Some probabilistic models of best, worst, and best–worst choices

By A. A. J. Marley, Jordan J. Louviere. 2005

711 citations recorded by the source corpus.

Subfields: Economics and Econometrics, General Decision Sciences, Management Science and Operations Research

View Paper: https://www.semanticscholar.org/paper/ebe0f6b64669c1374b5ac28403bac98452900467

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Economics and Econometrics

In General Decision Sciences

In Management Science and Operations Research

Citations in the corpus

Contemporary Guidance for Stated Preference Studies (2017, background): “CEs have included various types of binary andmultinomial choice formats, along with some types of ranking and best/worst scaling (Marley and Louviere 2005; Flynn 2010; Scarpa et al. 2011), although there has been some inconsistency in the literature regarding which elicitation formats are included…”
Comparison of best–worst and hedonic scaling for the measurement of consumer wine preferences (2009, methodology): “The statistical model underlying BWS assumes that the relative choice probability of a given pair is proportional to the distance between the two stimuli on the latent utility scale (Marley and Louviere 2005).”
Why school students choose and reject science: a study of the factors that students consider when selecting subjects (2017, methodology): “The BWS method has a strong mathematical basis and is well validated (Louviere, Lings, Islam, Gudergan, & Flynn, 2013; Marley & Louviere, 2005).”
Random regret minimization or random utility maximization: An exploratory analysis in the context of automobile fuel choice (2013, background): “…by Quiggin with regret and rejoice has similar properties to the best-worse (BW) processing rule that focuses on contrasts between alternatives (see Marley and Louviere 2005) ( )( )ln 1 expmi j m jm imR x xβ↔ = + ⋅ − In combination, this implies the following formulation for systematic regret:…”
Consumer preferences of wine in Italy applying best‐worst scaling (2009, methodology): “A solution to these problems has been given by Finn and Louviere (1992) through the so called Best:Worst (BW) method, which proved to be empirically successful, but it was not until 2005 that Marley and Louviere gave a rigorous methodological explanation for it (Marley and Louviere 2005).”
Models of best–worst choice and ranking among multiattribute options (profiles) (2012, methodology): “However, when the preference independent MA maxdiff model is developed in a random utility framework (see Marley & Louviere, 2005, for the maxdiff model in that framework), it is customary to work with sums (or weighted sums) of scales, and the scales are often (erroneously) considered to be…”
Changes in Personal Values in Pandemic Times (2021, methodology): “We computed relative importance scores for each value item for each respondent by subtracting the number of times an item was chosen as least important from the number of times the item was chosen as most important (Marley & Louviere, 2005).”
Probabilistic models of set-dependent and attribute-level best-worst choice (2008, methodology): “The latter form is the one that is often used when the model is derived as a random utility model (see Coast et al, 2006; Marley & Louviere, 2005).”
What You Don’t Know Might Hurt You: Some Unresolved Issues in the Design and Analysis of Discrete Choice Experiments (2006, background): “Marley and Louviere (2005, forthcoming) show how different ways of making such choices imply different models, and each model has different measurement properties, and in some cases, different degrees of bias.”
Conducting Discrete Choice Experiments to Inform Healthcare Decision Making (2012, methodology): “The underlying Theory and methods have been developed to allow theoretical properties of BWS were formally proven one to model the choices of single individuals, by Marley and Louviere[97] and Marley et al.”
An international comparison of retail consumer wine choice (2009, methodology): “4) and multiply the result by 100 (for the convenience of presentation). There is much more statistical rigour to the process, compared to the standard Likert-type scale. For a detailed discussion of the method and its application in the wine sector, see Cohen (2009) and Goodman et al.”