Conjoint survey designs are now common tools for studying decision-making, but a practical question arises for researchers: how many choice tasks can respondents do before satisficing meaningfully reduces response quality?
📋 How Respondents Were Tested
- Respondents were asked to complete as many as 30 conjoint choice tasks.
- Experiments were fielded on two common survey platforms: Amazon's Mechanical Turk (MTurk) and Survey Sampling International (SSI).
📊 What the Experiments Found
- Detectable increases in survey satisficing occurred as the number of tasks rose, but these increases were quite limited.
- Conjoint designs showed surprising robustness: even with dozens of tasks, response quality did not decline substantially in the study contexts examined.
🔎 Why This Matters
- The results provide practical guidance for survey designers weighing the trade-off between statistical power and respondent burden.
- In similar populations and contexts, researchers can assign multiple—potentially dozens of—conjoint tasks without expecting large losses in data quality, though monitoring satisficing indicators remains advisable.
