# Words, Meet Data

Published 11 years, 11 months past

At lunch today (okay, by now it’s actually yesterday), I had some leftover time.  Not quite enough to play a shift on Radar Chaos: Hawaii Edition, sadly, but enough to run down the answer to a question that had been bugging me for a couple of days: if Mitt Romney is right that the 47% of people who pay no federal income taxes are all Obama voters that he needn’t worry about, what would the polls look like?

Finding out was as simple as hitting up Wikipedia, Gallup, and Excel.  Combining the two sources of information, I came up with this:

Income range% of pop.ObamaRomney
Less than \$36,00040.1%56%37%
\$36,000 to \$89,99940.4%47%47%
\$90,000 or more19.5%45%50%

I used those income bands because they’re the divisions used in the Gallup results, and I got the percentage-of-population numbers from Wikpedia.  For simplicity’s sake, I decided to assume that Mitt was off by a bit and that only 40.1% of the electorate was lost to him — meaning that any person (or household) making more than \$36,000 a year but paying no federal income tax due to writeoffs, tax credits, and so on was still in play.

So once you combine the percentages in that table and add up the results, you get 50.2% of the vote for Obama, and 43.6% of the vote for Romney.  That was likely true immediately post-DNC but is a wider split than the most recent polls show, which are usually close to being tied at 47% each.  So we’ll note that I’ve given Obama 3% too many and Romney 3% too few, and apply that correction at the end.

The next step was to shift all of the 40% of voters below \$36,000 a year into the Obama column.

Income range% of pop.ObamaRomney
Less than \$36,00040.1%100%0%
\$36,000 to \$89,99940.4%47%47%
\$90,000 or more19.5%45%50%

Once you do the math on those figures, the results are 67.9% of the vote for Obama and 28.7% of the vote for Romney.  Apply the ±3% correction from before and you wind up with Obama getting about 65% of the vote and Romney getting about 32%.  That’s substantially different than polls have been showing, so I’m going to go out on a limb and say that Mitt is mistaken about the composition of the electorate.

(Of course, as noted before, I spotted Mitt almost 7% of U.S. households by only shifting the bottom 40.1% to Obama instead of the bottom 47%.  You’re welcome, Mitt.)

Anyway, that was my every-half-decade analysis of public data, which I like to conduct both to look at things from a new angle as well as celebrate what the web has made possible: the ability to look deeper, to analyze, to ask questions and find answers, to verify.  Even hypotheticals.

HTML allowed: `<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <em> <i> <q cite=""> <s> <strong> <pre class=""> <kbd> `