Hess

Stubbing out hypothetical bias: improving tobacco market predictions by combining stated and revealed preference data.

In health, stated preference data from discrete choice experiments (DCEs) are commonly used to estimate discrete choice models that are then used for forecasting behavioral change, often with the goal of informing policy decisions. Data from DCEs are potentially subject to hypothetical bias. In turn, forecasts may be biased, yielding substandard evidence for policymakers. Bias can enter both through the elasticities as well as through the model constants. Simple correction approaches exist (using revealed preference data) but are seemingly not widely used in health economics. We use DCE data from an experiment on smokers in the US. Real-world data are used to calibrate the scale of utility (in two ways) and the alternative-specific constants (ASCs); several innovations for calibration are proposed. We find that embedding revealed preference data in the model makes a substantial difference to the forecasts; and that how models are calibrated also makes a substantial difference.

Buckell, J. & Hess, S. (2019), Stubbing out hypothetical bias: improving tobacco market predictions by combining stated and revealed preference data. Journal of Health Economics, forthcoming.

 A control-function approach to correct for endogeneity in discrete choice models estimated on SP-off-RP data and contrasts with an earlier FIML approach by Train & Wilson.

It is common practice to build Stated Preference (SP) attributes and alternatives from observed Revealed Preference (RP) choices with a view to increasing realism. While many surveys pivot all alternatives around an observed choice, others use more adaptive approaches in which changes are made depending on what alternative was chosen in the RP setting. For example, in SP-off-RP data, the alternative chosen in the RP setting is worsened in the SP setting and other alternatives are improved to induce a change in behaviour. This facilitates the creation of meaningful trade-offs or tipping points but introduces endogeneity. This source of endogeneity was largely ignored until Train and Wilson (T&W) proposed a full information maximum likelihood (FIML) solution that can be implemented with simulation. In this article, we propose a limited information maximum likelihood (LIML) approach to address the SP-off-RP problem using a method which does not need simulation, can be applied with standard software and uses data that is already available for the stated problem. The proposed method is an application of the control-function (CF) method to correct for endogeneity in discrete choice models, using the RP attributes as instrumental variables. We discuss the theoretical and practical advantages and disadvantages of the CF and T&W methods and illustrate them using Monte Carlo and real data. Results show that, while the T&W method may be more efficient in theory, it may however fail to retrieve consistent estimators when it does not account properly for the data generation process if, e.g., an exogenous source of correlation among the SP choice tasks exists. On the other hand, the CF is more robust, i.e. less sensitive, to the data generation process assumptions, and is considerably easier to apply with standard software and does not require simulation, facilitating its adoption and the more extensive use of SP-off-RP data.

Guevara, C.A. & Hess, S. (2019), A control-function approach to correct for endogeneity in discrete choice models estimated on SP-off-RP data and contrasts with an earlier FIML approach by Train & Wilson. Transportation Research Part B, forthcoming.

Understanding attitudes towards congestion pricing: a latent variable investigation with data from four cities.

Numerous cities around the world are considering the implementation of road pricing to ease urban traffic congestion, following on from the success in cities such as London and Singapore. However, policy-makers are also all too aware of the generally negative public opinion toward such measures. This study makes use of data collected in four cities (two in Sweden, one in Finland, and one in France) using a very consistent survey probing for citizens’ attitudes toward pricing. We find very strong similarities across the four cities in terms of a number of underlying attitudinal constructs that help explain people’s answers in a hypothetical referendum on congestion pricing. The similarities across cities indicate that the increase in the opinion toward congestion pricing once they are introduced is not primarily an effect of changes in underlying attitudes, changes in how the underlying attitudes influence the support for congestion pricing, or differences in anticipated versus experienced or perceived self-interest. Instead, this effect seems to be caused by a status quo acceptance, tending to increase the support for the current situation.

Hess, S. & Börjesson, M. (2019), Understanding attitudes towards congestion pricing: a latent variable investigation with data from four cities. Transportation Letters, 11(2), pp. 63-77.

Modelling pedestrian crossing choice behaviour on Cape Town freeways: caught between a rock and a hard place?

Pedestrians and freeways are not supposed to coexist in any proximity to each other, yet in Cape Town, South Africa, the Freeway Management System (FMS) has recorded an alarming increase in pedestrian activity on its freeways in recent years, with a similar trend in (fatal) freeway pedestrian crashes. This paper reports on the development of a series of discrete choice models for pedestrian crossing decisions in a highly volatile and vulnerable freeway crossing environment. We hypothesize that freeway pedestrian crossing is driven by personal factors and the perceived risks associated with completing the crossing using a footbridge or (illegally) at-grade. We test this assumption by making use of stated choice and risk perception data collected from (adult) participants intercepted along three Cape Town freeways. Estimating mixed logit and hybrid choice models we look at the role of random heterogeneity and latent risk perception in the choice to cross at-grade or using a footbridge. The model estimates confirm that, as expected, crossing choice is largely influenced by a combination of built environment, vehicular and pedestrian traffic, next to some socio-demographic factors, but also risk perception. The study brings to light the seemingly opposite effect of some of the factors on risk perception and crossing choice. We also show that risk perception really only influences crossing choice in terms of the perception of at-grade crossing risk, where the impact is however non-trivial. Finally, we look at the implied relative sensitivities of the choice attributes within and between the crossing alternatives, as well between the three estimated models, amongst others demonstrating the power of the hybrid choice model over the other two. The results of the study can inform opportunities to counter the upward trend of fatalities and provide suggestions for policy-making that would lead to improved freeway crossing safety.

Dada, M., Zuidgeest, M. & Hess, S. (2019), Modelling pedestrian crossing choice behaviour on Cape Town freeways: caught between a rock and a hard place? Transportation Research Part F, 60, pp. 245-261.

Reliability in the German value of time study.

The German Federal Ministry of Transport and Digital Infrastructure is currently preparing the 2015 Federal Transport Investment Plan. Because this effort includes an update to the overall methodology of the ministry's cost–benefit analysis, both the value of reliability (VOR) and the value of travel time (VOT) for personal and business travel will be estimated. The VOT will replace a set of existing values, but the VOR will be estimated for the first time because they are not yet incorporated in the standard appraisal. A two-stage approach was used for data collection: first respondents reported about current trips (revealed preference), and then these responses were systematically varied to become the basis for stated preference experiments. This paper presents the findings of estimating the VOR. In the stated preference experiments the reliability of the travel modes was presented with different formats. The final model formulation differed in the definition of reliability for private and public transport. For car trips, saving travel time had more value for the respondents than reducing the variability. The calculated VOR for the mean expected unscheduled delay of public transport trips was slightly lower than the VOT; this result indicates that the reliability was less important to the respondents than the relevant travel time saving. A mean expected unscheduled delay of 1 min and 1 min of standard deviation are almost equivalent to 1 min of travel time saving (reliability ratio). Because this was the first official estimation of VOR and VOT for Germany, the values should be reconsidered and updated on a regular basis.

Ehreke, I., Hess, S., Weis, C. & Axhausen, K.W. (2015), Reliability in the German value of time study. Transportation Research Record, 2495 (1), pp 14-22.

Individual level models vs. sample level models: contrasts and mutual benefits

With a view to better capturing heterogeneity across decision makers and improving prediction of choices, there is increasing interest in estimating separate models for each person. Almost exclusively, this work has however taken place outside the field of transport research. The aim of the present paper is twofold. We first wish to give an account of the potential benefits of a greater focus on individual level estimates in transport applications. Secondly, we wish to offer further insights into the relative benefits of sample level and individual level models (ILMs) by drawing on a data set containing an unusually large number (144) of decisions on holiday travel per individual. In addition to comparing existing approaches, we also put forward the use of a novel technique which draws on the relative benefits of both sample level and ILMs by estimating ILMs in a Bayesian fashion with priors drawn from a sample level model. Our results show only limited differences between ILMs and conditionals from sample level models when working with the full set of choices. When working with more realistic sample sizes at the person level, our results suggest that ILMs can offer better performance on the estimation data but that this is a result of overfitting which can lead to inferior prediction performance. Our proposed Bayesian ILM model offers good intermediary performance. The use of best-worst data rather than simple stated choice, as is done commonly in published ILM work, does not lead to major changes to these findings.

Dumont, J., Giergiczny, M. & Hess, S. (2015), Individual level models vs. sample level models: contrasts and mutual benefits. Transportmetrica, 11 (6), pp 465-483.

Intra-respondent heterogeneity in a stated choice survey on wetland conservation in Belarus: first steps towards creating a link with uncertainty in contingent valuation.

Applications of discrete choice models in environmental valuation increasingly use a random coefficient specification, such as mixed logit, to represent taste heterogeneity. The majority of applications rely on data containing multiple observations for each respondent, where a common assumption is that tastes stay constant across choices for the same respondent. We question this assumption and make use of a model developed in the transport field which allows tastes to vary over choices for each consumer in addition to variation across consumers. An empirical analysis making use of a stated choice dataset for wetland conservation in Belarus shows that superior performance is obtained by allowing jointly for the two types of heterogeneity and that recovery of these intra-respondent variations is not possible using standard approaches, such as allowing for scale heterogeneity across tasks. We show also that intra-respondent heterogeneity can be especially high for attributes which respondents are unfamiliar with, and that a failure to account for it can substantially affect welfare estimates. We interpret this as an indication that this heterogeneity relates primarily to uncertainty. Finally, we offer initial insights into the relationship between intra-respondent heterogeneity and findings on uncertainty in a contingent valuation context.

Hess, S. & Giergiczny, M. (2015), Intra-respondent heterogeneity in a stated choice survey on wetland conservation in Belarus: first steps towards creating a link with uncertainty in contingent valuation. Environmental & Resource Economics, 60 (3), pp 327-347.

Incorporating environmental attitudes in discrete choice models: an exploration of the utility of the awareness of consequences scale.

Environmental economists are increasingly interested in better understanding how people cognitively organise their beliefs and attitudes towards environmental change in order to identify key motives and barriers that stimulate or prevent action. In this paper, we explore the utility of a commonly used psychometric scale, the awareness of consequences (AC) scale, in order to better understand stated choices. The main contribution of the paper is that it provides a novel approach to incorporate attitudinal information into discrete choice models for environmental valuation: firstly, environmental attitudes are incorporated using a reinterpretation of the classical AC scale recently proposed by Ryan and Spash (2012); and, secondly, attitudinal data is incorporated as latent variables under a hybrid choice modelling framework. This novel approach is applied to data from a survey conducted in the Basque Country (Spain) in 2008 aimed at valuing land-use policies in a Natura 2000 Network site. The results are relevant to policy-making because choice models that are able to accommodate underlying environmental attitudes may help in designing more effective environmental policies.

Hoyos, D., Mariel, P. & Hess, S. (2015), Incorporating environmental attitudes in discrete choice models: an exploration of the utility of the awareness of consequences scale. Science of the Total Environment, 505, pp 1100-1111.

Heterogeneous preferences toward landscape externalities of wind turbines – combining choices and attitudes in a hybrid model.

Expanding the share of renewable energy sources might substantially increase externalities as, for example, wind turbines may disturb the landscape and negatively affect biodiversity. This paper investigates the public׳s sensitivities towards these externalities by using discrete choice experiments and shows how preferences differ across inhabitants of our study region. As a further insight into the sources for these variations, a hybrid choice model is employed in order to incorporate individuals׳ latent attitudes in the estimated model. Our latent class structure allocates individuals to classes according to underlying latent attitudes that also influence the answers to attitudinal questions. We show that these underlying attitudes are a function of a number of socio-demographic characteristics, with young people, men with low income and those living closer to turbines having a stronger pro-wind power generation attitude. The inclusion of the attitudes in the class allocation component of the latent class model leads to a richer picture of people׳s valuations, revealing, for example, antagonistic preferences of two distinct groups of respondents, i.e. advocates and opponents of wind power generation.

Mariel, P., Meyerhoff, J. & Hess, S.  (2015), Heterogeneous preferences toward landscape externalities of wind turbines – combining choices and attitudes in a hybrid model. Renewable & Sustainable Energy Reviews, 41, pp 647-657.

Understanding the formation and influence of attitudes in patients’ treatment choices for lower back pain: testing the benefits of a hybrid choice model approach.

A growing number of studies across different fields are making use of a new class of choice models, labelled variably as hybrid model structures or integrated choice and latent variable models, and incorporating the role of attitudes in decision making. To date, this technique has not been used in health economics. The present paper looks at the formation of such attitudes and their role in patients' treatment choices in the context of low back pain. We use stated choice data collected from a sample of 561 patients with 348 respondents referred to a regional spine centre in Middelfart, Denmark in spring/summer 2012. We show how the hybrid model structure is able to make a link between attitudinal questions and treatment choices, and also explains variation of these attitudes across key socio-demographic groups. However, we also show how, in this case, only a small share of the overall heterogeneity is linked to the latent attitude construct. Despite their growing popularity, the key findings of the advanced model, despite a greater insight into the drivers of attitudes and small gains in efficiency, are no different from standard approaches which remain easier to apply.

Kløjgaard, M. & Hess, S.  (2014), Understanding the formation and influence of attitudes in patients’ treatment choices for lower back pain: testing the benefits of a hybrid choice model approach. Social Science & Medicine, 114, pp 138-150.

Heterogeneity assumptions in the specification of bargaining models: a study of household level trade-offs between commuting time and salary.

With many real world decisions being made in conjunction with other decision makers, or single agent decisions having an influence on other members of the decision maker’s immediate entourage, there is strong interest in studying the relative weight assigned to different agents in such contexts. In the present paper, we focus on the case of one member of a two person household being asked to make choices affecting the travel time and salary of both members. We highlight the presence of significant heterogeneity across individuals not just in their underlying sensitivities, but also in the relative weight they assign to their partner, and show how this weight varies across attributes. This is in contrast to existing work which uses weights assigned to individual agents at the level of the overall utility rather than for individual attributes. We also show clear evidence of a risk of confounding between heterogeneity in marginal sensitivities and heterogeneity in the weights assigned to each member. We show how this can lead to misleading model results, and argue that this may also explain past results showing bargaining or weight parameters outside the usual [0,1] range in more traditional joint decision making contexts. In terms of substantive results, we find that male respondents place more weight on their partner’s travel time, while female respondents place more weight on their partner’s salary.

O’Neill, V.L. & Hess, S. (2014), Heterogeneity assumptions in the specification of bargaining models: a study of household level trade-offs between commuting time and salary. Transportation, 41 (4), pp 745-763.

Impact of unimportant attributes in stated choice surveys.

Despite growing interest in the notion that respondents in stated choice surveys may make their decisions on the basis of only a subset of the presented attributes, the impact of any unimportant attributes on the estimates of other valuations is somewhat unclear. This paper presents evidence from a two stage survey where the second stage eliminates attributes deemed unimportant in the first stage. Our analysis shows no evidence of systematic differences between the results of the two stages. This leads to the conclusion that, up to a point where respondent burden may become an issue, analysts should include all attributes that may be relevant, and allow the respondent to filter out those that play no role.

Hess, S. (2014), Impact of unimportant attributes in stated choice surveys. European Journal of Transport and Infrastructure Research, 14 (4), pp 349-361.

A question of taste: recognising the role of latent preferences and attitudes in analysing food choices.

There has long been substantial interest in understanding consumer food choices, where a key complexity in this context is the potentially large amount of heterogeneity in tastes across individual consumers, as well as the role of underlying attitudes towards food and cooking. The present paper underlines that both tastes and attitudes are unobserved, and makes the case for a latent variable treatment of these components. Using empirical data collected in Northern Ireland as part of a wider study to elicit intra-household trade-offs between home-cooked meal options, we show how these latent sensitivities and attitudes drive both the choice behaviour as well as the answers to supplementary questions. We find significant heterogeneity across respondents in these underlying factors and show how incorporating them in our models leads to important insights into preferences.

O’Neill, V.L., Hess, S. & Campbell, D. (2014), A question of taste: recognising the role of latent preferences and attitudes in analysing food choices. Food Quality and Preferences, 32C, pp 299-310.

Understanding air travellers’ trade-offs between connecting flights and surface access characteristics.

This paper reports on a study which seeks to improve our understanding of how people choose between different kinds of flight at competing airports, and how their choices are affected by access conditions. In particular, using stated choice data collected in Scotland, it investigates whether improving surface access to regional airports that are in relatively close proximity to one another (Glasgow and Edinburgh) leads people to avoid taking indirect flights from their nearest airport in favour of direct flights from an alternative airport. In line with expectations, our estimation results from Cross-Nested Logit models show strong aversion to connecting flights, resulting in a willingness to either pay higher fares for direct flights or accept non-trivial increases in access time. For the latter, even without the potential new direct rail link between the two airports, current access times are such that a scenario where direct flights were only available at the non-home airport, a substantial share of passengers would choose to travel from the alternative airport.

Johnson, D., Hess, S. & Matthews, B. (2014), Understanding air travellers’ trade-offs between connecting flights and surface access characteristics. Journal of Air Transport Management, 34, pp 70-77.