Standard Conjoint Research Articles

Estimating and Using Individual Marginal Component Effects from Conjoint Experiments

AbstractConjoint experiments are quickly gaining popularity as a vehicle for studying multidimensional political preferences. A common way to explore heterogeneity of preferences estimated with conjoint experiments is by estimating average marginal component effects across subgroups. However, this method does not give the researcher the full access to the variation of preferences in the studied populations, as that would require estimating effects on the individual level. Currently, there is no accepted technique to obtain estimates of individual-level preferences from conjoint experiments. The present paper addresses this gap by proposing a procedure to estimate individual preferences as respondent-specific marginal component effects. The proposed strategy does not require any additional assumptions compared to the standard conjoint analysis, although some changes to the task design are recommended. Methods to account for uncertainty in resulting estimates are also discussed. Using the proposed procedure, I partially replicate a conjoint experiment on immigrant admission with recommended design adjustments. Then, I demonstrate how individual marginal component effects can be used to explore distributions of preferences, intercorrelations between different preference dimensions, and relationships of preferences to other variables of interest.

A choice experiment to identify the most important elements of a successful cancer screening program according to those who research and manage such programs.

The relationship between cancer screening activities in Europe and the health systems in which they are embedded varies, with some screening programs organized largely separately and others using existing health service staff and facilities. Whatever the precise arrangements, the opportunity for screening to achieve health gain depends on many elements interacting within and beyond the health system, from an accurate register identifying the target population to a means to ensure and monitor follow-up. A conjoint analysis was undertaken with 66 cancer screening experts from 31 countries taking part in EU-TOPIA (towards improved screening for breast, cervical and colorectal cancer in all of Europe) to identify priorities for an effective screening program, taking a whole system perspective. Ten attributes, each with two levels, were derived from a review of the literature and consultation with experts in cancer screening. Statistical software generated 12 profiles that were ranked by respondents and analyzed using standard conjoint analysis. The most important attributes were having up-to-date and evidence-based guidelines, followed by mechanisms for systematic monitoring of screening uptake, having a population register covering all of the eligible population and monitoring long-term outcomes. In discussions about the results, participants argued that quality assurance and adherence to guidelines were important, even though they generated low scores in the experiment. This difference may be due some attributes being interrelated, more wide-ranging or the sequential nature of establishing an effective screening program, with guidelines being the first stage of the process.

Can Conjoint Analysis Mitigate Social Desirability Bias?

How can we elicit truthful responses in surveys? Political scientists are often concerned about systematic survey misreporting on sensitive topics, known as social desirability bias (SDB). Conjoint analysis has become a popular tool to address this concern, despite the lack of systematic evidence showing its suitability for this purpose. We employ a novel experimental design to identify a fully randomized conjoint design's ability to mitigate SDB. Specifically, we compare a standard conjoint design against a partially randomized design where only the sensitive attribute is varied between two profiles in each task. Our design also includes a placebo condition designed to remove confounding due to the increased attention to the varying attribute under the partially randomized design. We implemented this experiment in a survey to uncover respondents' attitudes toward environmental conservation. We find suggestive evidence that conjoint analysis does, indeed, mitigate SDB.

Benefit-Based Conjoint Analysis

Firms develop products by manipulating the attributes of offerings, and consumers derive utility from the benefits that the attributes afford. While the field of marketing has long been aware of the distinction between attributes and benefits, it has not developed methods for understanding how attributes and benefits are related. This paper develops a benefit-based model for conjoint analysis that assumes consumers satiate on attributes that are perceived to provide the same benefit. A latent-variable model is proposed that estimates the map between attributes and benefits, and is applied to data from two conjoint studies involving a durable product and a household consumable. The model is shown to fit the data better, provide improved predictions, and lead to different product design implications than the standard conjoint model.Data, as supplemental material, are available at http://dx.doi.org/10.1287/mksc.2016.1003 .

Benefit-Based Conjoint Analysis

Firms develop products by manipulating the attributes of offerings, and consumers derive utility from the benefits that the attributes afford. While the field of marketing has long been aware of the distinction between attributes and benefits, it has not developed methods for understanding how attributes and benefits are related. This paper develops a benefit-based model for conjoint analysis that assumes consumers satiate on attributes that are perceived to provide the same benefit. A latent-variable model is proposed that estimates the map between attributes and benefits, and is applied to conjoint data from an industry study involving a household consumable. The model is shown to fit the data better, provide improved predictions, and lead to different product design implications than the standard conjoint model.

Making Conjoint Behavioural

Traditional choice-based conjoint methods are based on an unrealistically rational model of consumer decision-making. These methods work accurately only if we assume that consumers can process all the information given to them, weigh it up and make a calculated, accurate decision. Modern discoveries in behavioural economics make it clear that these assumptions are incorrect. To accurately understand consumers’ decisions and preferences, conjoint methods must be updated to include behavioural understanding. This paper presents five ways in which this can be done: rank-finding conjoint, goal-attribute conjoint, intangible-attribute conjoint, algorithmic conjoint and contextual conjoint. Each of these extensions to the standard conjoint method can explore a specific aspect of the decision-maker's psychology, and together they result in a much deeper and more accurate reading of consumer behaviour and desires.

Compensatory versus noncompensatory models for predicting consumer preferences

AbstractStandard preference models in consumer research assume that people weigh and add all attributes of the available options to derive a decision, while there is growing evidence for the use of simplifying heuristics. Recently, a greedoid algorithm has been developed (Yee, Dahan, Hauser & Orlin, 2007; Kohli & Jedidi, 2007) to model lexicographic heuristics from preference data. We compare predictive accuracies of the greedoid approach and standard conjoint analysis in an online study with a rating and a ranking task. The lexicographic model derived from the greedoid algorithm was better at predicting ranking compared to rating data, but overall, it achieved lower predictive accuracy for hold-out data than the compensatory model estimated by conjoint analysis. However, a considerable minority of participants was better predicted by lexicographic strategies. We conclude that the new algorithm will not replace standard tools for analyzing preferences, but can boost the study of situational and individual differences in preferential choice processes.

An Incentive-Aligned Mechanism for Conjoint Analysis

This article specifies, analyzes, and validates a rigorous and practical truth-telling mechanism (game) for conjoint applications. The mechanism requires only one real product variation and has truth telling in conjoint as its Bayesian Nash equilibrium, thus making it possible to incentive align participants in most conjoint applications. Using the iPod package as the context, the author shows empirically that the mechanism substantially improves purchase prediction compared with a standard conjoint procedure.

Context Effects in the Measurement of Subjective Workload: Prolonged Exposure to the Same or Different Contexts

The present study investigated the effect of context on the subjective measurement of workload using the continuous memory task. Twelve levels of task difficulty were manipulated either between-subjects or within-subjects. Participants performed the continuous memory task for 48 trials under either a single difficulty level (between-subjects context) or under all 12 difficulty levels (within-subjects context). Workload was measured using ModSWAT— a technique that uses the sum of the SWAT dimensions to form the workload composite rather than the standard conjoint measurement procedure. Results indicated no difference in performance as a function of context but that context affected ratings of subjective workload. Under the single context condition, participants rated the workload of the easiest conditions higher than if the same conditions occurred within context of other more difficult conditions. Results are discussed in terms of their implication for practical assessment of subjective workload.

Effect of Pairted Comparison Weighting and Weighting Context

The present study compared three subjective workload measurement techniques—a unidimensional scale (UW) and two mutidimenional scales (TLX and ModSWAT)–using a continuous memory task. ModSWAT is a technique that uses the sum of the SWAT dimensions to form the workload composite rather than the standard conjoint measurement procedure. Both weighted and unweighted workload measures were obtained for the two multidimensional scales. The paired-comparison technique used by TLX was employed to weight the three SWAT dimensions. For both multidimensional scales, weighting context was also varied. Paired comparison of the workload dimensions was made either prior to the task in a general context or within context of the task just performed. The major findings were that: (1) ModSWAT and UW were more sensitive than TLX to the task demands (i.e., difficulty) of the continuous recognition task; (2) weighting context affected the weights that were assigned the workload dimensions for both TLX and ModSWAT; (3) weighting context did not affect the weighted workload scores; and (4) the weighted and unweighted workload composites were very similar. Results were interpreted within context of averaging across multiple dimensions when several dimensions are insensitive to the task demands. Variability in the weights assigned contributed to the failure to find a effect of weighting.