Ambiguous Polling Question:
Starting in 1937, the Gallup Polling organization has run an occasional poll on U.S. attitudes towards the death penalty, Here's the main question Gallup asks:
Are you in favor of the death penalty for a person convicted of murder?You can see a chart of the answer over time here. I find the question puzzling, though. Consider some of the possible interpretations. It could mean:
Are you in favor of the death penalty for every person convicted of murder?Or,
Are you in favor of the death penalty for some people who are convicted of murder?Or,
Are you in favor of the death penalty for people who commit particularly heinous murders?Or,
Are you in favor of prosecutors having the option of seeking the death penalty for people who commit particularly heinous murders, if a jury first convicts and then agrees?Obviously a poll question can't capture every nuance. And perhaps most people would answer all of these questions the same way. Still, Gallup's question strikes me as unusually unclear.
Are you in favor of the death penalty for a person [you know] convicted of murder?
My advice to clients is that the data is usually only meaningful when comparing responses-over time, across brands, across demographic groups, etc.
So I have no idea how I'd answer Gallup's question.
The fact that we want to answer "sometimes" or "it depends" doesn't make the question ambiguous; it just makes it a bad question, one that turns a complex issue into "yes or no."
Hey, not all Americans share that opinion. In my home state, Wisconsin, we abolished the death penalty over 150 years ago, and haven't seen much reason to bring it back since. Even when we did have it, Wisconsin juries and judges were notorious for finding ways to avoid imposing it.
If that's a reason for supporting the death penalty, I'm surprised there aren't more Europeans in favor of it; a system where a life sentence works out to 14 years? What's up with that?
1. Are you in favor of the death penalty in any circumstance? Yes No
2. If you answered Yes, under what circumstances do you think a death penalty might be appropropriate; assume that a jury finds the defendant guilty beyond a reasonable doubt, and further finds no mitigating factors that might make a death sentence inappropriate for that defendant? Please check all that apply:
a. A murder for hire?
b. A multiple murder?
c. Murder committed in the course of a violent crime,
such as kidnaping, rape, armed robbery or a home invasion?
d. Murder of a child?
e. Murder involving torture or extreme cruelty?
f. Rape of a child?
g. Rape involving torture or extreme cruelty?
h. All intentional murders?
i. All rapes?
j. Illegal drug trafficing that results in a death?
k. Other crimes not listed?
Each time I heard that I resisted the temptation to reach for my cellphone and say something like this:
"I am a death penalty prosecutor. I have been responsible for handling a dozen capital habeas corpus cases. I have argued capital cases on behalf of the State in front of both the state Supreme Court and the Ninth Circuit Court of Appeals. I have actually been at the state prison 6 times to witness executions. On one of those occassions, I literally stood next to the executioner as he performed his duties."
"I think I have earned my capital punishment bona fides. I also think your position is insane."
The Gallup question has the same infirmities as this talk show host's "Let's kill them all" belief. There has to be a way to save the death penalty for the worst of the worst, while also expediting the process to give closure to the victims' families. Frankly, I think 12 years after the adoption of the AEDPA and the ensuing caselaw on the issue, we will see the process speed up, but only slightly.
I remember very little of my bar review class (having been admitted 16 years ago) except for one snippet of USC Professor Charles Whitebread's lectures on criminal law.
I will paraphrase the professor, but I remember him saying that the Multistate will always have a criminal law question in one of two basic fact patterns:
In one, you have a loving couple, devoted to each other. The husband has an excruciating illness. The doctors are not helping. The drugs are not helping. The husband begs his wife to kill him to put him out of his misery. She refuses. He begs again, and again--and gets weaker and weaker and is in more and more pain--until finally, out of compassion, the wife takes a gun in the house and shoots her husband dead.
In the other, there is a predator, who has raped and murdered a small child. But despite what the dead child's family believes to be overwhelming evidence, the jury inexplicably finds the perpetrator "not guilty." As the perp leaves the courtroom, he stops by the grieving parents and says, "I killed him. And I will kill again."
The father, having heard this goes home, gets a weapon, tracks down this monster, and shoots him dead.
What are these two shooters guilty of? Murder.
Why? (to directly quote Professor Whitebread) "Because it's the bar exam."
I wonder, if these hypotheticals were real, whether this radio talk show host would really think either of these two murderers deserves the death penalty.
I know I don't. And that's the problem with the Gallup question. It is overinclusive--and even though I am a death penalty prosecutor, my response to the question would have to be, "Not in every case."
I often advise my clients that it's better to keep using the imperfect question they've always used, simply because they can compare the results over a long time. Of course it's possible to calibrate a new survey question against the old, but that effort is rarely worth it unless the old question is very deeply flawed.
In this case, the ambiguity of the question strikes me as fairly minor compared to the value of the uniquely long history. There aren't very many survey questions which have been asked over 70+ years, and switching at this point would require a lot of very careful recalibration to preserve the ability to compare today's survey against the results from the 1940's.
Looking at the polling data, I am was surprised at the 7 point jump in capital punishment support in between March and December, 1972, from 50-41 support in March to 57-32 support in December. I have no idea as to what would cause that kind of spike--but the result can hardly be described as an anomoly, given both prior and subsequent polling numbers.
I am too young to remember 1972 all that clearly, but for those with long memories, what changed in 8 months?
Unlike many polling questions the question is not leading. There is a danger in adding more specifics that the question would bias the answers.
I will agree that for the answers to be truly informative it does need follow up, but the questions in a meaningful follow up would not have been the same 30 years ago as they are today.
I'd probably say no.
No, because this question is actually very ambiguous, as Orin pointed out. The only thinking people who can answer the question with a simple yes or no are those who believe the death penalty should be given to all murderers and those who think the death penalty should never be given. Fortunately, I fall into one of those categories, but most people hold a more nuanced view.
That said, it probably would be better to specify that the question is only referring to "some people who are convicted of murder."
Furman v. Georgia [death penalty is cruel and unusual] was decided on June 29, 1972
The problem is that the question is phrased more for a Lickert scale type answer (en.wikipedia.org/wiki/Likert_scale ) but the respondent is given only a minimal nominal scale to answer on (i.e. yes, no). The result is not only a less valid indicator of support for the death penalty but a statistically less valuable answer. (I'd suspect the lack of specificity also would hurt reliability, but that's another issue).
This is the biggest problem with Gallup and most other polls used in popular politics -- they usually use statistical measures that offer very little insight into the topic studied. Its not as important that x% of y believe z -- what is meaningful is that there is x correlation between believers of z and the level of y. The best you can usually do with nominal measures is to control for some other factor, but its still not something that groups like Gallup usually do for anything other than gender, party, or race.
That was my reaction also. The question seems designed to identify those who are philosophically opposed to the death penalty, such as myself, and those who are not.
I can easily answer no, and the number of people who answer no should generate a meaningful statistic.
Are people now just so much better informed on the issue than they were 50 years ago? Or do people no longer feel comfortable not having an opinion on something?
I would have to disagree with you on the appropriatness of disallowingexecution in your second example. Revenge seems like an entirely appropriate factor.
IOf course I would also say the first isn't even murder in the sense of unjustified killing. To use an even more extreme example in this vein I use the German cannibal case. The 'victim' volunteered to be killed and eaten, I cannot see even that as murder.
The biggest problem with life in prison being the maximum penalty, is what's to stop an inmate who's in fro life from killing another prisoner or a guard?
Given the procedural and substantive due process problems connected with the typical DP case, LWOPP is often a fine compromise.
LWOPP questions should be included in any DP attitudes survey that purports to be useful.
Also, the Catholic church [and many other churches] believe that it's wrong to execute even the most vile murderers. But why? If they believe what they teach, then someone on death row is in the best possible position: plenty of time for remorse and confession, then the fast track straight to heaven. Win-win. It's all good!
I think there are really two issues involved in this - one philosophical, the other practical. The first: do you believe imposing death as a punishment for the most serious crimes is morally justifiable? The second: do you trust the present government machinery to implement this policy fairly and effectively?
There's definitely some wiggle room for people who philosophically support the notion of capital punishment but do not like what they see in terms of its application.
I think the late 1972 spike mentioned by a previous poster illustrates this point. It was a backlash against the Supreme Court declaring the death penalty unconstitutional. While some polled may have had problems with the death penalty pre-1972, they were displeased with the Court's reaction.
Thank you. For some reason I didn't think of the most obvious.
Love the wonky stuff.
To elaborate on my previous post, I do think that the number of people who answer "no" generates a meaningful result. I agree with you that for people who respond with any answer other than no, little can be discerned from their responses given the nature of the question.
The biggest problem with the death penalty being the maximum penalty, is what's to stop a murderer from going on a killing spree? What are you going to do, kill him twice?
The problem that exists with any maximum sentence is that once you give it out, you can't do anything more, regardless whether that sentence is death or life imprisonment.
I am pro-capital punishment, but only if it could be exercised without error. Seeing as that is unlikely, I think the state necessarily needs to err on the side of caution for any close capital case and opt for life imprisonment.
Just like cooking, if you add too little, you can add more later. If you kill right away, you can't correct that mistake if DNA reveals that you've killed the wrong guy.
In the UK--which has abolished the death penalty--surveys at least as recent as 1999 shows 70% in favor of reinstituting capital punishment.
That European law makers do not wish to create laws to reflect popular opinion may or may not be a good thing in this case. In the general case, I find that situation to be a good excuse to avoid reliance on European or European-based laws for the interpretation of US law.
That's not quite right. I won't go into all the nuances, but basically, the Catholic Church still says that the death penalty is legitimate when necessary and when the bad doesn't outweigh the good. Then it goes on to say, with various qualifications, that the situation in the modern West is such that no case meets the test. But the death penalty is not taught as per se bad, and is not excluded for other places and other times. (In particular, the "other times" angle ensures that the Church can keep his teaching consistent with the last 2000 years of defending capital punishment -- the principle is still the same, but the application has changed.)
Also, the above are generally Catechism, i.e., binding teachings, but some of the stronger anti-penalty rhetoric came from John Paul II's personal (non-binding) writings.
Because the application of general principles to the American situation today is a matter of prudential judgment, Catholics may undoubtedly disagree with the conclusion as a matter of American policy, and still remain in good standing. So those who say that the Catholic Church is totally anti-death penalty, or that pro-death-penalty Catholics are defying their Church, are mistaken.
How would the average American's interpretation of that question have changed since 1937? Would it have been understood to mean something different by most people back then?
Unfortunately, because of the workings of the Internet, there's really no way for me to confirm that you are who and what you say you are.
However, assuming that you are in fact a prosecutor who has been involved in death penalty cases, well said.
Excuse me, but didn't a referendum on that very question win quite handily in Wisconsin? Who is the "we" who don't see much reason to bring it back?
Gallup's base question is useful only for measuring relative changes in death penalty support over time, simply because they have been asking it so long. A much better question for measuring the position of people relevant to current controversy is one they have asked for only a few years -- whether the death penalty is used too often, not often enough, or about right.
The sum of not often enough and about right -- people who support the death penalty in its present extent or want it tougher -- has run consistently close to 75%.
More on public opinion is here at C&C.
We should try very hard to see that people we put to death are guilty, but when a mistake happens, and it will, it can be justified the same way.
But, to be real value, it needs to be done quickly.
Thomas Jefferson had the right of it:
That would be true if, in fact, there was some societal benefit to the death penalty.
(Which, there isn't.)
Here is a careful, nuanced analysis of public opinion towards the death penalty in the U.S. over the last two generations.
(And no, I am not the author(s)).
"Second, as a matter of constitutional law, a majority of readers think such a law [imposing the death penalty for the rape of a child] is constitutional. Specifically, 54% of readers think such laws are constitutional, 20% thinks they are unconstitutional, and 25% doesn't know. Of course, it's not quite clear what this measures, as I didn't specify if I was asking for views on constitutionality that are normative (what the Justices should do), descriptive (what existing law requires), or predictive (what the Justices will do)."
So he's obviously well aware of the difficulties with conducting opinion polls...
Seriously, I blog semi-anonymously because the positions I take are my own—and do not necessarily reflect the views of the elected official who has the power to fire me without cause since I am exempt from civil service.
That said, I post here enough—and give enough "clues"— that someone without a life could probably unmask me. I have never lied on the VC even as I have masked, but I understand and respect your skepticism.
However, I would ask that you trust me. I am exactly who I say I am. Oh, and thanks for the compliment.
1. It is correct that the death penalty in Europe has been progressively eliminated from all states subscribing to the European Convention on Human Rights. Indeed the observer status of states such as the USA is only tolerated by way of waivers which are under attack.
2. It is equally true that European democracies are generally representative democracies where the legislators are ahead of public opinion. It is relatively easy to whip up public opinion to call for a return to executions - particularly after some particularly gruesome case -usually on the basis that incarceration is a "soft" option: "Prison is too good for these people" say blue-rinsed Tory matrons at party conference - also calling for a return to birching for juvenile delinquents.
3. In the UK we presently have a prescribed life sentence for murder. The judge fixes a minimum term 10-15-30 years and in extreme cases can be whole of life. Release on licence cannot be considered until that minimum terms is completed.
4. There is also a defence of "diminished responsibility" which, if successful, removes the need for any mandatory life sentence.
5. What has always struck me is that the appeals process regularly results in convictions being overturned - as does the subsequent fall back safeguard of our Criminal Cases Review Commission. For so long as there is the possibility of wrongful conviction, I do not see how the death penalty can ever be justified.
The public is dramatically ignorant of how death penalty cases really work. For example, I remember seeing surveys that asked what they think might mitigate against imposing the death penalty for any given murder. "Self defense" was the top answer. The lawyers reading this know that if you kill someone in self defense, it's not even a crime.
Actually, I agree with Edmund Burke, "Your representative owes you, not his industry only, but his judgment; and he betrays instead of serving you if he sacrifices it to your opinion."
However, Burke is also quoted as saying, "In effect, to follow, not to force the public inclination; to give a direction, a form, a technical dress, and a specific sanction, to the general sense of the community, is the true end of legislature."
I tend to agree with the former more than the latter. That said, the arrogance of presuming that legislators are necessarily "ahead of public opinion" is astounding.
our hypos remind me of a B.C. comic strip I once saw, that was something close to this:
- "I'd like you to do a Presidential poll"
- "What results do you want?"
- "Huh?"
- I need to know so I can ask the right questions"
Quite apart from anything else public opinion can be manipulated - how else did George W. Bush get re-elected ?
Says the "Dog"
Don't you remember whom he ran against? Sometimes that makes re-election much easier.
Gene will recall that from his paper on 'n'.
I once explained to a friend why murderers should be put to death, while unborn lives should be protected. Murderers are guilty, while the unborn are innocent. This seemed to come as a revelation to him.
I got a call asking if I trusted General Motors.
Trusted them to do what? For what? About what? No, said the caller, just, "do you trust them?"
I eventually declined to answer, it being a dumb question.
However, it is absolutely certain that the number of answers one way or another were used to demonstrate some absolute something or other.
Nuts.
How many polls on the public's view of Bush's conduct of the war ask if the individual being asked thinks it should be conducted far more actively and aggressively? Nope. Any answer other than "I think it's great!" is taken to mean opposition to the war.
I'd like to see a clear poll question, but my guess is that those commissioning the poll wouldn't want it.
I think this is largely, unfortunately, incorrect. If you monitor answers to the same vague question across groups, you may just see how different groups interpret a vague question. If you track answers over time, you may just see how people's understanding of a vague question has changed over time.
That is, the vagueness of the question inserts a variable that can correspond with, or not correspond with, how people feel about the thing you are trying to measure. At a minimum, it odds noise. At worst, it adds noise that correlates with or against the thing measured.