Choice Set Research Articles

Women's preferences for caesarean or vaginal birth with a perspective of future fertility: A discrete choice experiment.

To investigate pregnant women's preferences for risks of vaginal and caesarean birth, including possible impacts on future fertility. In this discrete choice experiment, low-risk nulliparous pregnant women recruited after 28 weeks of gestation evaluated eight choice sets, each between two different hypothetical births scenarios which intermixed the risks of planned caesarean or vaginal birth. Scenarios consisted of six attributes: pain, maternal health, neonatal health, risk of unplanned intervention, impact on fertility and risk of complications in the next pregnancy. All scenarios contained risks to neonatal health as neither vaginal nor caesarean birth guarantee an ideal outcome. Choice data were analysed using a conditional logistic regression model. Between June and September 2023, 211 participants, including 34 from pilot interviews, completed the questionnaire. Influential attributes were maternal health (conditional odds ratio [COR] 1.29, 95% CI 1.17 to 1.42, p<0.001) and risk of unplanned intervention (COR 1.37, 95% CI 1.24 to 1.51, p<0.001), favouring caesarean birth. Conversely, impact on fertility (COR 0.75, 95% CI 0.68 to 0.83, p<0.001) and complications in the next pregnancy favoured vaginal birth (COR 0.90, 95% CI 0.82 to 1.00, p = 0.045). Participants weighed the included morbidity risks of planned caesarean and vaginal birth in a low-risk pregnancy approximately equally. To facilitate an informed birth decision, clinicians should, apart from neonatal outcomes, particularly consider discussing impacts on fertility, maternal health and the risks of unplanned intervention or future pregnancy complications.

Diversity within universality: Explaining pandemic universal cash transfers in East Asia

AbstractThe response to the global COVID‐19 pandemic has prompted a surge in short‐term universal cash transfer programmes around the world. Notably, East Asian high‐income economies have been at the forefront of these initiatives. While the innovative nature of these universal cash injections has been emphasized, there is limited documentation regarding their characteristics, prospects, and underlying motivations. This article sheds light on the domestic political and institutional processes that led to the implementation and design of universal cash transfers in Hong Kong (China), Japan, Republic of Korea, Singapore and Taiwan (China). Overall, the analysis reveals that, within the framework of universality, a nuanced, diverse and dynamic set of operational choices emerge. A range of factors shaped the adoption and evolution of these programmes, including, for example, political pressures stemming from political party competition and efforts to maintain political legitimacy. In general, design parameters are not only defined in technocratic terms, but are negotiated politically.

The role of farming systems in the adoption of pest management practices: the case of the oriental fruit fly Bactrocera dorsalis in mango orchards in Senegal

Description of the suject. In Africa, pest pressure has led researchers, policymakers, and international development stakeholders to develop and disseminate innovative pest management strategies. However, farmers reject some pest management strategies due to economic constraints, agronomic problems, or incompatibilities with their strategies, knowledge bases, and visions of desirable change. Objectives. In this paper, we study the pest management practices employed by mango producers from different farming systems against the fruit fly Bactrocera dorsalis, as well as the factors determining the level of engagement of these producers in managing this pest. Method. We surveyed 304 Senegalese mango growers affected by the invasive fruit fly Bactrocera dorsalis to understand which farming systems’ features could explain the adoption level of fruit fly management tools through multivariate data analysis. Results. Three types of farming systems characterized by a coherent set of varietal choices, orchard management, harvest practices, commercial outlets, and fruit fly management practices are identified: (1) Intensive farming systems supplying the export market with the variety Kent and using a large set of orchard management tools and harvest practices to eliminate fruit flies, (2) Extensive farming systems supplying various varieties of mangoes with a dominance of Kent to both export and domestic markets, with access to fruit fly management tools from public services, (3) Gathering farming systems characterized by zero-input, high varietal diversity, difficult market access, non-selective harvest practices, and fruit fly management rarely practiced. Conclusions. The farming systems form a gradient of increasing use of fruit fly management tools and inputs, access to extension services, and export market orientation. Their coexistence in the same territories significantly contributes to the inefficiency of the control measures against Bactrocera dorsalis. Our results support the premise that in Africa, the adoption of pest management tools is more deeply integrated into the practices of certain producers compared to others, due to internal and external factors influencing the operation.

Generalizable Active Privacy Choice: Designing a Graphical User Interface for Global Privacy Control

The California Consumer Privacy Act and other privacy laws give people a right to opt out of the sale and sharing of personal information. In combination with privacy preference signals, especially, Global Privacy Control (GPC), such rights have the potential to empower people to assert control over their data. However, many laws prohibit opt out settings being turned on by default. The resulting usability challenges for people to exercise their rights motivate generalizable active privacy choice --- an interface design principle to make opt out settings usable without defaults. It is based on the idea of generalizing one individual opt out choice towards a larger set of choices. For example, people may apply an opt out choice on one site towards a larger set of sites. We explore generalizable active privacy choice in the context of GPC. We design and implement nine privacy choice schemes in a browser extension and explore them in a usability study with 410 participants. We find that generalizability features tend to decrease opt out utility slightly. However, at the same time, they increase opt out efficiency and make opting out less disruptive, which was more important to most participants. For the least disruptive scheme, selecting website categories to opt out from, 98% of participants expressed not feeling disrupted, a 40% point increase over the baseline schemes. 83% of participants understood the meaning of GPC. They also made their opt out choices with intent and, thus, in a legally relevant manner. To help people exercise their opt out rights via GPC our results support the adoption of a generalizable active privacy choice interface in web browsers.

Semiparametric Bayesian estimation of dynamic discrete choice models

We propose a tractable semiparametric estimation method for structural dynamic discrete choice models. The distribution of additive utility shocks in the proposed framework is modeled by location-scale mixtures of extreme value distributions with varying numbers of mixture components. Our approach exploits the analytical tractability of extreme value distributions in the multinomial choice settings and the flexibility of the location-scale mixtures. We implement the Bayesian approach to inference using Hamiltonian Monte Carlo and an approximately optimal reversible jump algorithm. In our simulation experiments, we show that the standard dynamic logit model can deliver misleading results, especially about counterfactuals, when the shocks are not extreme value distributed. Our semiparametric approach delivers reliable inference in these settings. We develop theoretical results on approximations by location-scale mixtures in an appropriate distance and posterior concentration of the set identified utility parameters and the distribution of shocks in the model.

EXAMINING VETERANS’ CHOICES OF HIGH-QUALITY MEDICARE SKILLED HOME HEALTH CARE AGENCIES

Abstract Medicare-certified skilled home health helps Veterans restore function post-hospitalization. Our objective was to understand the quality of VA contracted home health agencies (HHAs) available to Veterans (i.e., had at least 1 VA-paid episode of care in the 12 months) relative to the local market. For comparison, about one-third (30.4%) of traditional Medicare beneficiaries received care from high-quality HHAs (4- or 5-stars on quality of care) in 2015 (Schwartz et al., 2019). We used the Home Health CAHPS (Consumer Assessment of Healthcare Providers and Systems) 2016-2019 quarterly patient experience of care overall ratings (possible values: 1-5) to define a high-quality agency as those with 4- or 5-star ratings. We matched the quarterly ratings with the discharge date from a VA Medical Center to create HHA choice sets per eligible Veteran discharge. A choice set contains all contracted and non-contracted HHAs that serve a Veteran’s Zip Code. Among HHAs proximal to Veterans’ homes across all years (N= 10,507 unique HHAs), 54.8% were high-quality (n=12,449 HHA-years), 30.1% were 3-star quality (n=6,835 HHA-years), and 15.1% were low-quality (1- or 2- stars) (n=3,444 HHA-years). Among discharges, 53.2% (n=28,081) of the discharges went to high-quality HHAs and 15.7% (n=8,294) went to low-quality HHAs. Veterans tended to live in areas with access to high-quality home health care, and most discharges mapped to high-quality HHAs. However, there is room to improve Veterans’ access to high-quality skilled home health and more to understand about VA contracting practices and the HHA selection process.

Quantifying confidence shifts in a BERT-based question answering system evaluated on perturbed instances.

Recent work on transformer-based neural networks has led to impressive advances on multiple-choice natural language processing (NLP) problems, such as Question Answering (QA) and abductive reasoning. Despite these advances, there is limited work still on systematically evaluating such models in ambiguous situations where (for example) no correct answer exists for a given prompt among the provided set of choices. Such ambiguous situations are not infrequent in real world applications. We design and conduct an experimental study of this phenomenon using three probes that aim to 'confuse' the model by perturbing QA instances in a consistent and well-defined manner. Using a detailed set of results based on an established transformer-based multiple-choice QA system on two established benchmark datasets, we show that the model's confidence in its results is very different from that of an expected model that is 'agnostic' to all choices that are incorrect. Our results suggest that high performance on idealized QA instances should not be used to infer or extrapolate similarly high performance on more ambiguous instances. Auxiliary results suggest that the model may not be able to distinguish between these two situations with sufficient certainty. Stronger testing protocols and benchmarking may hence be necessary before such models are deployed in front-facing systems or ambiguous decision making with significant human impact.

Choice Modeling of Laundry Detergent Data for Sustainable Consumption

Consumer choice modeling takes center stage as we delve into understanding how personal preferences of decision makers (customers) for products influence demand at the level of the individual. The contemporary choice theory is built upon the characteristics of the decision maker, alternatives available for the choice of the decision maker, the attributes of the available alternatives and decision rules that the decision maker uses to make a choice. The choice set in our research is represented by six major brands (products) of laundry detergents in the Japanese market. We use the panel data of the purchases of 98 households to which we apply the hierarchical probit model, facilitated by a Markov Chain Monte Carlo simulation (MCMC) in order to evaluate the brand values of six brands. The applied model also allows us to evaluate the tangible and intangible brand values. These evaluated metrics help us to assess the brands based on their tangible and intangible characteristics. Moreover, consumer choice modeling also provides a framework for assessing the environmental performance of laundry detergent brands as the model uses the information on components (physical attributes) of laundry detergents. Through a comprehensive evaluation of product performance, including brand tangible estimation, we shed light on the sustainability attributes of laundry detergents, offering a roadmap for consumers and manufacturers alike to make more informed, environmentally responsible choices of laundry detergents based on their physical attributes. Knowing the estimates of the attributes for the laundry detergent products, manufacturers can modify their physical attributes, e.g., decrease the amount of the detergent needed for one wash while increasing the total weight of the laundry powder in the package. In this way, more ecology- and consumer-friendly decisions can be made by manufacturers of laundry detergents.

Linking farmers to markets: Assessing small-scale farmers' preferences for an official phytosanitary regime in the Kyrgyz Republic

Small-scale farmers are integral to Kyrgyz agriculture, contending with challenges such as low productivity, market entry barriers, and financial constraints. Most of these farmers are hindered by phytosanitary restrictions preventing direct exports of their products. Granting smallholders direct access to regional markets could significantly enhance their economic prospects. However, the specific policy measures that can encourage smallholders to participate in the official phytosanitary regime remain unclear. This study employs a randomized conjoint experiment, creating a hypothetical policy to measure the preferences of Kyrgyz fruit farmers regarding an official phytosanitary regime. Presenting choice sets with two randomly generated policy alternatives and an opt-out option, we collect data from 378 farmers across five villages. Each policy alternative comprises six attributes: (1) type of service provision, (2) provision of production inputs, (3) provision of interest-free loans, (4) provision of market information, (5) provision of phytosanitary inspection frequency, and (6) annual fees. The results reveal that attributes such as market information, interest-free loans, and production inputs—particularly virus-free seedlings—significantly increase farmers’ participation probability in systematization. However, farmers prefer the private sector as service provider over the government. Additionally, higher annual fees negatively impact the favorability of systematization. Overall, farmers express a preference for the hypothetical policy and are willing to pay at least 14,000 Kyrgyz som to participate in systematization. This study addresses the practical challenges faced by Kyrgyz farmers in meeting stringent phytosanitary standards, indicating the potential of a systematization policy to enhance market access. It lays the groundwork for future research investigating the economic impacts of specific policy attributes for Kyrgyz farmers.

Relationship between delay discounting, delay aversion and psychosocial domains of diabetes care

PurposeDelay discounting and aversion are important areas for diabetes management; however, little has been done to understand the relationship with psychosocial outcomes among adults with type 2 diabetes. MethodsThis study used data from 365 adults with type 2 diabetes to evaluate relationships between delay discounting and aversion and psychosocial outcomes. Delay discounting and aversion were measured with the validated Quick Delay Questionnaire. Psychosocial outcomes included depression, measured by the PHQ, anxiety by the GAD scale, perceived stress by the PSS, and social support by the Duke Social Support and Stress Scale. Multiple linear regression was used to assess the relationship between delay discounting and aversion on psychological health and social support controlling for relevant covariates. ResultsMean age of the sample was 61.8 years, 54.5 % were NHB, 41.8 % NHW, and 3.7 % Hispanic/Other. After adjusting for covariates, delay aversion was significantly associated with depression (beta = 0.35; p < 0.001), anxiety (beta = 0.52; p < 0.001), perceived stress (beta = 0.22; p < 0.001), and lower family support (beta = −0.62; p < 0.05). Delay discounting was significantly associated with depression (beta = 0.32; p < 0.001), anxiety (beta = 0.46; p < 0.001), and perceived stress (beta = 0.26; p < 0.001). LimitationsThis data is cross-sectional, future work should examine the longitudinal relationship while also including additional psychosocial outcomes. ConclusionsDelay discounting and aversion are significantly associated with poor psychosocial outcomes, including lower social support. As the body of evidence grows, additional research is needed to better understand the construct, mechanisms, and the impact of choice settings to better inform intervention development.

Alternate closed-form weibit-based model for assessing travel choice with an oddball alternative

Herein, a weibit-based model is proposed as an alternative to Recker's logit choice model with an “oddball” alternative, which explicitly focuses on a single alternative that has unique attributes to other conventional alternatives in the choice set [Recker, W.W. (1995). Discrete choice with an oddball alternative. Transportation Research, 29B, 201–212]. While retaining the closed-form probability expression, the proposed model handles the oddball alternative using a multiplicative random disutility function assuming Weibull distributed random components. The proposed model thus allows disutility-dependent perception variances of both the conventional and oddball alternatives and a flexible variance ratio between them, which are effective to reflect the way how individuals perceive varying magnitudes of travel disutility from different attributes. This gives the proposed model more flexibility to consider various heterogeneity issues, including the heterogeneous perceptions of conventional alternatives, heterogeneous perceptions of unique and common attributes of the oddball alternative, and heterogeneous features of conventional and oddball alternatives. The empirical application of the proposed model is explored to highlight its practical performance. The results demonstrate the superiority of the proposed model in the choice contexts where both oddball effect and heterogeneity issues need to be considered. The proposed model could further provide new behavioral insights into various decision-making scenarios of transportation networks, such as transportation mode choices in the current era of emerging technologies and destination choices in urban agglomerations.

A Preference-Based Value Assessment of the Fear of COVID-19 Contagion.

To assess the preference-based value of the fear of COVID-19 contagion. We conducted a web-based, cross-sectional discrete choice experiment among 544 US adults. We used a Bayesian efficient design to generate choice sets. Each choice set comprised two hypothetical COVID-19 vaccine options characterized by seven attributes: chance of COVID-19 infection, chance of having severe symptoms from COVID-19 infection, vaccine protection duration, chance of mild to moderate adverse events from vaccination, chance of serious adverse events from vaccination, chance of future exposure to COVID-19 after vaccination, and out-of-pocket cost. We used mixed logit (ML) and latent class (LC) models to analyze data. Furthermore, we calculated the willingness-to-pay for eliminating the chance of future exposure to COVID-19, shedding light on the value attributed to the fear of contagion. The ML model demonstrated all attributes, including the chance of future exposure to COVID-19, were statistically significant. The participants were willing to pay approximately $13,046 to eliminate the chance of future exposure to COVID-19 or their fear of contagion when COVID-19 was still pandemic. The LC model unveiled two participant classes with distinct preference weights for the chance of future exposure to COVID-19 and out-of-pocket cost attributes. Nevertheless, the chance of future exposure to COVID-19 exposure held a significant degree of importance in both classes. The chance of future exposure to COVID-19 exposure or fear of contagion was a significant element in the value assessment of COVID-19 vaccines. Further studies should be conducted to verify the value of fear of contagion and include it in the value assessment of healthcare technologies for infectious diseases.

Data management and techniques for best–worst discrete choice experiments

In this article, we present software that is suitable for use with Stata’s choice modeling suite of commands, which begin with cm. Within the context of choice models, we focus on best–worst data. In such data, respondents are presented a set of choices and are required to select a best and a worst choice from among the alternatives. Optionally, respondents may indicate an opt-out choice, in which no best or worst choice exists in the choice set. Such data are simplified versions of experiments in which respondents rank all the choices. Once best–worst data are collected, there are specific types of data expansions that analysts use to take advantage of both explicit and implicit information. The commands described in this article support data expansion and model estimation.

Patient message preferences to promote clinical conversations about chronic obstructive pulmonary disease (COPD): A discrete choice experiment.

The purpose of this study was to identify message features that motivate patients to initiate clinical conversations about Chronic Obstructive Pulmonary Disease (COPD). A secondary aim was to determine whether preferred message features vary by socio-demographic and behavioral characteristics. A discrete choice experiment was conducted in August 2020. Participants were asked to select which messages would motivate them to speak with a clinician about COPD. This included selecting messages across 8 choice sets, or a systematic combination of messages reflecting 6 attributes (e.g., susceptibility, call-to-action, emotion-frame, efficacy, message source, organizational support). The final sample was 928, which included adults (M=62.07; SD=10.14years old) who identified as non-Hispanic, white, and with at least some college experience. Message attributes ranked from most to least important were COPD susceptibility (25.53% [95% CI=24.39, 26.66]), message source (19.32% [95% CI=18.41-20.24]), COPD organization logo (19.13%; [95% CI=18.26, 20.01]), call-to-action (14.12%; [95% CI=13.40, 14.85], emotion-frame (13.24% [95% CI=12.55-13.94]), and efficacy (8.65%; [95% CI=8.20-9.09]). Participants preferred susceptibility messages about COPD signs/symptoms rather than risk behaviors related to smoking tobacco and environmental exposures. They also preferred messages from medical authorities (i.e., clinicians, COPD organization), a call-to-action that supports their autonomy in screening decision-making, and a message that conveys hope for living a healthy life with COPD and builds their self-efficacy to get screened. Differences in message preferences were detected according to age, gender, race, ethnicity, education level, and current vs. former smoking status. This study identified message features that motivate clinical conversations about COPD, especially those from subgroups who are disproportionately at-risk for its late-stage diagnosis.

OP24 Preferences Of Depressed And Depression-Prone Groups With Regard To Antidepressants In China: A Best-Worst Scaling Survey

IntroductionAntidepressants are one of the main treatment approaches for depression, and previous evidence suggests that consideration of patient preferences can improve their adherence to medication regimens. The objective was, therefore, to evaluate the preferences of depressed and depression-prone groups in China with respect to antidepressant medications.MethodsAn online survey with best-worst scaling choices was administered in depressed and depression-prone patients. The balanced independent block design generated 13 choice task profiles for each participant to answer, with each choice set consisting of four alternatives out of 13 antidepressant-specific attributes. Count analysis and a conditional logit model were used to estimate the relative importance of the 13 attributes and preference heterogeneity.ResultsThe analytical sample included 210 participants, comprising 49 individuals who had previous experience with depression and 161 who were depression prone. Participants in both groups preferred medications with a low risk of liver or kidney damage, headache or dizziness, and recurrence. There were significant differences in both groups regarding out-of-pocket costs and duration of medication. The K-means clustering further proved preference heterogeneity among the patients.ConclusionsOur study revealed patient preferences for antidepressant medication choices in China. Healthcare decision makers should consider and discuss patient preferences in the treatment decision-making process to improve patient adherence to and satisfaction with medications, and to ultimately improve patient outcomes.

OP60 Community Funding Prioritization Of Drug Treatments: A Discrete Choice Experiment

IntroductionAllocating government resources for drug treatments continues to be a challenge in health care, particularly given the increasing number of high-cost personalized drug treatments, finite resources, and aging population. Since taxpayers fund government health budgets, it is important to understand how they think funding should be distributed, considering attributes such as cost of drug treatments, risk of dying, commonality of medical conditions, and quality of life changes with drug treatments. The aim of this study was to understand what attributes of a medical condition and treatment determine a community’s willingness to fund new drug treatments. Two decision-making contexts were explored: the allocation of funds from a health budget and a willingness-to-pay (WTP) perspective.MethodsA representative sample of 500 Australian adults completed an online survey. The survey comprised two discrete choice experiments (DCEs) with different framing: an allocation of government funds and a WTP for drug treatments. The government allocation DCE allowed a choice between two hypothetical alternatives, each describing the medical condition and the drug treatment, while the WTP for funding showed one hypothetical alternative with an option to not fund the drug treatment. Seven DCE attributes, informed by a literature review, were displayed in each choice set relating to the medical condition (risk of dying, prevalence, and ages affected) and to the drug treatment (change in quality of life, additional life-years, availability of other drug treatments, and cost to the government and the taxpayer).ResultsThe resulting DCE model will establish the funding prioritization choices made by the general community. We expect changes in quality of life and risk of dying to be very important attributes. Of key interest for this study is how the difference in the decision-making context impacts preferences. We expect taxpayers to employ a stricter rule set in funding decisions when paying out of pocket (WTP exercise).ConclusionsThe findings from this research have implications for decision makers when aligning funding decisions with community preferences and values.

Comparative Analysis of Discrete Choice Approaches for Modeling Destination Choices of Urban Home-Based Trips to Work

Discrete choice models based on the Random Utility Maximization (RUM) approach are alternatives to traditional aggregated trip distribution models. However, the Random Regret Minimization (RRM) approach has emerged as an alternative to RUM for analyzing behavioral phenomena. This study compares the RUM and RRM approaches to model the destination choice of urban home-based trips to work, with a case study in the city of São Paulo, Brazil. The attributes of distance between traffic analysis zones (TAZs), total number of jobs, and number of jobs per type (industry and services) were considered, and the attributes referring to TAZs in commercial areas were tested in different models. We combined a method to estimate the optimal choice set size and a stratified sampling method to choose alternatives from available choices for a destination-choice problem with many alternatives. The RUM and RRM models produced similar results with respect to the final log-likelihood, [Formula: see text], Akaike information criterion, and Bayesian information criterion for the studied case. However, the RRM models required 52 min to be estimated on average, whereas estimation of the RUM models took only 31 s on average. Models derived from the RUM approach including the total number of jobs and distance produced higher accuracy, requiring less computational time to estimate the parameters.

Patient preferences, referral process, and access to specialized care. Is patient choice constrained?

ABSTRACT In most developed countries, patients have been encouraged to elect their preferred choice of health care provider. However, this is different for specialized care, where the patient’s referral could be defined as a two-stage decision process and their options are pre-selected by their general practitioner (GP). In this study, we estimate patient preferences while controlling for the pre-selection procedure, and we investigate whether patients are actively choosing their provider for cancer care. The French national hospital discharge database (Programme de Médicalisation des Systèmes d’Information, PMSI – MCO 2017) has been used for investigation. We estimated a multinomial choice model when choice sets are in fact unobserved, which is assumed to identify patient preferences, in a revealed preferences framework. Our findings provide evidence that patients consider factors other than distance to select their provider. The patient – hospital distance as well as the specialization profile of providers appears to be internalized in the pre-selection process, while patients rather consider waiting times, hospital quality, and other provider attributes to make their final choice. We also found that patients would be treated in higher-quality hospitals if they had the opportunity to choose among all available providers.

How crowding impedance affected travellers on public transport in the COVID-19 pandemic

In the aftermath of the COVID-19 pandemic, travel behaviour has changed significantly. Governors have introduced different transport policies to maintain the travel demand in the public transport system. Previous studies have developed the measurement of crowding impedance on public transport to determine the degree of transit use and the impact of the COVID-19 pandemic. This study explores the behavioural differences in crowding impedance to provide transport policies incorporating group segmentations. The D-efficient design process has structured a survey with a reasonable choice set, and questionnaires have been provided to identify the attitudinal groups of travellers effectively. The travellers are divided into four groups according to the values of factor loadings from the factor analysis: i.e., fear of disease, transit preference, time sensitivity, and auto preference. Multinomial logit models explore the behavioural differences in route and mode choices and calculate crowding multipliers. The results show that the group with a fear of disease comprises a high proportion of the elderly owing to their reluctance to expose themselves to infectious diseases; furthermore, the time-sensitive group exhibits less crowding impedance on public transport. Thus, the crowding multipliers differ between the groups and influence the relevant transport policies to promote public transport use. Policymakers are encouraged to introduce customized transport policies depending on the requirements of each group of travellers to cope with the adverse effects of the pandemic.

Caffeine’s complex influence on the attraction effect: a mixed bag of outcomes

Psychological state alterations induced by substance-related physiological mechanisms affect consumer decision-making. We examine the influence of caffeine—the world’s most popular psychostimulant—on the attraction effect. In three double-blinded experiments, we show that caffeine intake via coffee influences consumers’ preference for product options that asymmetrically dominate a decoy option in choice sets (i.e., the attraction effect). Using real products in consequential choice tasks, we show that high caffeine intake (200 mg) is associated with a larger attraction effect both on between-subjects and within-subjects levels and in free-choice as well as forced-choice decision tasks. On the contrary, we do not find support for caffeine’s influence on the attraction effect when considering intermediate levels of caffeine intake (125 mg) and hypothetical decisions. We discuss theoretical implications for context effect research and practical implications for marketers.