CBC Hierarchical Bayes

Overview

CBC/HB is a standalone tool for estimating individual-level results for CBC experiments or for discrete choice data not designed or collected using Sawtooth Software’s tools. CBC/HB is included with your purchase of CBC for Lighthouse Studio 9 and later. 

CBC/HB System

• Up to 1000 attributes or parameters per individual
• Up to 1000 alternatives per choice task
• Maximum of 1000 levels for any one attribute
• Up to 1000 tasks per respondent
• Effects-coding, dummy-coding, linear coding, or user-specified coding
• Can impose utility (monotonicity) constraints
• Can include first-order interactions
• Supports chip-allocation, best-worst CBC, MaxDiff, dual-response None
• CBC/HB is a component of Lighthouse Studio which requires Windows XP or later to run.

Methodology

Researchers and academics have argued that respondents are unique. Latent Class utility estimation takes an important step toward recognizing respondent heterogeneity, but stops short of achieving strong individual-level predictions.

The generally preferred method for analyzing CBC data is Hierarchical Bayes (HB) estimation. Having individual-level estimates improves the accuracy of market simulations and leads to better understanding of market structure and attribute importances than aggregate logit modeling.

The CBC/HB module leverages information from all respondents to estimate results for each individual. The individual-level utilities are estimated by a statistical simulation technique called Gibbs Sampling. 

HB generally improves predictive validity relative to aggregate logit and Latent Class. A long-standing complaint against aggregate logit has been its IIA assumption, often referred to as the red bus/blue bus problem. Very similar products tend to capture too much net share in aggregate logit market simulations. With individual-level modeling, this effect is less problematic. 

The CBC/HB standalone features estimation of individual-level utilites or linear functions (main effects and/or first-order interactions) for standard CBC, best-worst CBC, or allocation-based (constant-sum) CBC questionnaires. Typical run times for market research data sets are from 1 to 20 minutes. CBC/HB reads and writes data to/from text-only files. You do not need to have used CBC to collect the data as long as you arrange the data in the CBC-compatible .CSV format. The utilities from CBC/HB can be used within Sawtooth Software's market simulators.

The CBC/HB System benefits from a fast computer and plenty of available memory.