On Dec. 8, President Bush signed into law a bill that provides prescription drug benefits to seniors, and also changes the way seniors can receive Medicare coverage. Two weeks ago, a CNN/USA Today/Gallup survey asked respondents whether the changes "go too far, are about right, or do not go far enough?"
|
Do you think -- [Read A/B] -- go too far, are about right, or
do not go far enough? |
||||
|
Too far |
About |
Not far |
No |
|
|
A. New prescription drug benefits for Medicare recipients |
9% |
27 |
53 |
11 |
|
B. Changes made in Medicare coverage |
12% |
27 |
50 |
11 |
These results suggest that a bare majority of people believe both changes do not go far enough. But the percentage saying "not far enough" depends on the order in which the two items are mentioned.
When the items are read in the order shown in the table, prescription drugs followed by changes in Medicare coverage, more people say the changes were "not enough" than when the items are read in reverse order.
|
Do you think [A/B] go too far, are about right, or do not go
far enough? |
||
|
Order in Which Items Were Read: |
A then B |
B then A |
|
Response: |
Not far enough |
Not far enough |
|
A. New prescription drug benefits for Medicare recipients |
56% |
49 |
|
B. Changes made in Medicare coverage |
54% |
47 |
When the "new prescription drug benefits" item is mentioned first, followed by "changes made in Medicare coverage," a majority of respondents say each item does not go far enough -- 56% who say that about prescription drug benefits, and 54% about the changes in Medicare coverage.
But when the "changes in Medicare coverage" item is mentioned first and drug benefits second, only 47% say the changes in Medicare coverage do not go far enough, and just 49% say the drug benefits do not go far enough.
The difference in percentages between the two versions are statistically significant, though modest -- a seven-point difference for each item in the percentage saying "not enough." So, what is happening? Why do the percentages vary depending on which item is mentioned first?
Opinion Anchoring on Similar Issues
When the first question was asked about prescription drug benefits, respondents made their judgment (about whether the benefits went far enough or not) based on whatever information they could "retrieve" from their brain cells at that time.
But when the second question about changes in Medicare coverage was asked, respondents could then compare their reactions to the adequacy of coverage with their reactions to the adequacy of drug benefits. Overwhelmingly, whatever people said about the first item, they also said about the second one. The first item acted as the "anchor" for the second.
So, why the higher percentage when prescription drugs are mentioned first? Because people apparently tend to be more upset about prescription drugs than about changes in Medicare coverage -- or, at least their "top-of-mind" reaction is more negative. That conclusion is based on comparing the public's responses to each item when that item is mentioned first.
When prescription drugs are mentioned first, 56% say they do not go far enough. When changes in Medicare are mentioned first, only 47% say those changes do not go far enough. Thus, in a non-comparative context -- when people are first asked about each item without the other item having been mentioned -- more people say prescription drugs are inadequate than say that about the changes in Medicare coverage.
Sons and Daughters in Politics
Gallup found a similar pattern just over a decade ago, when Gallup asked Americans whether they would want their sons -- and separately, their daughters -- "to go into politics as a life's work."
When people were asked about their sons, and then their daughters, 22% said yes to politics for their sons and 24% for their daughters -- essentially no difference between sons and daughters.
But when people were asked about their daughters going into politics, and then their sons, 41% said yes for their daughters and 42% for their sons. Again, essentially no difference between sons and daughters, but a major difference in the percentage wanting politics for their children -- a difference that depended on whether sons or daughters were mentioned first.
Why the difference? We don't know for sure, but this may be the best explanation:
Form A respondents:
- When asked about sons first, respondents treated the question as mostly about politics -- and roughly 20% wanted politics for their sons.
- Then when asked about daughters, respondents felt obligated to say the same thing for daughters as for sons.
Form B respondents:
- When asked about daughters first, roughly 40% said yes to politics -- twice the number as when the questions were asked in reverse order. Why? Perhaps because it was unusual to ask about daughters going into politics (and not sons), and that may have triggered some respondents to say yes -- not so much because they liked politics, but because they supported their daughters doing anything men can do.
- Finally, when asked about sons after daughters, respondents felt obligated to say the same thing about their sons as they had just said about their daughters.
So, which set of responses in each of these examples represents the truth? The answer: Both. Despite the differences, each set represents an accurate reflection of the public's views, but in different contexts. Polling shows what intuitively we know anyway -- for many people, opinion is highly contingent on the social context in which it is expressed.