Spatial Dimensions of Stated Preferences

Pre-Conference Workshop

Date and Time: Wednesday 28 June, 9 a.m. to 4 p.m.

Venue:  Historical premises of the National and Kapodistrian University of Athens

Spatial dimensions of welfare estimation are increasingly being considered in stated preference studies. These include the relevance of spatial information for the design of the survey (e.g., spatially explicit choice tasks), the data collection process (e.g., spatial sampling), the analysis of choice data (e.g., estimating spatially informed discrete choice models such as the geographically weighted models), the use and aggregation of benefit estimates (e.g., when is knowledge of spatial heterogeneity required for cost-benefit analysis), and the presentation of results (e.g., willingness to pay (WTP) maps). However, there are still many open questions on when and how to incorporate spatial information and modelling in stated preference studies. Even if researchers suspect that spatial patterns are relevant, there is a lack of guidance on best practices to accommodate these patterns within stated preference survey design, value elicitation and data analysis.

This lack of guidance has led to inconsistent and sometimes ill-informed practices across the stated preference literature.  For example, researchers often fail to connect spatial components of sampling and survey design (e.g., spatial aspects of policy scenarios communicated explicitly or implicitly by questionnaires) with estimation methods that can accommodate them. There is also a need to better understand how spatial factors can influence behaviour, and how this behaviour is represented in analysis. Moreover, even when specific types of spatial patterns are suspected (e.g., spatial welfare heterogeneity), there is no clear consensus across the literature on how to incorporate it into statistical models. Multiple approaches exist, each with seeming advantages and disadvantages. Given this lack of clarity in the literature, developing a broader, more systematic understanding of these issues is a necessary starting point to move towards a more theoretically and statistically sound treatment of spatial issues within stated preference valuation.

This pre-conference workshop at EAERE 2017 in Athens directly follows and expands upon the thematic session Willingness to pay in space: New insights into the spatial dimension of stated preferences, held at the 2016 EAERE conference in Zurich. It responds to queries and requests received in the original thematic session, and specifically the need for: (a) in-depth discussion of the challenges and opportunities related to spatial dimensions of stated preference design and modelling, and (b) a pioneering set of “best practice guidelines” for the treatment of space within stated preference valuation.  It will engage a variety of experts in different areas of stated preference design and modelling to address ongoing challenges related to the role of space in stated preference studies, and provide an opportunity for participants at all levels of experience to learn about methods to model and accommodate spatial dimensions of stated preferences. 

9:00 Welcome and Introduction: The Relevance of Space for Stated Preference Design, Value Elicitation and Model Estimation (Robert Johnston)

9:15 When and Why Go Spatial in Stated Preference Analysis? An Overview and Outlook (30 minutes presentation + 15 minutes discussion)

Robert Johnston (Clark University)

Klaus Glenk (SRUC)

Jürgen Meyerhoff (Technische Universität Berlin)

Julian Sagebiel (Institute for Ecological Economy Research, Berlin)

10:00 Presentations on current topics concerning the spatial dimension of stated preferences (15 minutes presentation + 5 minutes discussion). Moderation: Julian Sagebiel.

  10:00 Mixed Logit, Geographically Weighted Choice Models, or One-Step Bayesian Estimation? Mikolaj Czajkowski and Wiktor Budziński (University of Warsaw)

  10:20 The Use of Semi-parametric Models to Account for Biases from Omitted Spatial Covariates. Brett Day (University of Exeter)

10:40 Coffee Break

11:00 Presentations continued (15 minutes presentation + 5 minutes discussion). Moderation: Klaus Glenk

  11:00 A new approach to capturing the spatial dimensions of value within choice experiments. Tomas Badura (University of East Anglia)

  11:20 Distance decay and spatial preference heterogeneity: Combining stated and revealed preference data. Jens Abildtrup (French National Institute for Agricultural Research)

  11:40 Does the functional form of distance decay depend on the different ecological endpoints that enter respondents’ (heterogeneous) utility functions? Søren Bøye Olsen (University of Copenhagen)

  12:00 Spatial preference heterogeneity in discrete choice site selection models. Roy Brouwer (University of Waterloo)

12:30 Lunch (will be provided)**

13:30 Panel Discussion: Questions and Frontiers in Spatial Dimensions of Stated Preferences - What are the questions that we need to answer, and what are the important research frontiers? (Moderation: Robert Johnston)

Panellists:

Roger H. von Haefen (North Carolina State University)

Ståle Navrud (Norwegian University of Life Sciences)

Toke Emil Panduro (University of Copenhagen)

Jette Bredahl Jacobsen (University of Copenhagen)

Eija Pouta (Natural Resources Institute Finland)

Each panellist spends about 10 minutes speaking on a specific topic, followed by 30 minutes of panel discussion with the audience.

14:45 Coffee Break

15:00 Group Discussion: Coordinating Approaches across the Literature and Steps towards Best Practice (Moderation: Jürgen Meyerhoff)

Key questions:

(1) When and how are spatial dimensions a first-order concern for analysis?

(2) Can we better coordinate myriad approaches across the literature and provide input on to when different approaches are relevant?

(3) How can be better leverage theory to guide spatial analysis? And when is spatial analysis purely empirical?

(4) Are there other key questions that have not yet been discussed?

(5) Are best practice guidelines needed?

15:50 Next Steps and Announcements: Jürgen Meyerhoff

16:00 Workshop concludes

*Literature

**Note: In order to cover expenses for coffee breaks and lunch, we ask each participant to pay 15 Euros (please bring the exact amount, as we will not have the opportunity to change money)