Conference Agenda

Choice experiments are a workhorse tool in environmental valuation, yet a fundamental design decision, the number of alternatives shown in each choice task, remains under-tested despite its potential to affect incentive properties, cognitive burden, and welfare estimates. We provide new evidence on choice-set size effects in public-good valuation by experimentally varying the number of non–status quo options from one to five in an otherwise identical choice experiment on Baltic Sea environmental improvements. To disentangle a mechanical “more opportunities to match” effect from genuine behavioural change, we propose a novel simulation-based test that uses the binary SQ-versus-one-alternative format as an incentive-compatible benchmark. Under preference stability, we simulate expected choice frequencies for larger choice sets and compare them with observed frequencies. The deviations increase systematically with choice-set size and are most pronounced for the cost attribute: high-cost alternatives are chosen substantially more often than predicted in the largest formats. These results indicate that expanding choice sets is not a neutral efficiency gain; beyond moderate sizes it can induce heuristics and materially bias welfare inference.