The Obama Department of Justice has used a mathematical trick in order to make its legal and public-relations case against South Carolina’s voter ID provisions. Officials of the DOJ intentionally created a false perception of the disparity between black and white voters in order to strike down the law that would haveve required photo ID or other proof for voters. Meanwhile, the media helped Eric Holder’s Department of Justice spin their intentionally misleading numbers to the general public.
This entire scam – there’s no better word for it – is based an interesting but totally irrelevant math quirk; when numbers are small, the difference between them is larger.
An Easy Math Trick
To understand the math behind this, let’s start simply and look at two very small numbers – 1 and 2.
If you had one dollar and I had two dollars, I have one dollar more than you. However, if I wanted to try to impress someone, just telling them that I was a dollar wealthier probably wouldn’t work. In an effort to sound more impressive, I could find a way to pump myself up by claiming that I had twice as much money as you. It’s true, of course – 2 is twice as much as 1 — but just knowing that I had twice as much money as someone else doesn’t really paint the whole picture.
If I want to get extra-fancy, I could also say “I have 100% more money than that person.” This is saying the same (misleading) thing as “I have twice as much” in a slightly different way. There’s a simple equation for determining this percentage difference for any two numbers.
(X / Y) – 1
So in this case, ( 2 / 1 ) – 1 = 1.
To get the ‘percent’ we just move the decimal point to the right two times and we get 100%.
Here’s the quirky part — as numbers get bigger, this difference decreases. Now imagine that you have $100 and I have $101. I still only have one dollar more than you but because the numbers are larger, I can’t pull my ‘twice as much money’ claim trick. In fact, I only have 1% more money than you.
(101 / 100 ) – 1 = .01.
Move the decimal and you’ll see that 101 is 1% more than 100. And of course, saying ‘I have 1% more money‘ doesn’t sound that impressive.
In both cases, I only had one more dollar but the lower the numbers, the more impressive I can make the difference sound.
Now, in both of these examples, the simpler way to express our financial differences is to say “I have a dollar more than you” or by spelling it all out, such as “I have two dollars and you have one dollar.” The whole bit about ‘twice as much money’ or ‘100% more’ is just a way to obfuscate the truth that I only have one more dollar and it works only when the numbers are low.
Thus ends the math lesson.
This intentional obfuscation is exactly what the Federal Government and the Holder Justice Department did to the people of South Carolina. They used this same mathematical quirk to hide the truth about the actual statistics and the media dutifully repeated the ‘Holder Math’. Unfortunately, the state of South Carolina was also bamboozeled by this simple math game and failed to expose what the Obama DOJ was up to.
Now that you’re armed with the knowledge that the difference between low numbers is exaggerated, let’s set the record straight.
The 20% Pufferfish
For a wider overview of the problems with the Department of Justice’s objection, you’ll want to read J. Christian Adams’ summary piece published on the tail of the DOJ’s pre-Christmas announcement. Adams says…
In the objection letter, DOJ said that South Carolina did not meet its burden to prove that photo identification laws did not have any discriminatory effect. Notice the word “any,” more on that later. The data show, according to DOJ, that 1.6 percentage points more voting blacks don’t have a driver’s license than whites. Roughly 10 percent of blacks registered to vote don’t have a photo ID, and 8.4 percent of whites don’t. That represents a “discriminatory effect” under the statute.
And later in the same article…
The letter (from the DOJ) says “minority registered voters were nearly 20% more likely to . . . be effectively disenfranchised.” A difference of 1.6 percent between black and white is now 20%. This statistical sleight of hand was necessary because the actual difference of 1.6 percent (10% vs 8.4%) was laughable to the public. Like a puffer fish that bloats its size to scare predators, the DOJ did the same thing with a 1.6% point difference to scare off critics.
Looking at the letter sent by the DOJ to South Carolina, you’ll see that they brought this 20% figure out twice. First they claim…
When disaggregated by race, the state’s data show that 8.4% of white registered voters lack any form of DMV-issued ID, as compared to 10.0% of non-white registered voters. In other words, according to the state’s data, which compare the available data in the state’s voter registration database with the available data in the states DMV database, minority registered voters were nearly 20% more likely to lack DMV issued ID than white registered voters, and thus to be effectively disenfranchised by act R 54’s new requirements.
Later, the Department of Justice reemphasizes the 20% figure…
Moreover, the state did not provide any data whatsoever refuting the fact, demonstrated by the states earlier data, that minority registered voters are about 20% more likely than white registered voters to lack DMV issued identification.
Once the DOJ had decided on this Holder Math ruse, they simply counted on the media to not even question the 20% number but to repeat it. Here are just a few examples of the media playing up the 20% figure…
Here’s Charles Savage in the New York Times – the article makes no mention of the actual numbers of 10% versus 8.4%, just the 20% difference…
In a letter to the South Carolina government, Thomas E. Perez, the assistant attorney for civil rights, said that allowing the new requirement to go into effect would have “significant racial disparities.”
He cited data supplied by the state as showing that there were “81,938 minority citizens who are already registered to vote and who lack” such identification, and that these voters are nearly 20 percent more likely be “disenfranchised” by the change than white voters.
Here’s ABC News….
“Minority registered voters were nearly 20 percent more likely to lack DMV-issued ID than white registered voters, and thus to be effectively disenfranchised,” Perez wrote, noting that the numbers could be even higher since the data submitted by the state doesn’t include inactive voter.
Here’s Ryan Reilly at Talking Points Memo…again, no mention of actual 1.6 percentage difference but instead playing up the intentionally misleading ‘20%’ figure.
Officials in DOJ’s Civil Rights Division found a significant racial disparity in the data provided by South Carolina, which must have changes to its election laws precleared under Section 5 of the Voting Rights Act, because of past history of discrimination. The data demonstrated that registered non-white voters were 20 percent more likely than white voters to lack the specific type of photo identification required to exercise their constitutional rights, according to a letter sent to South Carolina and obtained by TPM.
So, since the raw data shows only a 1.6% difference between black and white voters where did Department of Justice poll their 20% figure from?
They relied on the fact that the number 10 is 20% larger than the number 8.4. This is that quirk about small numbers that we discussed at the very beginning of the article. It’s true mathematically that 10 is 20% larger than 8.4 ( 10 / 8.4 – 1) but it’s completely irrelevant
(Actually, the correct amount is 19% but rounding the number up is the least scammy thing here)
In order to really understand just how misleading this is, let’s take the exact same data from South Carolina and express it in a slightly different way. Rather than look at the (low numbers of) people who don’t have ID, let’s look at the larger numbers of people who do . If it’s true that 10% of black voters do not have DMV-issued photo ID, then it’s true that 90% of them do. With white voters, it’s 91.4% who have the ID.
So we have the exact same data represented here but because the numbers are higher, look at the difference. 90 is only 1.8% LOWER than 91.4 because (91.4/90) – 1 = about 1.8%. This is crucial to grasp; it’s equally true to say that 10 is 20% higher than 8.4 OR to say that that 90 is 1.8% lower than 91.4. BOTH are true mathematically but neither one matters.
The difference between the two numbers is completely misleading. Why didn’t the Holder DOJ say that the number of black voters with ID is only 1.8% lower than the number of whites with ID? Obviously, because it didn’t suit their purpose.
The number that matters is that there’s a 1.6% real difference, based on the data used – or to spell it out, 10% of black voters without DMV-ID versus 8.4% of white voters. The 20% figure had to be manufactured for some reason and the obvious assumption is that the DOJ is trying to hide the fact that the real difference is so small. They know public opinion is against them so they used a semi-clever trick to fool the state of South Carolina.
This intentional gaming of the statistics by a legal branch of the Federal government should be deeply troubling. The fact that ‘respected’ media like Charles Savage in The New York Times, ABC News and Talking Points Memo went along for the ride is very disturbing. And the fact that this ruse was done in order to protect the Obama administration is downright frightening.