Join us for Sawtooth Software's Summer Training Event in the charming mountain town of Park City, Utah. This event features the following workshops: Choice Modeling Workshop (CBC/MaxDiff), Market Simulator A-Z, Turbo Choice Modeling, and MBC. Sign up to attend two or four days of this event today!Register Now
The summer training event will include two sessions. Each session has two courses to choose from. Attend one session for $1,500 or two sessions for $2,500.
Come and join us for a two-day Choice Modeling workshop taught by the experts at Sawtooth Software. This is an introductory workshop aimed at those completely new to choice modeling or with limited experience hoping to gain a deeper understanding. We’ll pull back the curtains and demystify how everything works from start to finish.
Participants should bring a Windows-capable laptop computer with them, as we’ll explore programming and analysis of both a Choice-Based Conjoint and MaxDiff project using Lighthouse Studio, our capable Windows-based survey platform. We will also touch a little bit on Adaptive CBC (ACBC) opportunities within Lighthouse Studio, as well as demo for you new developments for CBC and MaxDiff within our web-based Discover platform.
The Turbo Choice Modeling Event is an intensive two-day meeting covering advanced choice modeling topics. Six panelists will each present two or three papers with ample time built in for Q&A, for group discussions, and for meeting with the speakers and other attendees during breaks and meals. Here's a sampling of some of the speakers and some of the topics they'll be presenting:
Metropolis-within-Gibbs vs. HMC vs. Variational Bayes for estimating HBMNL models.
Imposing sign and ordering constraints on HBMNL part-worths estimated with HMC: a comparison of approaches.
Weights in HB Estimation
Modeling with the Posterior Distribution: Deriving Predictive Point Estimates via Empirical Bayes
A New Approach to Evoked Sets Using Continuous Approximations of Maximum Utilities
Beyond Serial Cross Effects: A Simultaneous Nested Logit Estimation of Menu Based Conjoint
Can We Do HB 'On the Fly'
Standard Errors and Confidence Intervals for HB Results
How sparse is too sparse? An investigation into the relationship between nK and accuracy within MaxDiff
Simulations to Compare Convergence of Sawtooth Software’s CBC/HB and Stan’s HMC
Does the NOL Effect Fade after 4 Levels of a Quantitative Attribute? (with Zachariah Hewett)
Prohibitions in CBC? Trading Off 1-way vs. 2-way Level Balance to Improve D-Efficiency
Highlights from the 2023 Analytics & Insights Summit
Three approaches to WTP
Dominance in CBC Surveys
How Many CBC Questions Per Respondent? A Parameter Recovery Experiment (with Christina Miller)
LC-MNL vs HB+Cluster for Choice-Model Segmentation: A Structure Recovery Experiment (with Cameron Halversen)
A CBC Study about Survey Taking
Excel tips and tricks useful when dealing with CBC data
The market simulator is where the results of a conjoint analysis come alive and we make predictions of choice (shares of preference summing to 100%) among competing product alternatives and the None option. Sawtooth Software's conjoint analysis market simulator for the desktop (running either in Lighthouse Studio or as a standalone Windows application) is very powerful, with numerous settings and options.
This thorough, hands-on workshop will spend two days covering all features and aspects of the "Choice Simulator" for conjoint analysis. It will be hands-on, with attendees doing exercises using practice datasets to gain experience with the simulator. In addition to basic simulations, we'll cover such things as sensitivity analysis, optimization search, reporting revenues and profits, exporting simulators to Excel, and adjustments to market simulations to account for incomplete product distribution and awareness. Participants should bring a Windows-capable laptop computer with them.
Real-world products often are not sold in the discrete choice CBC format (take one of multiple pre-configured product options) but rather allow the buyer to configure/customize their purchase via multiple selections and multiple choices. Obvious examples include menus at restaurants, many technology/software purchases, a la carte offerings, and mixed bundling cases.
In this 2-day session we’ll cover the theory behind MBC, then we’ll jump in and analyze at least six different MBC data sets together. These examples highlight different uses of MBC in a variety of contexts, including a mixed bundling case and Situational Choice Experiments (SCE). Once you learn what MBC can do, you’ll see that CBC is just a specific (discrete choice) case within a broader family of multi-check choice experiments.
This 2-day hands-on workshop is for advanced choice researchers who already have a firm foundation in CBC and who are experienced with econometric modeling, including coding of the design matrix (dummy coding) for logistic regression.