Real-world Estimation: Estimation Modes and Seeding Effects

More people live in Ethiopia (64 million) than in the Untied Kingdom (59 million). Located at 34 ° north latitude, Atlanta, Georgia is just one degree north of Tijuana Mexico. At the moment, Bill Gates is worth about $64 billion; the GDP of Tunisia (with its population of 9.7 million people) was $63 billion in 2000. Montreal is about 2900 kilometers from Edmonton, Alberta; the distance between Edmonton and the city of Chihuahua in Mexico is about 2800 kilometers. At $48,000 a new Honda $2000 cost more than a new BMW Z3 sports car which goes for $45,000, etc, etc. To many people, these facts taken separately, or in tandem, are surprising, even counterintuitive. From a cognitive perspective, this is an interesting response because it suggests that people do have intuitions about many different real-world quantities and that these intuitions can be very wrong. At the same time, exposure to numerical facts like these can be highly informative. This latter point has been made a number of times using a method called seeding the knowledge-base (Bostrom & Brown, 2001; Brown, 2001; Brown, Friedman, & Lee, 2001; Brown & Siegler, 1993, 1996, 2001; Friedman & Brown, 2000a,b; Friedman, Kerkman, & Brown, in press-b; LaVoie, Bourne & Healy, in press; Murray & Brown, 2001; Walbanm, 1997). In the usual seeding experiment, participants first provide numerical estimates for a set of items; they then learn the actual values of a subset