We refer to this as partial data analysis. The estimation of the same model on each sub-sample provides insights into potential effects of the variable of interest. This approach is applied in relation to three design variables on the data for the last national VTTC study in the UK, using state-of-the-art model specifications. The results show several ways in which the estimated set of VTTC can be affected by the levels of SC design variables. We conclude that model estimates (including the VTTC and covariates) are different in different settings. Hence by focussing the survey on specific settings, sample level results will be affected accordingly. Our findings have implications for appraisal and can inform the construction of future SC experiments.

show abstract

Section: Exercise 2-data Split By Dtsupporting

confidence: 87%

Understanding valuation of travel time changes: are preferences different under different stated choice design settings?

2016

View full text Add to dashboard Cite

show abstract

“…A substantial number of studies have found that the VTT increases with the size of the travel time savings in stated choice experiments (see for instance Bates and Whelan (2001), Hultkrantz and Mortazavi (2001) and Cantillo et al (2006)). To further explore the properties of the value of time distribution and the role of ∆t, a local logit regression is also performed in (log(∆t),log(V))-space.…”

Section: Local Logitmentioning

confidence: 99%

Catching the tail: Empirical identification of the distribution of the value of travel time

Börjesson¹,

Fosgerau²,

Algers³

2012

Transportation Research Part A: Policy and Practice

View full text Add to dashboard Cite

Recent methodological advances in discrete choice analysis in combination with certain stated choice experiments have allowed researchers to check empirically the identification of the distribution of latent variables such as the value of travel time (VTT). Lack of identification is likely to be common and the consequences are severe. E.g., the Danish value of time study found the 15% right tail of the VTT distribution to be unidentified, making it impossible to estimate the mean VTT without resorting to strong assumptions with equally strong impact on the resulting estimate. This paper analyses data generated from a similar choice experiment undertaken in Sweden during [2007][2008] in which the range of trade-off values between time and money was significantly increased relative to the Danish experiment. The results show that this change allowed empirical identification of effectively the entire VTT distribution. In addition to informing the design of future choice experiments, the results are also of interest as a validity test of the stated choice methodology. Failure in identifying the right tail of the VTT would have made it difficult to maintain that respondents' behaviour is consistent with utility maximization in the sense intended by the experimenter.

show abstract

“…Despite a growing number of studies focusing on these issues (see for example Cantillo et al 2006, Hensher 2006b, Swait 2001, in the majority of empirical studies, the entire domain of every attribute is treated as relevant to some degree and included in the utility expressions for every individual. In response to a concern over the assumption of relevancy, a series of papers by Hensher and colleagues in particular have focussed on the role that a range of attribute processing rules (or heuristics) might play in influencing the empirical identification of the preferences of individuals and populations.…”

Section: Introductionmentioning

confidence: 99%

Using conditioning on observed choices to retrieve individual-specific attribute processing strategies

Hess

Hensher

2010

Transportation Research Part B: Methodological

166

171

View full text Add to dashboard Cite

With the growing reliance on Stated Choice (SC) data, researchers are increasingly interested in understanding how respondents process the information presented to them in such surveys. Specifically, it has been argued that some respondents may simplify the choice tasks by consistently ignoring one or more of the attributes describing the alternatives, and direct questions put to respondents after the completion of SC surveys support this hypothesis. However, in the general context of issues with response quality in SC data, there are certainly grounds for questioning the reliability of stated attribute processing strategies. In this paper, we take a different approach by attempting to infer attribute processing strategies through the analysis of respondent-specific coefficient distributions obtained through conditioning on observed choices. Our results suggest that a share of respondents do indeed ignore a subset of explanatory variables. However, there is also some evidence that the inferred attribute processing strategies are not necessarily consistent with the stated attribute processing strategies. Additionally, there is some evidence that respondents who claim to have ignored a certain attribute may simply have assigned it lesser importance. The results produced by the inferring approach not only lead to slightly better fit but also more consistent results.

show abstract

A discrete choice model incorporating thresholds for perception in attribute values

Cited by 82 publications

References 18 publications

Understanding valuation of travel time changes: are preferences different under different stated choice design settings?

Understanding valuation of travel time changes: are preferences different under different stated choice design settings?

Catching the tail: Empirical identification of the distribution of the value of travel time

Using conditioning on observed choices to retrieve individual-specific attribute processing strategies

Contact Info

Product

Resources

About