tag:blogger.com,1999:blog-4509736829271344872.post443680772421324363..comments2024-03-02T03:24:43.931-05:00Comments on Advanced Analytic Techniques: Bayesian Reasoning Method For Intelligence Using Natural FrequenciesKristan J. Wheatonhttp://www.blogger.com/profile/02566135545863154089noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-4509736829271344872.post-86440545280193855312014-10-06T10:54:50.432-04:002014-10-06T10:54:50.432-04:00Harrison,
All three experiments used Bayesian tec...Harrison,<br /><br />All three experiments used Bayesian techniques. In the first experiment consisting of two groups, the independent variable was the use of natural frequencies in the wording of the problem instead of traditional statistical language. The dependent variable was the number of participants who were able to answer the question correctly, both in the natural frequencies<br />group and the traditional statistical language group.<br /><br />The natural frequencies group outperformed the traditional frequencies group as shown in the picture.<br /><br />In experiments 2 and 3, the performance of participants using an experimental structured natural frequency tree method was compared to the performance of the natural frequency group in experiment 1.<br /><br />Participants in experiments 2 and 3 using the experimental structured natural frequency tree method did not perform better than the natural frequency group in experiment 1 at a statistically significant level. Performance was comparable. However, participants in experiments 2 and 3 reported an improved understanding of Bayesian reasoning at a statistically significant level.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-4509736829271344872.post-26587708802103213812014-10-05T23:17:03.160-04:002014-10-05T23:17:03.160-04:00Ricardo,
The experiment found that students using...Ricardo,<br /><br />The experiment found that students using Bayes did not have a significantly larger forecasting accuracy than those using the frequency tree method. Do you think this is a valid method that improves forecasting accuracy?Anonymoushttps://www.blogger.com/profile/11771375528076623770noreply@blogger.comtag:blogger.com,1999:blog-4509736829271344872.post-30798395611809577552014-10-05T22:13:59.118-04:002014-10-05T22:13:59.118-04:00Kyle,
Lee performed three different experiments. ...Kyle,<br /><br />Lee performed three different experiments. The first one tested performance on Bayesian problems that were worded using natural frequency compared to Bayesian problems worded using traditional statistical language.<br /><br />The second experiment tested an experimental Bayesian reasoning method using the natural frequency tree format.<br /><br />The third experiment tested the same experimental Bayesian reasoning method using the natural frequency tree format again to address concerns about the second experiment's methodology and results.<br /><br />In the first experiment Bayesian reasoning as described by Ulrich Hoffrage and Gerd Gigerenzer outperformed Bayesian reasoning using traditional statistical language.<br /><br />In experiments 2 and 3 the experimental natural frequency tree format failed to outperform the Bayesian reasoning as prescribed by Hoffrage and Gigerenzer. However the natural frequency tree format did not decrease forecasting accuracy. The natural frequency tree format just failed to yield better results than letting analysts use their own form of calculation to solve Bayesian problems like in the Hoffrage/Gigerenzer group from the first experiment.<br /><br />Appendix G of the article contains a step by step approach to the frequency tree method with visuals.<br /><br />Lee considered the possibility that the participants did not have the background to understand Bayesian, but discounted this hypothesis because out of the 66 combined freshman and sophomore participants in experiment 2, only 29 had not yet taken a statistics class.<br /><br />Neither experiments 2 or 3 found statistically significant improvements of forecasting accuracy utilizing the natural frequency tree method compared to simply wording Bayesian problems in natural frequencies.<br /><br />Participants did report a statistically significant improvement of their understanding of Bayesian after learning the natural frequency tree method so there was an auxiliary benefit. Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-4509736829271344872.post-41745318742048647472014-10-05T22:12:40.434-04:002014-10-05T22:12:40.434-04:00This comment has been removed by the author.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-4509736829271344872.post-20398090746500531582014-10-05T12:15:05.907-04:002014-10-05T12:15:05.907-04:00Even though Lee found that Bayesian is easy to tea...Even though Lee found that Bayesian is easy to teach, could it be possible that the students used in the experiment truly didn't understand Bayesian? That was quite a discrepancy between the two studies.Anonymoushttps://www.blogger.com/profile/13109479698281674218noreply@blogger.com