This explains a lot. The questions asked in any survey have a lot to do with the answers received. The CBS poll showing that “evangelicals” oppose the president’s position on the Schiavo case is no different.
Suppose a patient is in a coma, doctors say brain activity has stopped and the patient is getting food and water through a feeding tube. Should a close family member have the right to tell the doctor to take away the feeding tube and let the person die, or not?
Terri Schiavo is not in a coma and her brain activity has not stopped. This question is deceptive and leads those who haven’t taken the time to read the particulars of the case to believe that Terri Schiavo is brain-dead.
This question is followed by two more about patients in “a coma or persistent vegetative state”. Then comes question #13, which begins:
Terri Schiavo has been in a persistent vegetative state since 1990.
Really? Not according to a fair number of experts. Although the courts have ruled otherwise, there is substantial disagreement on this point. It is critical to how respondents to the poll answer the question and the questions following regarding whether Terri Schiavo should be starved to death. Take away that assertion and you change the tone of the question–and probably the answer you get.
In addition, the survey consistently uses the word “let” when “make” is more accurate: “Let her die” rather than “make her die”.
Terri Schiavo wasn’t dying until she was denied food and water by court order. If I lock my teenage daughter in a room without food and water, is that “letting” her die? Do you think a court would accept that as my defense?
Not on your life. But Terri Schiavo is different, isn’t she? She can’t move or ask for help. In America, that means her legal guardian can “let her die” with impunity.
Excuse me while I puke.
If CBS had phrased the question, “should Terri Schiavo be starved and dehydrated until she is dead”, I have no doubt the results would have been very different.
Then there is the issue of sample size. 737 total respondents are judged sufficient to gauge the mood of the nation. For a local area, maybe. I’d like to know the geographic dispersion of the sample. It can’t have been too great, because men were underrepresented; the 294 responses from men were weighted to equal 354 in the calculation of the percentages.
Here’s the kicker: The key questions about whether Terri Schiavo should be allowed food and water were only answered by a “partial sample” of 321 respondents.
Look: When I was the music director of KZ-93 in Peoria, Illinois, we’d survey about 300 people a week for music preferences. Anything less than that was considered statistically insignificant. For Peoria. A metro area of about 325,000.
And, since we were a Top 40 station, we were only interested in teens (12-17) and young adults (18-34). If we oversampled one demographic, our callers kept dialing until we got enough responses in each key age group to give us a fairly scientific survey.
Instead, CBS weights its responses. In other words, one person in an undersampled group counts for more in the final analysis than one in an oversampled group.
How many people surveyed were “evangelicals”? 26%, about 192. (22% of men, whose surveys were given more weight than womens’.) Definitely not a statistically reliable measure of American evangelicals, no matter how you parse it.
In short, I don’t trust the survey. Sample size is too small, and the survey questions were slanted.
Don’t buy the media spin. There’s no way 2/3 of self-described American evangelicals prefer to see Terri Schiavo condemned to die of thirst.