Wednesday, May 1, 2013

Facts have a hard time swaying opinion

I am working on a project right now that looks at the durability of books as a proxy for the utility and durability of ideas, knowledge and influence. The supposition is that the epistemological integrity of information that is reported real-time is suspect but that it improves over time. Reports in the instant are highly inaccurate and unreliable, daily reports better, weekly and monthly better yet. By the time a book is written, months and years later, the factual information has been scrubbed and is more reliable. Not necessarily completely or even mostly reliable, just more reliable.

Megan McArdle has a couple of blog posts which, while not a book, illustrate this scrubbing process and provides a nice summary of the events thirteen years ago around the Florida recount. It also illustrates the durability of first impressions formed in the heat of the moment versus the factual reality based on later data. The accusation that the Florida vote was stolen has been long put to bed but I suspect that were you to poll a knowledgeable sample of the electorate, you would find a substantial portion, if not a majority, who believe that had not the Supreme Court issued its decision, that the outcome would have been different.

Following all the ebbs and flows at the time, and being particularly concerned about the nakedly distorting (from a statistical perspective) courses of action which were being put forward, I was eventually confident that the Supreme Court had made the right decision and that, after all the dust had settled, it would likely be found that Bush was indeed the winner, though with different numbers.

McArdle's posts affirm this; What if the Supreme Court Had Declined to Hear Bush v. Gore? and What if the Supreme Court Had Turned Down Bush v. Gore? All the details are in the posts, particularly the first one, but the upshot was that Bush would have won the recount under any requested recount procedure at the time. I somehow missed, or more likely have forgotten, the details of a newspaper consortium led research project: EXAMINING THE VOTE: THE OVERVIEW; Study of Disputed Florida Ballots Finds Justices Did Not Cast the Deciding Vote which verified that that was the case.

While this particular example is a political one, it is simply an example of a broader truth. First impressions matter because they influence the arc of a narrative being created in the moment. Later factual data can be revealing and can reverse our understanding of the facts and the events. However, facts have a hard time swaying opinion.

No comments:

Post a Comment