|
on Discrete Choice Models |
| By: | Simonsen, Nick (Aarhus University) |
| Abstract: | Discrete choice experiments are widely used in the social sciences to understand individuals’ stated preferences. Given their popularity, many diQerent software packages, such as Ngene, are available to create discrete choice experiment designs. However, integrating these designs into online platforms like Qualtrics can be time-consuming and often challenging. This paper addresses these challenges and provides a detailed tutorial on converting Ngene outputs into a Qualtrics-compatible format using custom code and integrating it into Qualtrics to create a fully functional online discrete choice experiment. The provided code simplifies the process, making it more accessible and manageable for researchers. This approach minimizes time and eQort and potential human errors when integrating the design into the online survey platform. Lastly, the paper provides code for extracting and formatting the post-experiment choice data into a meaningful data format that can be used for data analysis. In sum, this paper equips researchers with tools to easily deploy discrete choice experiment designs created using the Ngene software to Qualtrics and data extraction post-data collection. |
| Date: | 2025–02–12 |
| URL: | https://d.repec.org/n?u=RePEc:osf:osfxxx:rj7yh_v1 |
| By: | Ertian Chen |
| Abstract: | Estimating dynamic discrete choice models with large state spaces poses computational difficulties. This paper develops a novel model-adaptive approach to solve the linear system of fixed point equations of the policy valuation operator. We propose a model-adaptive sieve space, constructed by iteratively augmenting the space with the residual from the previous iteration. We show both theoretically and numerically that model-adaptive sieves dramatically improve performance. In particular, the approximation error decays at a superlinear rate in the sieve dimension, unlike a linear rate achieved using conventional methods. Our method works for both conditional choice probability estimators and full-solution estimators with policy iteration. We apply the method to analyze consumer demand for laundry detergent using Kantar's Worldpanel Take Home data. On average, our method is 51.5% faster than the conventional methods in solving the dynamic programming problem, making the Bayesian MCMC estimator computationally feasible. The results confirm the computational efficiency of our method in practice. |
| Date: | 2025–01 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2501.18746 |
| By: | Paul H. Y. Cheung; Yusufcan Masatlioglu |
| Abstract: | We explore the influence of framing on decision-making, where some products are framed (e.g., displayed, recommended, endorsed, or labeled). We introduce a novel choice function that captures observed variations in framed alternatives. Building on this, we conduct a comprehensive revealed preference analysis, employing the concept of frame-dependent utility using both deterministic and probabilistic data. We demonstrate that simple and intuitive behavioral principles characterize our frame-dependent random utility model (FRUM), which offers testable conditions even with limited data. Finally, we introduce a parametric model to increase the tractability of FRUM. We also discuss how to recover the choice types in our framework. |
| Date: | 2025–01 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2502.00209 |
| By: | Zack Dorner (Lincoln University); Steven Tucker (University of Waikato); Abraham Zhang (Adam Smith Business Svchool, University of Glasgow); Anna Huber (University of Natural Resources and Life Sciences, Vienna) |
| Abstract: | Information represents the "third wave" of environmental policy. Existing evidence shows consumers increase their willingness to pay (WTP) for environmentally friendly products with clear labelling. However, there is a gap in the literature regarding whether consumers have a willingness to engage (WTE) with detailed information, for example, through a Digital Product Passport (DPP). This technological innovation is part of the European Union's new circular economy action plan. In our theoretical model, a green consumer decides whether to invest in information on how to mitigate their environmental damage, but at a cognitive cost. We test the model in a lab experiment selling an environmentally friendly toothbrush, but information about its environmental credentials is only available through a DPP. We find education on the DPP's purpose is key to increasing revealed WTE when a DPP is available. Participants with a high stated WTE engage with the DPP regardless; the increase in revealed WTE comes from those with a lower stated WTE. Engagement with the DPP, in the case that it contains positive environmental information, increases WTP. The policy implications of our results are that education about the purpose of the DPP is required in order to increase the likelihood of actual consumer engagement with it, as long as it is user friendly. However, engagement with a DPP may not lead to further shifts in environmental orientation and behavior. Our study also demonstrates novel measures of WTE, and how these can be used to understand pro-environmental consumer behavior in a theoretically informed manner. |
| Keywords: | Circular economy; digital product passport; consumer behavior; ecolabel; green consumerism; information-based instruments; pro-environmental behavior |
| JEL: | C92 D12 D18 D63 D64 D91 |
| Date: | 2025–02–21 |
| URL: | https://d.repec.org/n?u=RePEc:wai:econwp:25/03 |
| By: | Cai, Liang; Song, Guangwen; Zhang, Yanji |
| Abstract: | Objectives Although the social disorganization tradition emphasizes the role of neighborhood context in shaping delinquent behaviors and neighborhood crime, researchers have rarely considered the influence of neighborhood context on criminals’ decision of where to offend. This study explicitly examines how concentrated disadvantage in both the origin and destination neighborhoods structures burglars’ preference for street physical disorder and spatial familiarity. Methods We measure observed and perceived physical disorder from 107, 858 street view images using computer vision algorithms. Geo-referenced mobile phone flows between 1, 642 census units are used to approximate offenders’ potential spatial knowledge about target neighborhoods. Discrete choice models are estimated separately for burglars from disadvantaged and non-disadvantaged neighborhoods (N=1, 972). Results While burglars residing in non-disadvantaged neighborhoods are not sensitive to physical disorder in non-disadvantaged target neighborhoods, they strongly avoid disadvantaged neighborhoods with disorder. Conversely, residents of neighborhoods with concentrated disadvantage swiftly act upon street disorder in better-off neighborhoods but not in disadvantaged neighborhoods. These tendencies to react to the presence of physical disorder on the street are also contingent on burglars’ potential familiarity with the target environment. Conclusions We highlight the importance of larger neighborhood structural characteristics and their interactions with spatial knowledge and environmental conditions such as visual signs of disorder, in criminal decision making. Physical disorder is not uniformly indicative of decay across neighborhoods and offenders. This divergent decision-making may also partially explain spatial heterogeneity of crime. Moreover, spatial knowledge is most effective in triggering or deterring actions in places that are categorically different from offenders’ residential spaces. |
| Date: | 2024–12–22 |
| URL: | https://d.repec.org/n?u=RePEc:osf:socarx:rcny3_v1 |
| By: | John List |
| Abstract: | Contingent valuation is a widely used method for estimating the value of nonmarket commodities. Yet, a persistent issue is whether responses to Contingent Valuation Method (CVM) questions accurately reflect true values. Recent studies indicate that hypothetical bias is a significant factor that creates a gap between intentions and actions. I use a novel approach within non-market valuation - a List Experiment in the field - to test whether it can attenuate the hypothetical bias observed within CVM surveys. Using data from 400 subjects in a field experiment, I find initial promising results. |
| Date: | 2025 |
| URL: | https://d.repec.org/n?u=RePEc:feb:framed:00809 |
| By: | Roberto-Rafael Maura-Rivero; Marc Lanctot; Francesco Visin; Kate Larson |
| Abstract: | Reinforcement Learning from Human Feedback (RLHF), the standard for aligning Large Language Models (LLMs) with human values, is known to fail to satisfy properties that are intuitively desirable, such as respecting the preferences of the majority \cite{ge2024axioms}. To overcome these issues, we propose the use of a probabilistic Social Choice rule called \emph{maximal lotteries} as a replacement for RLHF. We show that a family of alignment techniques, namely Nash Learning from Human Feedback (NLHF) \cite{munos2023nash} and variants, approximate maximal lottery outcomes and thus inherit its beneficial properties. We confirm experimentally that our proposed methodology handles situations that arise when working with preferences more robustly than standard RLHF, including supporting the preferences of the majority, providing principled ways of handling non-transitivities in the preference data, and robustness to irrelevant alternatives. This results in systems that better incorporate human values and respect human intentions. |
| Date: | 2025–01 |
| URL: | https://d.repec.org/n?u=RePEc:arx:papers:2501.19266 |