Use Of Discrete Choice Models Research Articles

Welfare with Imperfect Foreknowledge: The Case of Risk Aversion

We study the use of discrete choice models in the context of recreation when there may be discrepancy between the anticipated (ex ante) and the experienced quality of site attributes due to imperfect information. Unlike previous literature on the topic, we relax the assumption of risk neutrality. Adopting a model with quasi-linear preferences, we characterize the disparity between ex ante and ex post utility by decomposing expected utility into a term stemming from bias in beliefs and a component representing willingness to pay for reducing uncertainty in site attributes. Our results show that ex post corrections of welfare losses are attenuated by greater dispersion in beliefs about amenities.

Tourists’ travel time values using discrete choice models: the recreational value of the Teide National Park

The value of travel time is one of the most important factors in recreation demand models. Traditionally, the most common approach for its calculation has been the use of different proportions of the wage rate; however, criticisms of this method abound because in a recreational trip the relevant measure is the opportunity cost of leisure time rather than work time. In this paper, we adopt a novel approach in the literature using discrete choice models based on short-term decisions and independent of the labor market. We obtain the value of travel time through the trade-off between time and money considered by the tourist visitors when choosing the transport mode and we present the first calculation of the recreational value of the Teide National Park. Specifically, using a Revealed Preference survey of 801 park visitors, we estimate Mixed Logit models accounting for random preference heterogeneity, derive travel time values and incorporate them into a Zonal Travel Cost Method. This approach allows us to estimate different time values depending on transport mode and stage of the trip and shows that the use of discrete choice models instead of the wage rate approach has a strong impact on the recreational value calculated.

Air itinerary shares estimation using multinomial logit models

ABSTRACTThe main goal of this study is the development of an aggregate air itinerary market share model. In order to achieve this, multinomial logit models are applied to distribute the city-pair passenger demand across the available itineraries. The models are developed at an aggregate level using open-source booking data for a large group of city-pairs within the US air transport system. Although there is a growing trend in the use of discrete choice models in the aviation industry, existing air itinerary share models are mostly focused on supporting carrier decision-making. Consequently, those studies define itineraries at a more disaggregate level using variables describing airlines and time preferences. In this study, we define itineraries at a more aggregate level, i.e. as a combination of flight segments between an origin and destination, without further insight into service preferences. Although results show some potential for this approach, there are challenges associated with prediction performance and computational intensity.

A DISCRETE CHOICE MODEL FOR FIRM LOCATION DECISION

Where to locate? It is one of the most important question in locating a business in a city. In the city center, business or firms are functioning as a dominant attractor of employment and also employment locations which linked the land use and transportation system. The objective of this paper is to describe the location model of firms in Kuala Lumpur area. Two important determinants of location choice model in this study are the accessibility measures and the suitability analysis indicators. The model focuses on the statistical technique for analyzing discrete choice data by using econometric and Geographic Information System software. The findings in this paper show that agriculture, mining, electricity, gas and water, transport and finance firms' type are mostly located outside of Kuala Lumpur's Central Business District area. Meanwhile, manufacturing, construction and wholesale firms' type are located in the Central Business District area. The result of this study will highlight the use of discrete choice models in the analysis of firm location decisions which will be a foundation to facilitate town planners and decision makers to understand the firm location decisions in their region.

A DISCRETE CHOICE MODEL FOR FIRM LOCATION DECISION

Where to locate? It is one of the most important question in locating a business in a city. In the city center, business or firms are functioning as a dominant attractor of employment and also employment locations which linked the land use and transportation system. The objective of this paper is to describe the location model of firms in Kuala Lumpur area. Two important determinants of location choice model in this study are the accessibility measures and the suitability analysis indicators. The model focuses on the statistical technique for analyzing discrete choice data by using econometric and Geographic Information System software. The findings in this paper show that agriculture, mining, electricity, gas and water, transport and finance firms' type are mostly located outside of Kuala Lumpur's Central Business District area. Meanwhile, manufacturing, construction and wholesale firms' type are located in the Central Business District area. The result of this study will highlight the use of discrete choice models in the analysis of firm location decisions which will be a foundation to facilitate town planners and decision makers to understand the firm location decisions in their region.

Discrete choice modelling of natal dispersal: ‘Choosing’ where to breed from a finite set of available areas

Summary Classic natal dispersal studies focus mainly on distance travelled. Although distances capture some of the main selective pressures related to dispersal, this approach cannot easily incorporate the properties of the actual destination vs. the available alternatives. Recently, movement ecology studies have addressed questions on movement decisions in relation to availability of resources and/or availability of suitable habitats through the use of discrete choice models (DCMs), a widely used type of models within econometrics, which explains individual choices as a function of the properties of a finite number of alternatives. In this contribution, we show how the dispersal discrete choice model (DDCM) can be used for analysing natal dispersal data in patchy environments given that the natal and the breeding area of the disperser are observed. We test this method using a case study on Great Tits (Parus major) in an archipelago of small woodlots. Our results show that DDCMs are able to capture the results of classic distance‐based approaches and simultaneously allow testing hypotheses on how departure and settlement are affected by variables that characterize the disperser, the natal patch and the breeding area, as well as their interactions. DDCMs can be applied to any other species and system that uses some form of discrete breeding location or a certain degree of discretization can be applied.

A discrete choice model based on random utilities for exit choice in emergency evacuations

This work presents a behavioural model based on discrete choice models (DCMs) in order to simulate decision maker behaviour during an emergency situation. The proposed model is a probabilistic exit choice model that allows understanding the heterogeneity of the different decision maker’s tastes. A stated preference survey was designed and realized with a sample of decision makers in order to achieve this objective. Thus, the obtained data was used to model the optimal mixed logit model. The use of DCMs allows to know the behaviour of different decision maker types during an evacuation process. The case study highlights the importance of the influence of other decision makers on the decision-making process. This work could be used as the starting point in the development of behavioural models which could be implemented in the current tools for the simulation of emergency evacuation.

Introduction to the Virtual Issue: Recent Advances in Discrete Choice Methods in Political Science

A “discrete choice” situation occurs when a political agent (i.e., voter, political leader, country) faces a choice or a series of choices from among a finite set of discrete options. These types of situations are frequently studied by political scientists, but it is only in the last decade that our understanding and use of discrete choice models has really started to evolve. While even into the mid-1990s it was not unusual to see linear regression applied to binary or ordinal dependent variables (e.g., Gerber 1996), now discrete choice models are part of the standard toolkit for applied researchers, and methodological training in the use of these models is required in many graduate programs. The leading edge of the field has also advanced rapidly. In the last decade political methodologists have moved beyond the “welcoming” of discrete choice methods from economics and other fields (Beck 2000) to the development of new discrete choice methods. In this virtual issue I have selected six articles from Political Analysis that I feel best exemplify this trend. These articles are:

Modelling human perception of static facial expressions

A recent internet based survey of over 35,000 samples has shown that when different human observers are asked to assign labels to static human facial expressions, different individuals categorize differently the same image. This fact results in a lack of an unique ground-truth, an assumption held by the large majority of existing models for classification. This is especially true for highly ambiguous expressions, especially in the lack of a dynamic context. In this paper we propose to address this shortcoming by the use of discrete choice models (DCM) to describe the choice a human observer is faced to when assigning labels to static facial expressions. Different models of increasing complexity are specified to capture the causal effect between features of an image and its associated expression, using several combinations of different measurements. The sets of measurements we used are largely inspired by FACS but also introduce some new ideas, specific to a static framework. These models are calibrated using maximum likelihood techniques and they are compared with each other using a likelihood ratio test, in order to test for significance in the improvement resulting from adding supplemental features. Through a cross-validation procedure we assess the validity of our approach against overfitting and we provide a comparison with an alternative model based on Neural Networks for benchmark purposes.

Standard and Shuffled Halton Sequences in a Mixed Logit Model

Modeling consumer choice in different areas has lead to an increase use of discrete choice models. Probit or Multinomial Logit Models are often the base of further empirical research of consumer choice. In some of these models the equations to solve have no closed-form expression. They include multi-dimensional integrals which can not be solved analytically. Simulation methods have been developed to approximate a solution for these integrals. This paper describes the Standard Halton sequence and a modification of it, the Shuffled Halton sequence. Both are simulation methods which can reduce computational effort compared to a random sequence. We compare the simulation methods in their coverage of the multi-dimensional area and in their estimation results using data of consumer choice on grocery store formats.

On the use of discrete choice models for causal inference

Methodology for causal inference based on propensity scores has been developed and popularized in the last two decades. However, the majority of the methodology has concentrated on binary treatments. Only recently have these methods been extended to settings with multi-valued treatments. We propose a number of discrete choice models for estimating the propensity scores. The models differ in terms of flexibility with respect to potential correlation between treatments, and, in turn, the accuracy of the estimated propensity scores. We present the effects of discrete choice models used on performance of the causal estimators through a Monte Carlo study. We also illustrate the use of discrete choice models to estimate the effect of antipsychotic drug use on the risk of diabetes in a cohort of adults with schizophrenia.

Constructing Quality-Adjusted Price Indices: A Comparison of Hedonic and Discrete Choice Models

The Boskin report (1996) concluded that the US consumer price index (CPI) overestimated the inflation by 1.1 percentage points. This was due to several measurement errors in the CPI. One of them is called quality change bias. We compare two methods in this paper which can be used to correct for quality change bias, namely the hedonic method and a method based on the use of discrete choice models. We compare the underlying micro-economic models of the two methods as well as their empirical implementation. Although the discrete choice model has not been used often to calculate quality-adjusted price indices, past research shows that it might be beneficial to do so. JEL Classification: C43, D11, D21, C25

The use of discrete choice models in the determination of community choices in public issue areas impacting on business decision making

Responsible business are increasingly recognizing the need to be sensitive to the local environmental implications of decisions taken in the course of establishing and operating a business. Rather than rely on the pressures of lobby groups to direct business behavior in relation to community concerns, a preferred strategy is to identify the preferences and choices of the community as a whole and to use information from a representative cross-section from the community to aid in making environmentally linked decisions that maximize the benefits to the affected community. This study demonstrates how discrete-choice models can be used to identify community choices between alternative traffic management schemes designed to improve the traffic impacts of business decisions such as the location of factories, offices, and retail outlets. We also indicate the loss of orthogonality, which is unavoidable when discrete-choice models are estimated on disaggregate data as distinct from grouped data.

The use of discrete choice models in the determination of community choices in public issue areas impacting on business decision making

Responsible business are increasingly recognizing the need to be sensitive to the local environmental implications of decisions taken in the course of establishing and operating a business. Rather than rely on the pressures of lobby groups to direct business behavior in relation to community concerns, a preferred strategy is to identify the preferences and choices of the community as a whole and to use information from a representative cross-section from the community to aid in making environmentally linked decisions that maximize the benefits to the affected community. This study demonstrates how discrete-choice models can be used to identify community choices between alternative traffic management schemes designed to improve the traffic impacts of business decisions such as the location of factories, offices, and retail outlets. We also indicate the loss of orthogonality, which is unavoidable when discrete-choice models are estimated on disaggregate data as distinct from grouped data.

A Graphical Interpretation of Probit Coefficients

Although the use of discrete choice models (particularly the probit model) is widely taught, it is the method rather than the interpretation of results that is emphasized. This article provides a graphical technique for interpretation that will be useful in statistics and econometrics courses.

Marginal welfare measures for discrete choice models

This letter deals with the use of discrete choice models for applied welfare analysis. An approximate formula is presented for the marginal willingness to pay for a change in the attributes of any of the discrete alternatives which is simple to calculate and has an intuitive interpretation.