Tversky and Kahneman (1974) - Anchoring Bias

Aim

Investigate how anchoring bias affects numerical estimation when solving a mathematical problem

Procedure

Sample: high school students

Randomly assigned to two conditions:

1. Ascending condition: estimated the value of 1x2x3x…x8 in 5 seconds

2. Descending conditions: estimated the value of 8x7x6x…x1 in 5 seconds

Researchers hypothesised that the first number in the sequence would serve as an anchor, influencing estimations

Findings

Mean estimate (ascending condition): 512

Mean estimate (descending condition): 2250

Actual value: 40,320

Conclusion

Anchoring bias affects decision-making by relying too heavily on the initial piece of information presented

Even in a short time frame, the first number influenced ppts' numerical estimates

Evaluation

(+) Reliability - used a standardised procedure

(+) High internal validity - clear cause-and-effect relationship between anchoring and estimation is established

(-) Low ecological validity - the task is artificial and doesn't reflect real-world decision-making

(-) Low internal validity - the usage of an independent measures design means that participant variability may have influenced results