GALLUP NEWS SERVICE
PRINCETON, NJ – Black Americans are much less likely to support the current war against Iraq than they were to support the 1991 Persian Gulf War. Other groups whose support for the current conflict is dramatically lower include ideological liberals, political independents, low-income households, and 18-29 year olds. Aside from political independents, these groups tend to be Democratic in their political orientation. A special statistical analysis of the data shows that, as today, support for the 1991 conflict was most strongly related to presidential job approval. While race, partisanship and political ideology were all important factors then as now, the effect of race is much stronger today than in 1991.
Support For The War By Group, 1991 Versus 2003
A 优蜜传媒analysis published last week explored support for the current war with Iraq among demographic groups, finding support lowest among blacks and liberals, and highest among Republicans, conservatives and men. Those same patterns generally held in 1991, although overall support may have been slightly higher in 1991 than it is now. Currently, about 7 in 10 Americans say they "favor the war with Iraq." Back in 1991, about 8 in 10 approved of the "U.S. decision to go to war with Iraq in order to drive the Iraqis out of Kuwait."
Given differences in question wording, one cannot necessarily conclude that support for the war is lower now than it was in 1991. The 1991 question included a reason for the United States' taking military action ("to drive the Iraqis out of Kuwait"), while the 2003 question does not. Some other polling organizations' questions give a reason for military action, and find current support running in the high 70% range. However, other measures with identical question wordings asked in both years, such as presidential approval and approval of the president's handling the situation in Iraq, were also about 10 percentage points higher in 1991 than today.
As such, it is possible to compare the 1991 and 2003 support numbers at the outset of each war to illustrate where differences in support lie between the two conflicts.
The following table details the differences in support by group between the previous war with Iraq and the current situation. Keep in mind that some of the differences between 1991 and 2003 may be due to question wording, and as a result the figures may not be directly comparable:
Support for U.S. Wars with Iraq by Group, 1991 and 2003
|
Group |
% Favor, 2003 |
% Approve, 1991 |
Difference between 2003 and 1991 |
|
Overall |
72 |
81 |
-9 |
|
Men |
78 |
86 |
-8 |
|
Women |
66 |
76 |
-10 |
|
White |
78 |
83 |
-5 |
|
Black |
29 |
59 |
-30 |
|
18-29 years old |
66 |
82 |
-16 |
|
30-49 years old |
75 |
84 |
-9 |
|
50-64 years old |
69 |
83 |
-14 |
|
65 years and older |
73 |
71 |
+2 |
|
Reside in east |
66 |
79 |
-13 |
|
Reside in Midwest |
73 |
80 |
-7 |
|
Reside in south |
71 |
85 |
-14 |
|
Reside in west |
77 |
81 |
-4 |
|
Less than H.S. education |
67 |
76 |
-9 |
|
High School Graduate only |
73 |
83 |
-10 |
|
Some college education |
77 |
83 |
-6 |
|
College graduate |
65 |
80 |
-15 |
|
Income less than $30,000 |
58 |
77 |
-19 |
|
Income $30,000-$49,999 |
79 |
85 |
-6 |
|
Income $50,000 and over |
77 |
86 |
-9 |
|
Conservative |
84 |
90 |
-6 |
|
Moderate |
70 |
77 |
-7 |
|
Liberal |
44 |
72 |
-28 |
|
Republican |
93 |
92 |
+1 |
|
Democrat |
66 |
79 |
-13 |
|
Independent |
53 |
73 |
-20 |
|
Children under 18 |
76 |
84 |
-8 |
|
No children under 18 |
70 |
79 |
-9 |
Among the key findings:
- The biggest single shift in war support among the groups analyzed is for blacks, whose support for the current war effort (29%) is less than half what it was in 1991 (59%). By comparison, whites have shown just a modest drop in support.
- Men and women show about an equal decline in support between the two wars – 8 points for men and 10 points for women, comparable to what is observed for the nation as a whole
- Older Americans – those age 65 and older – are one of the few groups that show no decline in support between the two wars. Seventy-three percent of older Americans support the current war, compared with 71% support for the previous war with Iraq.
- Each of the remaining age groups shows significantly lower support for the current conflict, with the largest drop among 18-29 year olds (82% in 1991 compared with 66% now). Americans between the ages of 50 and 64 show a 14-point decline, from 83% support in 1991 to 69% today.
- In 1991, there were essentially no differences in support by region. Today, residents in the east show somewhat lower support, and residents in the west show somewhat higher support. But, in all regions support is lower compared with the 1991 levels.
- Among educational groups, the largest change in support (15 percentage points) is seen among college graduates. This group includes Americans with a post-graduate education, who have been among the least supportive groups of the current military action in Iraq (college graduates with no post graduate education have been among the more supportive groups). In 1991, 优蜜传媒did not distinguish between American college graduates who did and did not have any post-graduate education, so a comparison of post-graduates in 1991 and 2003 is not possible.
- Americans in low-income households show decidedly less support for the current war with Iraq than for the previous one. Just 58% of Americans whose household income is $30,000 or less now favor the war with Iraq, compared with 77% approval of the decision to go to war with Iraq in 1991.
- Among ideological groups, liberal support for the current war is dramatically less than it was for the last Persian Gulf War, 44% compared with 72%. Conservatives' support levels are only slightly lower currently (84%) than in 1991 (90%).
- Republicans show essentially no change in support between the current and past Iraqi conflicts (93% for the current conflict and 92% for the previous one). Democrats show a decline of 13 percentage points, from 79% to 66%. Political independents show an even larger decline, from 73% to 53%.
In 1991, 91% of Americans who approved of the elder George Bush as president approved of the decision to send troops to the Persian Gulf to drive the Iraqis out of Kuwait. That compares with just 29% approval among those who disapproved of the president. In the current conflict, 93% of those who approve of George W. Bush favor the war with Iraq, while just 20% of those who disapprove of him favor the war. Previous 优蜜传媒analysis has shown that support for the war was strongly related to approval of George W. Bush as president, and in fact presidential approval made the greatest independent contribution to support for the war.
Presidential Job Approval, Race Most Strongly Related to War Views
Many of the groups that are showing less support for the current war – blacks, liberals, low-income households, more highly educated people -- generally come from a Democratic or liberal political persuasion. In a previous article, a 优蜜传媒statistical analysis showed that taking into account many possible factors related to war approval, the president's job rating was most strongly related to support for the war, followed by race, ideology and partisanship. A similar analysis of the 1991 data reveals essentially the same results – the most important predictors of support for the war in 2003 are generally the most important predictors of support for the war in 1991.
In both years, presidential approval far and away shows the strongest independent relationship to support for the war, taking into account the effect of other variables. However, presidential approval seems to have had an even greater impact in 1991 than it does in 2003. For the first war, the likelihood of approving of the decision to go to war with Iraq was 25 times higher among those who approved of the elder George Bush as president than among those who disapproved. For the current war, the likelihood of favoring the war is about 7 times greater among those who approve than among those who disapprove of the job George W. Bush is doing as president.
One notable difference in the correlates of support concerns race, which has a substantially greater independent effect on war attitudes now than it did in 1991. Today, blacks are nearly five times more likely than non-blacks to oppose the war, on average, taking into account the effect that other variables have on war support. In 1991, blacks were only twice as likely as non-blacks to oppose the war. Partisanship and ideology, important factors today and in 1991, were slightly more important than race in determining one's view on the war in 1991, even though they were less strongly related to support for the war than they are currently.
The following table shows the average increase in likelihood of favoring the war for each of the variables tested, in each year, with the variables ordered according to the strength of their influence in 2003. Figures around 1.0 (or –1.0) indicate no difference in likelihood, while numbers further away from 1.0 indicate increasing likelihood. Negative numbers indicate increased likelihood of opposing the war.
|
Variable |
Relevant Comparison(s) |
Increased Likelihood of Favoring War, 2003 |
Increased Likelihood of Approving of War, 1991 |
|
Presidential Approval |
Approve vs. Disapprove |
7.1 |
25.2 |
|
Race |
Black vs. Non-Black |
(-4.8) |
(-2.0) |
|
Ideology |
Conservative vs. Liberal |
3.6 |
2.8 |
|
Moderate vs. Liberal |
1.8 |
1.4 |
|
|
Party Identification |
Republican vs. Democrat |
3.4 |
2.8 |
|
Income |
Less than $30,000 vs. All others |
(-3.0) |
(-2.5) |
|
Gender |
Male vs. Female |
1.9 |
1.4 |
|
Education |
Postgraduate vs. All others |
(-1.4) |
-- |
|
High school or less vs. All others |
(-1.4) |
-- |
|
|
College graduate vs. All others |
-- |
(-1.2) |
Note: When taking into account the effect of other variables, age and region no longer show meaningful differences, and so are not included in the table. Statistics shown are the odds ratios found when using a logistic regression with opinion on the war as the dependent variable.
Survey Methods
These results are based on telephone interviews with combined randomly selected national sample of 2,028 adults, 18 years and older, conducted March 22-23 and March 24-25, 2003, and 2,030 adults, 18 years and older, conducted January 17-20 and January 23-26, 1991. For results based on these samples, one can say with 95 percent confidence that the maximum error attributable to sampling and other random effects is plus or minus 3 percentage points. In addition to sampling error, question wording and practical difficulties in conducting surveys can introduce error or bias into the findings of public opinion polls.
Support for U.S. Wars with Iraq by Group, 1991 and 2003, Sorted by Difference in Support
|
Group |
% Favor, 2003 |
Sample Size 2003 |
% Approve, 1991 |
Sample Size, 1991 |
Difference in support between 2003 and 1991 |
|
Overall |
72 |
2028 |
81 |
2030 |
-9 |
|
Black |
29 |
141 |
59 |
159 |
-30 |
|
Liberal |
44 |
369 |
72 |
622 |
-28 |
|
Independent |
53 |
667 |
73 |
663 |
-20 |
|
Income less than $30,000 |
58 |
507 |
77 |
823 |
-19 |
|
18-29 years old |
66 |
301 |
82 |
478 |
-16 |
|
College graduate |
65 |
787 |
80 |
576 |
-15 |
|
50-64 years old |
69 |
517 |
83 |
384 |
-14 |
|
Reside in south |
71 |
656 |
85 |
595 |
-14 |
|
Reside in east |
66 |
457 |
79 |
504 |
-13 |
|
Democrat |
66 |
584 |
79 |
656 |
-13 |
|
Women |
66 |
1064 |
76 |
1015 |
-10 |
|
High School Graduate only |
73 |
515 |
83 |
690 |
-10 |
|
30-49 years old |
75 |
844 |
84 |
841 |
-9 |
|
Less than H.S. education |
67 |
129 |
76 |
216 |
-9 |
|
Income $50,000 and over |
77 |
960 |
86 |
483 |
-9 |
|
No children under 18 |
70 |
619 |
79 |
1223 |
-9 |
|
Men |
78 |
964 |
86 |
1015 |
-8 |
|
Children under 18 |
76 |
394 |
84 |
799 |
-8 |
|
Reside in Midwest |
73 |
474 |
80 |
517 |
-7 |
|
Moderate |
70 |
744 |
77 |
448 |
-7 |
|
Some college education |
77 |
587 |
83 |
538 |
-6 |
|
Income $30,000-$49,999 |
79 |
455 |
85 |
540 |
-6 |
|
Conservative |
84 |
892 |
90 |
938 |
-6 |
|
White |
78 |
1673 |
83 |
1756 |
-5 |
|
Reside in west |
77 |
441 |
81 |
411 |
-4 |
|
Republican |
93 |
750 |
92 |
690 |
+1 |
|
65 years and older |
73 |
353 |
71 |
310 |
+2 |
