Talk:Outcome bias

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Wiki Education Foundation-supported course assignment[edit]

This article was the subject of a Wiki Education Foundation-supported course assignment, between 1 September 2020 and 18 December 2020. Further details are available on the course page. Student editor(s): Farrell101.

Above undated message substituted from Template:Dashboard.wikiedu.org assignment by PrimeBOT (talk) 05:56, 17 January 2022 (UTC)[reply]

Utilitarian Ethics[edit]

If outcome bias was extrapolated to ethics (as with other cognitive biases, such as omission bias), would this not seriously hurt the case for a utilitarian system of morality? ~ Baker —The preceding unsigned comment was added by 65.0.246.3 (talk) 13:35, 4 February 2007 (UTC).[reply]

Historian's fallacy[edit]

The historian's fallacy seems very similar. --Mrwojo 06:54, 30 September 2007 (UTC)[reply]

Possible unattributed copying and/or non-existent references[edit]

I have reverted most of the changes in this series of edits by @LibertarianLiechtenstein, for multiple reasons:

  1. Most of the changes to the first paragraph appear very similar to this article on Investopedia. The first paragraphs of both are nearly identical, starting at "arises when a decision is based on the outcome..." and ending at "...does not involve the distortion of past events." This Wayback Machine archive from 2014 indicates it isn't a case of Investopedia copying Wikipedia.
  2. The changes to the sports paragraph replaced valid references and content with a dead source and a redundant statement.
    • More specifically, the "Gauriot19" reference was changed from being about a paper in the Review of Economics and Statistics to one in the Journal of Sports Science and Medicine (the primary author's last name stayed the same, but that might be a coincidence). The DOI and PMID for the new reference are both dead, the title is very unlikely to be correct or close to it if it is correct (Google searches for site:jssm.org "outcome bias" and site:jssm.org "outcome effect" return nothing), and the paper isn't in JSSM's archives assuming the year/volume and pages are correct. That reference has already been removed by another editor.
    • The final sentence of that paragraph was changed from "Another study found that professional basketball coaches are 'more likely to revise their strategy after a loss than a win... even when a loss was expected and even when failure is due to factors beyond the team's control.'" to "This is an example of how the outcome bias can cloud people's judgment and cause them to focus on the final result rather than the actual process and performance of the player.". In my opinion, the latter conveys no new information and is just filler.
  3. The new paragraphs added have reference problems as well. For the Annals of Internal Medicine reference, its DOI is dead and its PMID is unrelated. Google searches for site:acpjournals.org "outcome bias" and site:acpjournals.org "outcome effect" return only irrelevant results. The paper also isn't in the journal's archives. For the National Bureau of Economic Research reference, its URL is dead and weirdly formatted. The "w7721" part probably refers to a working paper, but the actual NBER paper of that number is irrelevant and has a different URL scheme. (There is a w7732 that is relevant, but it doesn't support the claims of the added paragraphs.) The "nber.org/digest/apr01/..." format is actually used for digest articles published in April 2001, but none of the articles in that digest are relevant.
  4. It is possible that the new content, from the sports paragraph onwards, was generated using a large language model and thus likely containing AI-generated "hallucinations". It would explain the seemingly non-existent references, as AIs tend to make up plausible-sounding fake references when prompted to cite their sources, but it would also make all of that information completely unverifiable. I used every detector mentioned in this Signpost article (except for OpenAI's one, because it requires an account), and all of them indicated the added text was likely AI generated (except Content at Scale, which indicated around 50-50).

I am not certain that these changes are definitely plagiarism or fabrications, unintentional or not. LibertarianLiechtenstein's response would be helpful, since they were the editor who made these changes. If there's something I've overlooked, please tell me. Coolclawcat (talk) 07:53, 20 June 2023 (UTC)[reply]