Sandbox: Difference between revisions

From ChanceWiki
Jump to navigation Jump to search
Line 1: Line 1:
==Politics and porn (What's the matter with Kansas?)==
[https://source.opennews.org/en-US/learning/distrust-your-data/  Distrust your data]<br>
by Jacob Harris, opennews.org, 22 May 2014
Harris identifies 6 ways to make mistakes in reporting data:
*Sloppy proxieson
*Dichotomizing
*Correlation does not equal causation
*Ecological inference
*Geocoding
*Data naivete
His prime example is a story that was widely circulated via social media, featuring the following scatterplot
<center>[[File:Porn_politics.png‎  | 500px]]</center>
Kansas is a clear outlier.  Harris credits a reader of Andrew Sullivan's blog for the following explanation of the geocoding problem:
<blockquote>
What happened here was that a large percentage of IP addresses could not be resolved to an address any more specific than “USA.” When that address was geocoded, it returned a point in the centroid of the continental United States, which placed it in the state of—you guessed it—Kansas!
</blockquote>
Kansas aside, the red/blue divide is still striking. The "ecological fallacy" here is similar to Durkheim's (see Chance News 92 [http://test.causeweb.org/wiki/chance/index.php/Chance_News_92#Simpson.E2.80.99s_paradox_and_the_ecological_fallacy here] for more discussion), where he noted that the more Protestant the Prussian province,  the larger the suicide rate--but it turns out that the suicides were actually committed by Catholics, not Protestants.  The possible analogy here: in Democratic states it may be the Republicans who are frequenting pornography web sites.
[http://andrewgelman.com/2014/05/24/buzzfeed-porn-kansas-results-cant-good/ More commentary] can be found on Andrew Gelman's blog.
Submitted by Paul Alper


==Reproducibility==
==Reproducibility==

Revision as of 01:33, 10 June 2014

Reproducibility

When studies are wrong: A coda
by George Johnson, New York Times, 7 March 2014


Submitted by Bill Peterson