A new suite of free, open-source software tools aims to make an intimidating family of survey techniques far easier to teach and learn. The tools, described in a paper published in the journal Discover Data, take the form of graphical menus for the statistical programming language R and cover three widely used stated preference methods: dichotomous choice contingent valuation, discrete choice experiments, and three types of best–worst scaling. The central idea is simple but consequential for classrooms: students can run complete analyses through point-and-click dialogs instead of writing code, while the underlying R commands are still displayed for them to study.

Stated preference methods are survey techniques for eliciting how people value goods and services, including things that are never bought and sold in markets. They have been adopted internationally across marketing, transportation, health and healthcare, agriculture and food, and environmental research. The preference information they generate allows researchers to estimate the economic value of goods or services and to forecast consumer demand. Governments have used them to evaluate the environmental impacts of public policies or projects in the United States, the United Kingdom, and Japan, while companies have applied them in business contexts such as designing new products and services.

The author of the new paper, Hideo Aizaki, argues that despite this reach, students often struggle when these methods appear in lectures. Applying a stated preference method involves multiple steps, from designing choice sets to extracting preference information from estimated results, and each step draws on knowledge from a different discipline. Experimental design theory is needed to build choice sets, survey methodology to run the questionnaire, econometrics to analyze responses, and microeconomics to interpret the results. Users must understand all of these steps before they can apply the methods effectively, which raises the barrier for newcomers considerably.

Software compounds the problem. Many free and open-source packages for stated preference methods operate primarily through command-line interfaces, so students unfamiliar with programming face the psychological burden and time cost of learning the methods and coding skills simultaneously. The appropriate package often varies depending on the specific method and the steps involved, meaning students may need to master several tools. Existing graphical, menu-driven software for these methods exists but is typically commercial and proprietary, leaving a gap for a unified, free, non-programming platform suitable for undergraduate teaching.

The solution described in the paper is a series of five plugin packages for R Commander, a long-established graphical user interface for R that has been distributed from the Comprehensive R Archive Network since 2003. The five packages correspond to dichotomous choice contingent valuation, discrete choice experiments, and the three variants of best–worst scaling, known as Case 1, Case 2, and Case 3. All are freely available from CRAN under the GNU General Public License and install like any other add-in package. Once installed and loaded, each adds its own menu item to the R Commander menu bar.

Each package contains a set of dialogs that guide users through the full workflow of a stated preference study. One dialog generates an experimental design and converts it into choice sets according to conditions such as attributes, their levels, and the number of choice sets. Another converts those choice sets into a series of questions and displays them. A further dialog collects responses to the questions and exports them to a CSV file, while a separate dialog merges the choice sets and the responses into a single dataset for analysis. A model-fitting dialog then estimates a choice model, and three types of post-estimation dialogs calculate and display willingness to pay measures and confidence intervals, forecast choice probabilities, and compute preference shares for items.

The technical scope is deliberately limited to introductory-level techniques. Choice set design relies on orthogonal designs or balanced incomplete block designs, and the econometric analysis is restricted to the conditional logit model for discrete choice experiments and best–worst scaling, or the binary choice model and its expansions for contingent valuation. All models are estimated by maximum likelihood. The author argues this restriction is pedagogically defensible: because every model in the series uses maximum likelihood estimation, basic knowledge of that method is enough for students to grasp the statistical basis of the tools, whereas advanced choice models such as random parameters or latent class models would require estimation methods likely beyond undergraduate econometrics.

Several design features target the classroom directly. The response-collection dialog lets students answer the survey questions themselves, save their answers as CSV files, and then analyze their own pooled data in class after the lecturer merges the files. Dialogs for designing choice sets and creating analysis datasets can export their outputs to external files, so a course spread over two days can save a design on the first day and import it on the second. In the best–worst scaling packages, a slider, a table, and radio buttons automate the naming of response variables, reducing typing errors in datasets where the number of response variables tends to be large. Predefined variable names and classified variable lists in the model-fitting dialogs help prevent students from selecting the wrong variables.

Because the series depends on R Commander, students see the R code generated by each of their menu actions in the script and output windows. The author notes that this functionality helps novices understand how R code is written, and that it is particularly useful for learning R’s model formula notation, since the code for fitting a model appears on screen as soon as variables are selected through the dialog. The output produced through the menus is identical to what the corresponding R functions would produce in the console with equivalent settings. Existing R Commander tutorial materials, plus a freely available online manual for the new series published under a Creative Commons license, can further reduce the cost of preparing teaching materials.

The paper is candid about limitations. The available methods are introductory only, which may disappoint those seeking statistical efficiency-based design methods or models that capture heterogeneity in preferences among individuals. The series also cannot be used appropriately without understanding each method; it removes programming barriers but does not reduce the methodological knowledge required for proper use. Most notably, the author states that empirical evidence demonstrating that the series improves the teaching and learning of stated preference methods is lacking, and that its effectiveness has not been empirically tested, though it has proven useful in the author’s own undergraduate instruction of contingent valuation, discrete choice experiments, and Case 1 best–worst scaling. Future work, the paper concludes, should evaluate the series’ effects on student learning, develop free and open-source GUI tools for advanced users, and explore alternative teaching approaches suited to the AI era.

Subject of Research: A free R-based graphical user interface series for teaching introductory stated preference methods

Article Title: R-based graphical user interface series for teaching and learning introductory stated preference methods: dichotomous choice contingent valuation, discrete choice experiments, and best–worst scaling

Article References: Aizaki, H. (2026). R-based graphical user interface series for teaching and learning introductory stated preference methods: dichotomous choice contingent valuation, discrete choice experiments, and best–worst scaling. Discover Data, 4(1), Article 4. https://doi.org/10.1007/s44248-026-00099-x

Image Credits: AI Generated

DOI: 10.1007/s44248-026-00099-x

Keywords: stated preference methods, R Commander, contingent valuation, discrete choice experiments, best-worst scaling, graphical user interface, econometrics education, choice set design, open-source software, R programming, survey methods, teaching tools