EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
3169. Accounting for heuristics in reputation systems. An interdisciplinary approach in aggregation processes
Invited abstract in session WC-40: Experimental economics and game theory 2, stream Experimental economics and game theory.
Wednesday, 12:30-14:00Room: 96 (building: 306)
Authors (first author is the speaker)
| 1. | René Fahr
|
| Management, Paderborn University | |
| 2. | Dirk van Straaten
|
| Department Management , Paderborn University | |
| 3. | Vitalik Melnikov
|
| Software Innovation Campus Paderborn, Paderborn University | |
| 4. | Eyke Hüllermeier
|
| Institut für Informatik, Ludwig-Maximilians Universität München | |
| 5. | Behnud Mir Djawadi
|
| Department Management , Paderborn University |
Abstract
Product ratings help customers to evaluate the quality of a product prior to their purchasing decision in online markets. To overcome information overload, the vast amount of reviews is aggregated to a single scalar value, typically the arithmetic mean. However, customers may not base their purchasing decision on the arithmetic mean but on other aggregation functions such as the amount of 1-star and/or 5-star ratings. Our research question therefore asks to what extent the arithmetic mean really reflects how customers aggregate product ratings. In our incentivized laboratory experiment, participants are shown bundles of different products, and subsequently, asked to rank them according to their preferences. Analyzing the rank data with a Placket-Luce model, we find that the arithmetic mean indeed best explains the obtained data. However, our analysis on an individual level reveals more heterogeneous aggregation patterns. In particular, a significant share of participants exhibits a binary bias. That is, moderate 2-star and 4-star categories get over- and the extreme 1-star and 5-star categories under-weighted. We also observe minor clusters who prefer products which have the lowest amount of 1-star ratings. The way our participants aggregate the ratings is neither affected by variation of provided information nor by individual characteristics such as experience, risk attitudes, or demographics.
Keywords
- Behavioural OR
- Management Information Systems
- Decision Support Systems
Status: accepted
Back to the list of papers