Skip Navigation
Archive

Voter ID Supporters Need Statistics 101

Bad statistical practices – like old habits – die hard. Supporters of voter ID requirements are at it again, this time misinterpreting a new set of election results in Georgia.

  • Sundeep Iyer
July 7, 2011

Any good student of Stat­ist­ics 101 will tell you that correl­a­tion does not imply caus­a­tion. Appar­ently, many voter ID support­ers never got the memo.

Two and a half years ago, Justin Levitt wrote on this  blog about how some proponents of voter ID require­ments were assert­ing that strin­gent ID laws in Geor­gia and Indi­ana did not depress turnout in 2008. Those proponents thought they had found their magic bullet: turnout in Geor­gia and Indi­ana was higher in 2008 than in 2004, despite the imple­ment­a­tion of strict ID laws in the interim.

Mr. Levitt gave them a simple stat­ist­ics lesson. Even if turnout increases at the same time as the adop­tion of a new voter ID law, there may be some­thing other than the voter ID law – Mr. Levitt iden­ti­fied campaign mobil­iz­a­tion, in partic­u­lar – that caused the turnout increase. In other words, correl­a­tion does not imply caus­a­tion.

Bad stat­ist­ical prac­tices – like old habits – die hard. Support­ers of voter ID require­ments are at it again, this time misin­ter­pret­ing a new set of elec­tion results in Geor­gia. In response to E.J. Dion­ne’s Wash­ing­ton Post column on vote suppres­sion efforts across the United States, Geor­gi­a’s Secret­ary of State wrote to the Post’s edit­ors about how an increase in black turnout between 2006 and 2010 showed that voter ID laws do not suppress turnout. Hans von Spakovsky repeated the asser­tion on NPR and in USA Today, and Ohio House Speaker William Batchelder picked up the same message in defend­ing Ohio’s proposed voter ID require­ment. Citing the Geor­gia stat­ist­ics in a see-this-could­n’t-be-that-bad sort of way has become a cent­ral talk­ing point among proponents of voter ID laws.

Once again, these proponents have mistaken simple correl­a­tion for caus­a­tion. You don’t need to be a stat­ist­i­cian to know that without controlling for other factors that might influ­ence turnout, the asser­tion that Geor­gi­a’s voter ID require­ment didn’t depress turnout is mean­ing­less—at best unscientific, at worst just plain wrong.

By compar­ing Geor­gi­a’s turnout with turnout in other similar states that do not have voter ID require­ments, it is possible to control for other factors that influ­ence turnout. Of course, the compar­ison will never be perfect. But it can provide valu­able context.

Consider nearby North Caro­lina, which does not yet have a voter ID require­ment on the books (Governor Bev Perdue recently vetoed a proposed voter ID bill). In 2006, just 28.8% of registered black voters turned out; in 2010, turnout among black voters was 40.4%. Thus, relat­ive to 2006 black turnout, North Caro­lin­a’s level of black turnout in 2010 repres­en­ted a 40.2% increase. Compare those turnout figures with Geor­gi­a’s. In 2006, when there was no voter ID require­ment, 42.9% of registered black voters turned out; in 2010, after the state’s restrict­ive ID require­ment became law, turnout was 50.4%. So relat­ive to its 2006 black turnout, Geor­gi­a’s level of black turnout in 2010 consti­tuted just a 17.5% increase.

In other words, the black turnout jump in North Caro­lina, a state without voter ID laws, was more than twice the size of the jump in Geor­gia, a state with strin­gent voter ID laws. When appro­pri­ately contex­tu­al­ized, Geor­gi­a’s voter ID law no longer looks quite so harm­less.

The notion that Geor­gi­a’s turnout stat­ist­ics alone show that voter ID laws do not depress turnout is fanci­ful and absurd. It is only when we reori­ent ourselves with basic stat­ist­ical prin­ciples—and common sense—that the real story told by Geor­gi­a’s turnout begins to emerge.