Conjoint Study Design Generator
Create optimized fractional factorial designs for choice-based conjoint (CBC) studies. Define attributes and levels, specify constraints, and generate balanced experimental designs ready for data collection. Uses Latin Hypercube Sampling and D-efficiency optimization for commercial-grade study designs.
DESIGN METHODOLOGY
What is Fractional Factorial Design?
A fractional factorial design strategically selects a subset of all possible product configurations to test. Instead of showing respondents every combination (which could be thousands of scenarios), we use statistical design principles to create an efficient set that still allows us to estimate all attribute effects.
Key principles: Orthogonality (attributes vary independently), balance (each level appears equally often), and minimal overlap (reduce repeated alternatives within tasks).
Design Quality Metrics
- D-Efficiency: 0-100% score measuring design optimality (85%+ = commercial-grade)
- Level Balance: Each attribute level appears proportionally across profiles
- Orthogonality: Attributes vary independently to prevent confounding
- Space Coverage: Latin Hypercube ensures profiles span attribute space
Typical Parameters
Tasks per Respondent: 8-15 scenarios (balance data quality vs. fatigue)
Alternatives per Task: 3-5 products (cognitive load constraint)
Attributes: 3-8 features (more attributes = larger sample needed)
Levels per Attribute: 2-6 values (avoid single-level attributes)
None Option: Recommended for realistic choice modeling
STEP 1: DEFINE ATTRIBUTES & LEVELS
Add 3-8 attributes. Each attribute should have 2-6 levels.