In Simple Words
Imagine evaluating four houses in a row; if viewing an expensive mansion first causes someone to bid higher on a small cottage afterwards, the valuation tool is broken, but if the bidding price for the cottage remains identical regardless of house order, the measurement tool is robust.
Farmers valuing the same rainfall policy
- Actual policy shown first
- Mean bid Rs.69.0
- Actual policy shown last
- Mean bid Rs.67.9
Randomized presentation order in 2010. No statistically significant difference; this does not prove complete order independence. Other contracts omitted.
See source →Research question
Valuation elicitations using multi-contract BDM auctions in field settings risk order-dependent anchoring biases, where initial contract presentations artificially anchor subsequent willingness-to-pay bids.
Participants
N = 1,978 smallholder farmers in rural Andhra Pradesh, India.
Study Design
Randomized contract presentation sequence (Ordering 1: Real, Exit, mm deficit, Basis risk vs. Ordering 2: Basis risk, mm deficit, Exit, Real) in an incentive-compatible BDM field game.
Results
Mean willingness-to-pay bids were virtually identical across orderings (Real policy: Rs. 69.0 in Ordering 1 vs. Rs. 67.9 in Ordering 2; Basis risk policy: Rs. 38.9 vs. Rs. 39.1), confirming complete order independence.
Takeaway
Randomize and test presentation sequences when deploying incentive-compatible BDM auctions to ensure contract valuation measurements are free from cognitive anchoring biases.