How to lie with Statistics
Jan. 13th, 2012 10:21 am![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png)
When I was in high school our AP Biology teach (Mr Patterson) assigned us a book entitled How to Lie With Statistics. The book started with the infamous 4 out of 5 dentists recommend sugarless gum for their patients who chew gum and proceeded to introduce students to statistics and how they can easily be manipulated. I was reminded of this when I saw a friend posting an infographic on Facebook.
This graphic (put out by the White House) is based on BLS statistics that can be found at BLS.Gov. Go to the Database and Tools and select the Data Retrieval Tools, then Select Employment and the Top Picks for Employment, Hours, and Earnings (National). The specific data being looked at is Manufacturing Employment.
All well and good, and the 334K number does reflect the variance from the Jan 2010 number 11,465K manufacturing jobs and the Dec2011 number of 11,790K manufacturing jobs. This is a nice improvement of 334K jobs. Of course this is down from the Jan 2001 number of 17,114K manufacturing jobs which was the high point for such jobs (the high point was 17,637K jobs in March of 1998…so from its high point, the number had slipped just over 500K jobs.
This chart starts in 1991 and goes to 2011.
A while back I had seen a similar graphic that touted how President Obama had turned around the job creation debacle and things were rosy.
Note that how when President Obama took office, that the economy is losing close to 800K jobs a month and by mid 2010 it is creating jobs at 50-200K a month. I noted that the chart only goes back to Dec 2007 and doesn’t show job creation in the mid years of President Bush’s administration. So I created a chart that went back further.
Note, that this chart shows that when President Bush came into office in Jan of 2001 the economy was hemorrhaging jobs and by 2003 it was turned around and creating at a clip that is a bit better (not much, just a bit) than current trends. Then there is the 08 crash and it gets very nasty. I did this to show that taking a snapshot in time can be illustrative, but often exaggerates something that may, or may not be important. In this case it shows that while the recession of 08-09 was bad…it came on the heels of a five year job growth span that it turn came on the heels of a two year recession. So, I did the same chart back 40 years for comparison.
We see that the normal state is growth, and the dips are short, relatively severe, but offset by long periods of growth. This is true until the 21st century when growth is lessened and the dips longer. At the time I mentioned to my friend that it would be interested to see how this compared to who controlled Congress/The White House.
After he posted the 334K infographic I decided to do a quick run using number from 2001-2011. I took the job creation numbers and I gave credit to each party for what they had control of. For example in Jan 2001 the GOP controlled the House and the Dems controlled the Senate. Thus in that month 16K jobs were lost and each party got credit for half (-8K each). In May of 2006 the GOP controlled both houses and thus got full credit for all 11K jobs created that month. In December of 2007 there were 84K jobs created and the Dems controlled both houses and got full credit for both. here is what that chart looks like
Pardon the X axis but it was a mess trying to get that right. As you can see this chart is designed to share the credit and split the blame. Overall from 2001 the numbers look like this:
GOP Job Creation: 6,281K jobs
Democrat Job Creation: –7,058K jobs.
OK, so, this is just as bad because it ignores the President, so I modified the formula to give the President half the credit and Congress half the Credit (each house getting 1/4 of the total). When you graph that data it looks like this…
OK, that looks the same, but the numbers change a bit:
GOP Job Creation: 3,140.5K jobs
Democrat Job Creation: –3,529L jobs
Presidential Impact: –388K jobs (which is 949K for President Bush and –1,337.5K for President Obama).
My friend would say that the first year shouldn’t count against President Obama (I don't necessarily agree, but we will also slice of 2001 for President Bush) and the numbers look like this:
President Bush: –701K jobs created.
President Obama: 1,194K jobs created
What does all that mean? Not much if you think back to the title of this post…or everything if you are trying to make a political point using said statistics. The big problem is that too many people are not willing to actually look at the full data, or think about other reasons why things happened. Should the President be blamed for everything? Likely not, but, much like a Quarterback…they will end up taking the heat when times are bad the credit when things are good. Was President Clinton responsible for the boom of the 90s…or was it the GOP Congress? Your answer will depend upon your politics.