Predicting Field Experiment Results in a Lab
|Autor:||Mgr. Kateřina Chadimová|
|Rok:||2017 - letní|
|Vedoucí:|| PhDr. Lubomír Cingl Ph.D.
|Typ práce:|| Diplomová
|Abstrakt:||This thesis is aimed at forecasting of experimental results in a lab environment,
investigating often discussed external validity of laboratory experiments. We
run a novel laboratory experiment in which the subject pool is asked to make
predictions on results of a certain field experiment. The collected data is analyzed
using different accuracy measures, arriving at several interesting results.
First, the forecast among the 94 subjects is quite informative about the actual
treatment effects although its accuracy substantially varies based on a type of
accuracy measure and a particular treatment. Second, the average forecast is
either more accurate or at least comparable to the mean individual forecast,
proving the presence of “wisdom-of-crowds” effect.