Wednesday, May 19, 2010

The Unethical Use Of Statistics

Here is the problem with unethical usage of statistics:

Veronique de Rugy of Big Government referenced a study on the compensation of both public and private sector employees during a time of recession and came to the conclusion that public sector earnings rose 8% greater then their private sector counterparts at a rate of 1.92% to the private sector's 1.78%. Based on these findings, de Rugy comes to the conclusion that it must be great to be a public sector employee at a time of recession.

The problem is that when dealing with such small numbers, figures can be made to appear greater then what they really are. This is like investing in penny stocks - the price of the stock is sometimes worth tenths of a penny but as soon as the slightest increase takes place, the investor would experience large gains, but the increase in those particular stocks are not representative of the entire market. The research de Rugy cites essentially does the same, making the minute difference in pay seem greater then what it really is.

Check out this misuse of statistics coming from John Smithson, who cites the most recent Rasmussen report on the presidential approval without linking to the source - Smithson compares those who "strongly approve" to the "total disapprove" column, making the disparity appear greater then what it really is.  Keep in mind that there is a margin of error of +/- 3 percentage points as well.



I would also like to point out that Rasmussen has engaged in unethical polling in the past, making the results of any of their other surveys very questionable to say the least.  Eric Boehlert of Media Matters had written about the pollster in the past, pointing out one particular poll in which Rasmussen asked only Republicans of their opinion of Sarah Palin but passed off their results as representative of all political ideologies.

"What kind of polling firm, while trying to take the country's temperature about politics, only questions Republicans?"  Boehlert asked.

Apparently misusing statistics is alive and well among conservative activists like Smithson and de Rugy...

1 comment:

  1. Oh my god! A .14% difference! Has the world gone mad??!!

    ReplyDelete

Please share your thoughts and experiences in relation to this post. Remember to be respectful in your posting. Comments that that are deemed inappropriate will be deleted.