RE: Statistical analysis

[Dan Goldstein <DGoldstein -at- nuot -dot- com>]

Fortunately, there's no money involved. :-)

As for the other caveats below, this is exactly why I asked for a volunteer with experience in statistical analysis. I'm assuming that my own analysis would be not only unqualified, but probably biased as well.

So far, there are three exclusions that will be described clearly in the final report. The volunteer statistician might question any of those exclusions or might suggest making additional ones.


-----Original Message-----
From: Peter Neilson
Sent: Monday, April 11, 2016 9:51 AM
To: techwr-l -at- lists -dot- techwr-l -dot- com
Subject: Re: Statistical analysis

While we are caveatating...
- Fallacy of assigning causes to the observed coincidences (false correlation).
- Sample size problems even when it looks like there aren't any.
- Lumping incommensurables together to create a decently sized sample set.
- Selection of data by unidentified causes.

In some agenda-driven studies (for instance, funded to provide a particular result) outliers or data that is "obviously bad" are eliminated. This has the result of sharpening a peak (reducing standard deviation), or even creating a peak where none previously existed. In extreme cases researchers might ignore a bimodal distribution. Fortunately we are all too smart to fall into any of those blunders.


^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Visit TechWhirl for the latest on content technology, content strategy and content development | http://techwhirl.com

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

You are currently subscribed to TECHWR-L as archive -at- web -dot- techwr-l -dot- com -dot- 

To unsubscribe send a blank email to
techwr-l-leave -at- lists -dot- techwr-l -dot- com


Send administrative questions to admin -at- techwr-l -dot- com -dot-  Visit
http://www.techwhirl.com/email-discussion-groups/ for more resources and info.

Looking for articles on Technical Communications?  Head over to our online magazine at http://techwhirl.com

Looking for the archived Techwr-l email discussions?  Search our public email archives @ http://techwr-l.com/archives