In a Winthrop University poll, 66 percent of South Carolina residents said the state legislature made the right decision to remove the Confederate flag, while 30 percent said it was the wrong decision and 5 percent had no opinion. Although several national surveys have gauged attitudes on the issue, the new poll is the first live-interviewer South Carolina survey published since a tense debate over removing the flag.
The survey marks a major change from a Winthrop poll last year, which found that residents preferred to keep the flag aloft rather than take it down by a nearly 2-to-1 margin (61 percent to 33 percent). When a similar question was asked in the latest survey, the margin was about 2 to 1 in opposition to the flag flying (61 percent to 32 percent).
Whites and blacks differ sharply on the issue, but both groups have shifted significantly in the past year. Nearly half of whites now oppose the flag flying at the state house, up 27 percentage points from last fall. In a separate question, 56 percent of whites say state lawmakers were right to remove the flag.
African Americans have been united, with opposition to the flag’s display rising from 61 percent in 2014 to 90 percent in the latest survey. Similarly, the share of black South Carolinians who support flying the flag dropped from 27 percent last year to 4 percent today.
Republicans are still divided on the flag; 50 percent say the state made the right decision in removing it from state house grounds, while 45 percent say the opposite. Asked their preferred outcome, 49 percent of Republicans said they would like to see the flag fly, while 43 percent do not. Democrats are more united, with 83 percent supporting the decision to remove the flag and 81 percent saying it should no longer fly at the state house.
While Gov. Nikki Haley (R) was vocal in pushing for the flag’s removal – a now-popular decision – her approval rating shows no significant shift since this spring in the Winthrop poll. Fifty-five percent of South Carolinians approve of her performance as governor, similar to 53 percent before the flag debate began. But her approval dipped from 78 percent to 68 percent among Republicans and GOP-leaning independents — a modest drop-off that may be a result of her opposition to the flag’s display.
The Winthrop poll suggests a sharp change in public opinion on the Confederate flag, which seems to be a plausible reaction to the shootings in Charleston and intense public debate when many lawmakers from both parties came out in opposition to its display.
But another possibility could explain some change in the results: The Confederate flag debate in South Carolina might have increased the social stigma of supporting the flag. Supporters of keeping the flag were shamed during debates, and the idea of the Confederate flag as a symbol of racism was a core argument of the debate.
The impact can be a "social desirability" bias in results — the tendency for respondents to avoid reporting embarrassing or potentially offensive answers to interviewers. The Winthrop poll used live interviewers to conduct the survey — which has several benefits to surveys (being able to dial cellphones is one) — but it also can make respondents less willing to report opinions they think will offend the interviewer.
Quantifying social desirability bias is difficult outside an experimental setting, but one result mentioned prominently in Winthrop’s release raises an eyebrow. When asked whether they approved or disapproved of the flag being displayed before this summer, 49 percent of respondents said they disapproved. That level of opposition is significantly higher than the 32 percent in a 2014 survey who said the flag should not be flown on state house grounds. The result might be borne out of respondents' unwillingness to admit they once supported the flag or themselves do not recognize their change in attitude.
Winthrop political science professor Scott Huffmon tweeted his interpretation of the apparent discrepancy:
The Winthrop Poll was conducted Sept. 19-27 among a random sample of 963 adults reached on landline and cellular phone numbers in South Carolina. The overall margin of sampling error is plus or minus 3.2 percentage points; among whites, the error margin is 3.8 points; among African Americans, the error margin is six points.
Peyton M. Craighill contributed to this report.