International Choice Modelling Conference, International Choice Modelling Conference 2015

Font Size: 
Does the place of residence affect land use preferences? – Evidence from a discrete choice experiment in Germany
Julian Sagebiel

Last modified: 18 May 2015


In discrete choice experiments (DCE), individuals choose between alternatives which vary in their attributes. Frequently, a status quo alternative is included in the choice set as a reference point, allowing for calculation of welfare estimates. In some contexts, the status quo is similar to all respondents or framed as an opt-out alternative, for example the valuation of a specific recreational site (e.g., Colombo, Hanley, & Louviere, 2009). Yet in some applications a common status quo alternative is not feasible as respondents have different status quo situations (Hess, Rose, & Hensher, 2008; Stathopoulos & Hess, 2012). Consider a DCE to value changes in local landscape. Spatially distributed respondents reside in different landscapes, implying unique status quo situations. In the literature, it is argued that such differences impact estimated welfare changes. This may be reference point bias, loss aversion or diminishing marginal utility.

For example, Glenk (2011) asked respondents in a DCE on environmental services in Indonesia about their perception of the status quo and used this data in model estimation. He estimated separate utility parameters for those, who regard the attribute levels as an improvement (i.e., those who have a low status quo level) or a loss.


In this paper, the effect of different residence locations (i.e., different reference points) on preferences for land use changes is investigated. The main aim is to study whether respondents living in different landscapes, distinguished for example by the share of forest, the degree of biodiversity etc., behave differently in DCEs. Thereby, three types of differences are tested: differences in variance scale (Swait & Louviere, 1993), differences in willingness to pay and preferences (Poe, Giraud, & Loomis, 2005), and, differences in the status quo effect (Scarpa, Ferrini, & Willis, 2005). The first type, differences in variance scale, imply that some respondents are more certain in their decisions than others. For example, a respondent living in a nature rich area is more aware of landscape attributes than a respondent from an urban area, thus making more informed choices. The second type is concerned with the reference point. Here, the individual status quo, as a reference point for the decision impacts the willingness to pay. For example, a respondent living in a forest rich area has a lesser willingness to pay for an increase in forest share than a respondent living in an urban landscape. The third type, the status quo effect, means that respondents view the status quo option differently from the other options. It might be the case that differences between landscape categories exist when, for example, respondents living in a flawless nature have a preference that things remain as they are, regardless of the attribute levels of the other alternatives.

The data for the analysis come from an online DCE survey with 1409 respondents in Germany. The land change scenarios were represented as a change in their local surroundings ‑‑ a 15 kilometer radius around the respondents' principle residence. Thus, each respondent is subject to an individual status quo. In the DCE, two alternative landscape scenarios were presented with a third alternative comprising the status quo of the respondents i.e., all attributes were described with as today. In total there were six attributes describing the local landscape in terms of agricultural land use, biodiversity, recreational uses and forest share. An interactive map allowed the respondents to indicate the coordinates of their principal residence. The coordinates were used to form landscape categories based on spatial data from the Federal Agency for Nature Conservation, Germany (BfN). The categories distinguish between the share of forest, whether the area is urban or rural, indicators for the degree of biodiversity and share of agriculture. In total four of these categories were used to classify the respondents’ status quo situation.

The analysis is structured as follows: In a first step, nonparametric tests are used to explore whether the landscape categories are correlated with exogenous variables like age, gender etc. This is important to control for endogeneity issues. Next, separate mixed logit models for each landscape sample are estimated Thereby, the Louviere and Swait test is applied to investigate scale differences and the Poe et al. test is used to infer differences in willingness to pay values. Finally, a large model with all samples is estimated. Using a mixed logit specification, interaction terms with the landscape categories and the alternative specific constant for the status quo option are integrated into the models. A significant alternative specific constant indicates the existence of a status quo effect. The interactions give information how this effect differs between the landscape categories.



Colombo, S., Hanley, N., & Louviere, J. J. (2009). Modeling preference heterogeneity in stated choice data: an analysis for public goods generated by agriculture. Agricultural Economics, 40(3), 307–322. doi:10.1111/j.1574-0862.2009.00377.x

Glenk, K. (2011). Using local knowledge to model asymmetric preference formation in willingness to pay for environmental services. Journal of Environmental Management, 92(3), 531–541. doi:10.1016/j.jenvman.2010.09.003

Hess, S., Rose, J. M., & Hensher, D. A. (2008). Asymmetric preference formation in willingness to pay estimates in discrete choice models. Transportation Research Part E: Logistics and Transportation Review, 44(5), 847–863. doi:10.1016/j.tre.2007.06.002

Poe, G. L., Giraud, K. L., & Loomis, J. B. (2005). Computational Methods for Measuring the Difference of Empirical Distributions. American Journal of Agricultural Economics, 87(2), 353–365. doi:10.1111/j.1467-8276.2005.00727.x

Scarpa, R., Ferrini, S., & Willis, K. (2005). Performance of Error Component Models for Status-Quo Effects in Choice Experiments. In R. Scarpa & A. Alberini (Hrsg.), Applications of Simulation Methods in Environmental and Resource Economics (S. 247–273). Springer Netherlands. Abgerufen von

Stathopoulos, A., & Hess, S. (2012). Revisiting reference point formation, gains–losses asymmetry and non-linear sensitivities with an emphasis on attribute specific treatment. Transportation Research Part A: Policy and Practice, 46(10), 1673–1689. doi:10.1016/j.tra.2012.08.005

Swait, J., & Louviere, J. (1993). The role of the scale parameter in the estimation and comparison of multinomial logit models. JMR, Journal of Marketing Research, 30(3), 305–314.

Conference registration is required in order to view papers.