Monday, March 03, 2008

Misusing Statistics

Our good friend over at Crawford’s Take is frustrated.  She says:

If a black man and a white woman with the same qualifications are applying for a job in America, the facts are that the white woman will get the job nearly every time. This is the way America works. It is one of the things that is wrong in America, but it exists none-the-less… Now granted if they hired the woman, she’d be paid 30% less than the white man, but the point still remains.

Ahhh… the danger of statistics.  Let’s look at the assertions in isolation.  First:

...white woman will get the job nearly every time.

Is that true?  I couldn’t find any statistics on this, but I’ve hired a few people in my career and know many people who have.  I haven’t seen any such trend.  If anything, if the qualifications are the same, the black person tends to get hired more often because it “diversifies” the work force and puts the company in a better position to get government contracts where racial factors are taken into account.  Then there’s this:

...she’d be paid 30% less than the white man…

That’s outright false.  First of all, the actual “gender gap” is about 24%, not 30%, but semantics aside, it doesn’t apply in this circumstance.  The “gender gap” is a global statistic of wages.  It takes all of the wages earned by women compared to all of the wages earned by men, divide it by the number of people involved to normalize it, and you get your percentage.  While an interesting statistic, it does not measure the relative disparity in a given job.

The gender gap measurement takes into account a host of factors, including the kinds of jobs women and men tend to occupy, the length of time in the workforce, etc.  The fact that women dominate low wage jobs like house cleaning and child care brings down the average wage.  On the flip side, the fact that men dominate fields like engineering and executive management brings up their average.  Then you throw in factors like the tendency of women to take time off of work for children and such and their average wage is driven even lower.  These are cultural or societal norms that can be debated, but they don’t lend to Crawford’s point. 

Crawford puts forth the supposition that if a white woman and a black man apply for the same job (presuming that they are equally qualified), that the woman will get the job but be paid 30% less.  She is inserting a global statistic regarding aggregate wages into a specific scenario where it doesn’t apply.  It’s a fun way to spin her position, but it’s a meaningless application of statistics.

(2) Comments
Posted by Owen at 2314 hrs
Culture + Politics + Politics - General

  1. So if Clinton somehow gets elected president, it will save us some tax money from her salary?

    Posted by Matt on March 04, 2008 at 0936 hrs


  2. Wasn’t it former prime minister Benjamin Disraeli who said there were three kinds of lies:  lies, damn lies, and statistics?

    Posted by .(JavaScript must be enabled to view this email address) on March 04, 2008 at 1040 hrs


Commenting is not available in this channel entry.