Liberty Hill

Race and Polling, Part 3: The Reverse Bradley Effect

    

    

                Interestingly, there is perhaps more evidence that the 2008 election may have a reverse Bradley effect for some states -- that black voters are being under-counted and states with a high black demographic may actually have higher numbers for Obama than what the polls suggest.  The hypothesis is that black voters were reluctant to disclose to typically white pollsters their voting preferences.

     Based on polling data and actual results from the 2008 primaries, Nate Silver argues that: “There is nowhere in the country where we have reason to subtract from Obama's poll numbers based on the Bradley effect. (Yes, Obama has underperformed his polls in some "white" states -- but he has overperformed in others that are whiter than Kurt Rambis). . . On the other hand, there is one specific group of states where we might want to add to Obama's polls based on the Reverse Bradley Effect, which are Southern states with high African-American populations.”

     Taking the findings from Dan Hopkins’ study, Nate surmises that certain polling methods can cause black voters to understate their preference for a black candidate when they perceive they are being surveyed by a white interviewer.  Nate’s theory is supported in part by studies that demonstrate the race of the interviewer can affect responses.  Since we know from studies that people are very good at guessing the race of the interviewer, Nate argues that black voters may not want to tell the predominantly white interviewers their voting preference (fear that the white interviewer might be racist or not wanting to be seeing as fitting into a stereotype), and will usually respond as undecided. 

     Interestingly, Nate and other polling analysts have noticed that pollsters like Rasmussen, SurveyUSA, and PPP that use automated callings scripts did not have much of a reverse Bradley Effect.  While not conclusive, this is intriguing evidence that the mostly white interviewers do cause black voters to understate their preference for a black candidate.  Also, based on this hypothesis, Nate’s analysis of the polling date in March led him to predict that NC and VA, former Bush states, would ultimately go to Obama.  Recent polls, particularly from Rasmussen, SurveyUSA, and PPP are now showing Obama with leads in those states.

     So, to conclude, race may very well be a factor in the 2008 Presidential election.  But statistical analysis of the data does NOT show that the polling is under-reporting votes for Obama.  If Obama remains ahead in the electoral votes as the current polls are indicating, the fear that Obama needs a cushion of votes to compensate for the Bradley effect simply does not hold water.  Of course, a cushion of votes for any candidate is preferable to protect against other types of polling errors, lower voter turn-out, unexpected events, etc.  There’s enough to worry about in the weeks running up to the election, regardless of your preferred candidate.  I just wouldn’t spend too much of that angst on the Bradley effect.

     These posts were researched and written by Liberty Hill’s Chad Finlay. See yesterday and Friday October 3 for the earlier posts.

                                                                                                                                                                                                                                                    UPDATE:  A study just released by University of Washington researchers has found evidence of the Reverse Bradley Effect in 13 states during the primary, mostly amongst the Southeastern states Nate Silver also identified.

                                                                                    Other good articles on Bradley effect:

The Disappearing Bradley Effect

On Race Based Voting

Bad Math and the Bradley Effect

Race and Polling, Part 2: Why The Myth Persists

     Chad Finlay, a new addition to the Liberty Hill staff, has done some analysis of the so-called Bradley Effect. His first post went up on Friday. Here’s part 2. According to Nate Silver of the polling analysis site FiveThirtyEight.com, the “myth” of the Bradley effect persists for four reasons:

                                                                                       

                                    1. Misunderstanding the Bradley effect. Denying the existence of the Bradley effect does not mean denying that some people vote on the basis of race. Some people will not vote for Obama because he is black. Some people will vote for Obama because he is black. The Bradley Effect is not an argument about whether people vote based on race. It's an argument about whether people will lie to pollsters.

2. Confusing Past with Present. There is fairly strong academic evidence that the Bradley effect used to exist in the 1980s and early 1990s. However, the evidence is just as strong that it does not exist any longer. The people who vouch for the existence of the Bradley Effect are not wrong so much as they are relying on dated evidence.

3. Confusing Exit Polls with pre-Election Polls. Unlike the normal, pre-election polls, exit polls conducted on the day of the election did substantially overstate Barack Obama's margins throughout the primaries. This is something to keep in mind at about 5 PM on November 4, when Matt Drudge and Jim Geraghty begin to leak exit poll results, but not when discussing pre-election polling.

4. Cherry Picking Results. The notion of the Bradley effect gained a lot of currency after the New Hampshire primary, when Hillary Clinton did much better than anyone expected and won the state. However, the 8.9-point gap separating the pre-election polls and the actual results in New Hampshirerepresented only the seventh-largest error in the primaries. There were bigger discrepancies in Iowa, South Carolina, Alabama, Georgia, Wisconsin and Mississippi, all of which favored Barack Obama. These discrepancies did not receive as much attention as New Hampshire because they did not change the outcome of the election. But mathematically speaking, they were just as important.

Tomorrow, we’ll post Chad’s insights into the reverse Bradley Effect.

Race and Polling, Part 1: The Impact of the Bradley Effect

     I can’t turn the radio on without hearing speculation about “The Bradley Effect.” How much should Obama's popularity be discounted because of racism white voters are unwilling to acknowledge to pollsters? 

     The smartest and most comprehensive analysis of the Bradley Effect I’ve seen comes from Liberty Hill’s own Chad Finlay. Here’s his research and analysis on The Bradley Effect and the Reverse Bradley Effect. Over the next few Fridays, we’ll be posting Chad’s analysis.

     The Bradley effect was named after the 1982 California gubernatorial race in which Tom Bradley was projected to win based on exit polls, but ultimately lost. Post-election analysis posited that voters over-reported voting for Bradley to avoid appearing prejudiced against a black candidate. 

     Despite the attention paid to the Bradley effect, in the 2008 presidential election, the Bradley effect DOES NOT seem to be a factor.  Data from the 2006 Congressional elections (most notably Harold Ford who received more votes than the polls indicated) and the 2008 primaries do not show evidence of the Bradley effect being in play.  There was also an important new study released in August by Harvard Professor Dan Hopkins that analyzes the affect of race and gender on elections.  Since the 1996 elections, Dan did not find any evidence of the Bradley effect – “Before 1996, the median gap [between polls and actual results] for black candidates was 3.1 percentage points, while for subsequent years it was -0.3 percentage points.”

     Check back next Monday for Chad's research on why The Bradley Effect persists as myth if not as reality.