International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
Including Value Orientations in Choice Models to Estimate Benefits of Wildlife Management Policies
Sandra Notaro, Gianluca Grilli, Daniel Campbell

Last modified: 28 March 2017


The use of attitudinal questions in stated Choice Experiments (CE) has always represented a matter of interest for many analysts. However, many argue that the common practice of including such indicators in the deterministic part of the utility function may lead to endogeneity bias (Hess and Stathopoulos, 2014). For this reason, hybrid models recently gained importance (Ben-Akiva et al., 2002; Bolduc et al., 2005; Bolduc and Alvarez-Daziano, 2010). The idea of hybrid models is that individual attitudes are treated as latent variables. The latent variable, explained by personal socio-demographic characteristics, is used to model a certain number of attitudinal indicators, treated as proxies for the underlying individual attitude, and the choice model at the same time. In this framework, the answers to attitudinal questions are included in the model but they are not in the deterministic part of the utility function, thus solving endogeneity issues.

In our contribution, we make use of the hybrid methodology to model value orientations of tourists towards wildlife. Value orientations are human mind’s psychological concepts that serve as intermediary between values, ideologies of a person and his behaviour and opinion about a specific issue (Manfredo, 2008).

Several studies have demonstrated the predictive power of wildlife value orientation in explaining people attitudes toward wildlife management options, wildlife viewing, hunting and fishing (Manfredo et al., 2009; Vaske et al., 2011), but none of these studies have used an economic approach. As well wildlife value orientations can be seen as at the root of most conflicts in public opinion over conservation and wildlife management policies (Teel et al., 2010).

In the literature it is possible to find two main orientations, mutualism and domination. Mutualism is strongly related to the importance of wildlife’s non-consumptive value. Mutualist people tend to treat animals at the same level as humans, thus struggling for animal rights and opposing to killing and exploitation. Conversely, dominant behavior is strongly related to a utilitarian vision of the human-wildlife relationship. A person with this value orientation will have no problem to accept control actions that can provoke the death or harms animals or simply restricts animals’ freedom (Tell et al., 2010).

To the best of our knowledge, our study is the first that uses values orientations as latent variable in a hybrid choice model. In order to collect individual value orientations, we used a reduced form of the questionnaire proposed by Manfredo (2008); 14 proposition were included, 7 related to the dominant orientation and 7 to the mutualist orientation, collected on a 7-points scale. In order to simplify the model and reducing the number of parameters to be estimated, two indicators per each orientation were selected, those one showing the highest values for the two value orientations.

The empirical study concerns the management of three threatened species in the Italian Alps, the wolf (Canis lupus), the lynx (Lynx lynx) and the aurora salamander (Salamandra atra aurorae). The field CE was carried out in Summer 2015, interviewing 420 tourists face-to-face. Four attributes were selected, three non-monetary attributes and the cost. Non-monetary attributes were the size of wolf, lynx and Aurora salamander populations (number of individuals). The cost associated with each alternative was an entrance fee to natural areas. Local wildlife is jointly managed by local administration and institutions of natural and protected areas. The latter usually suffer from lack of funding, thus contribution from users may represent a solution for self-financing wildlife-related activities. Combinations of attribute levels were arranged into choice cards by means of a Bayesian efficient design (Rose and Bliemer, 2009). Priors were based on data from a pilot of 63 tourists. Each respondent was required to answer 12 choice tasks, composed by two alternatives plus the status quo, though the best-worst method (Louviere and Islam,  2008; Scarpa et al., 2010). Considering that each respondent provided 24 observations, it was possible to acquire a total of 10,080 observations for the econometric analysis. Analyses were conducted by means of a latent class logit to model choice data, while values orientations, used as indicators, were modelled with ordered logit. The latent variable enters the model as a variable affecting class membership. The methodology was proposed by Hess et al. (2012) and enables the possibility to understand how latent attitudes affect choices through the probability of being allocated in a certain class (Hoyos et al., 2015; Mariel et al., 2015). The joint likelihood function depends on parameters of the utility function describing choices, class membership parameters, parameters of the attitudinal questions and parameters of socio-demographic characteristics, included in the latent variables. Components of this likelihood function were estimated simultaneously using R (R Core Team, 2013).

Preliminary results suggest that latent variables play a role in class allocation, in particular for what concerns the dominant attitude. An increase in the significance levels of the parameters between the simple latent class and the hybrid model is experienced, thus confirming that the hybrid modeling may improve the model precision. Concerning preferences for wildlife populations, there are differences between classes, both in term of coefficient and related WTP. This is another evidence of the effectiveness of the latent variable to model attitudes towards wildlife. In particular, people belonging to the first class showed a positive WTP for the wolf, although declining as the population size increases. Negative WTP were instead recorded for lynx and salamanders. In the second class, people showed positive WTP for the population of the three animals. In general, results suggest differences between the hybrid model and the simple latent class taken as reference model and a general improvement of the results, thus indicating that value orientation as attitudinal variable play a role in modeling preferences. The resulting improved model is useful for policy-makers, because they can rely on better data for anticipating possible conflicts over proposed wildlife management policies. This knowledge can also be useful for adapting communication over these policies in order to increase the success of conservation policies.


Keywords: Choice experiment, hybrid model, value orientations, latent attitudes



Ben-Akiva, M., McFadden, D., Train, K., Walker, J., Bhat, C., Bierlaire, M., Bolduc, D., Boersch-Supan, A., Brownstone, D., Bunch, D.S., Daly, A., de Palma, A., Gopinath, D., Karlstrom, A., Munizaga, M. a., 2002. Hybrid Choice Models : Progress and Challenges. Mark. Lett. 13, 163–175. doi:10.1023/A:1020254301302

Bolduc, D., Alvarez-Daziano, R., 2010. On estimation of hybrid choice models, in: Choice Modelling: The State-of-the-Art and the State-of-Practice: Proceedings from the Inaugural International Choice Modelling Conference. Emerald Group Publishing. p. 259.

Bolduc, D., Ben-Akiva, M., Walker, J., Michaud, A., 2005. Hybrid choice models with logit kernel: applicability to large scale models. Integr. Land-Use Transp. Model. Behav. Found. Elsevier, Oxford 275–302.

Hess, S., Beharry-Borg, N., 2012. Accounting for latent attitudes in willingness-to-pay studies: the case of coastal water quality improvements in Tobago. Environ. Resour. Econ. 52, 109–131.

Hess, S., Stathopoulos, A., 2014. Linking response quality to survey engagement: a combined random scale and latent variable approach, in: 93rd Annual Meeting of the Transportation Research Board. pp. 1–28.

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. Sci. Total Environ. 505, 1100–1111. doi:10.1016/j.scitotenv.2014.10.066

Manfredo, M.J., 2008. Who cares about wildlife? Springer.

Manfredo, M.J., Teel, T.L., Henry, K.L., 2009. Linking society and environment: A multilevel model of shifting wildlife value orientations in the western United States. Soc. Sci. Q. 90, 407–427.

Mariel, P., Meyerhoff, J., Hess, S., 2015. Heterogeneous preferences toward landscape externalities of wind turbines - Combining choices and attitudes in a hybrid model. Renew. Sustain. Energy Rev. 41, 647–657. doi:10.1016/j.rser.2014.08.074

Louviere, J. J., and T. Islam. 2008. A Comparison of Importance Weights and Willingness-to-Pay Measures Derived from Choice-Based Conjoint, Constant Sum Scales and Best-Worst Scaling.  Journal of Business Research 61: 903–911.

R Core Team, R.C., 2013. R: A language and environment for statistical computing.

Rose, J.M., Bliemer, M.C.J., 2009. Constructing Efficient Stated Choice Experimental Designs. Transp. Rev. 29, 587–617. doi:10.1080/01441640902827623

Scarpa, A.R., Notaro, S., Louviere, J., Raffaelli, R., 2010. Exploring Scale Effects of Best/Worst Rank Ordered Choice Data to Estimate Benefits of Tourism in Alpine Grazing Commons. Am. J. Agric. Econ. 1–28.

Teel, T.L., Manfredo, M.J., Jensen, F.S., Buijs, A.E., Fischer, A., Riepe, C., Arlinghaus, R., Jacobs, M.H., 2010. Understanding the cognitive basis for human-wildlife relationships as a key to successful protected-area management. Int. J. Sociol. 40, 104–123.

Vaske, J.J., Jacobs, M.H., Sijtsma, M.T.J., 2011. Wildlife value orientations and demographics in The Netherlands. Eur. J. Wildl. Res. 57, 1179–1187.

Conference registration is required in order to view papers.