Premium
An experimental investigation of plural analysis: Estimating the market share of compact disc players in 1990
Author(s) -
Winterfeldt Detlof Von,
Eppel Thomas,
Ford Carla
Publication year - 1988
Publication title -
behavioral science
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.371
H-Index - 45
eISSN - 1099-1743
pISSN - 0005-7940
DOI - 10.1002/bs.3830330303
Subject(s) - confidence interval , econometrics , market share , statistics , market penetration , plural , mathematics , economics , engineering , linguistics , philosophy , finance , electrical engineering
This article describes an experiment on individual forecasting and estimation. It explores how people reconcile results from different models or analysis approaches that estimate the same variable. Seventy‐seven subjects were asked to estimate the market share of compact disc players in 1990. The estimates were made (a) intuitively, (b) using a market penetration model, and (c) using a scenario model. Subjects' uncertainty over the estimate was substantial, with a 90% credible interval covering a range of 50% of possible market shares. The average 90% credible intervals did not decrease significantly after the presentation of the market penetration model or the scenario analysis. However, the uncertainty ranges tended to increase with an increase in the degree of discrepancy between the initial estimate and the scenario estimates. No such effect was observed for the market penetration model. Independently of the discrepancy between subjects' estimates and the model estimates, the confidence in subjects' post‐model estimates always tended to increase. Subjects had more confidence in the scenario model than in the market penetration model and least confidence in their own initial guesses. When asked to reconcile the different estimates, subjects most often chose the scenario estimate or created a mixture.