PRINCETON NJ -- Gallup is now reporting three different death penalty questions, with a spread in responses between 81% in favor of the death penalty at the top end to 52% at the bottom end.
Here they are:
- 65% support the death penalty in cases of murder
- 52% favor the death penalty for murder in contrast to the explicit alternative "life imprisonment with no chance of parole"
- 81% favor the use of the death penalty in the case of Timothy McVeigh
Does this mean that the American public is mixed up or confused? No. This is an excellent illustration of how much we can learn by examining the different responses of the public when we ask questions about a topic in different ways.
This basic point is well worth emphasizing. Most humans do not have filing cabinets in their brains from which they deftly select a definitive response when asked about a topic or policy issue. Most humans have developed basic patterns of reactions to stimuli that we generally call attitudes. When we ask about a specific topic, the human has a tendency to respond in a given way, but the precise nature of that response is often dictated by the circumstances in which the question is asked (including, for example, the gender and race of the interviewer) and the exact nature of stimulus words included in the question.
Thus, it is clear that humans have a more complex set of attitudinal responses surrounding a given topic than is often ascribed to them. Our research on abortion attitudes, for example, shows that a significant number of Americans put themselves into a middle category -- they favor restrictions on abortion, but are not in favor of either a total ban on abortion, nor making abortion totally legal in all circumstances. What this means: the response to a question about abortion is dependent on exactly how the question is phrased and more specifically on what the question says about the circumstances in which the abortion could take place.
Financialengines.com has made a name for itself in displaying the probabilities of certain financial outcomes based on different scenarios. One’s net worth upon retirement, the site shows, can be very roughly estimated as a series of probable outcomes, each with associated probabilities -- rather than a particular dollar figure. In similar fashion, I think it is useful, at least heuristically, to look at an American’s attitudes about a given topic as a tendency or probability of responding in a certain way. Responses to a topic tend to be focused in one direction or the other, but the exact nature of the response is dependent on circumstances. Generalized across the American population, a precise estimate of a population parameter depends strongly on the nature of the question and the environment in which it is asked. Thus, there is almost never just one number that adequately represents the population’s position on a topic of interest. There is a range of responses.
In some instances, there is less need to stress the range of probable responses because the question is worded only one way and there is only one environment in which it will be asked -- the voting booth. In this situation, we don’t care as much about the spread of responses as we do about the response to one specific question. A good example of this is the 1998-99 impeachment procedures against Bill Clinton. At one point the House, and then the Senate, was faced with just one question: a yes or no vote first on impeachment, and then on conviction. But normally, things aren’t that neat.
So, what we try to stress here at Gallup is a more holistic approach to understanding attitudinal structures. There is often a range of opinion on an issue, and -- as is true for any science -- our task is to represent that range in a way that allows us to best understand what is going on.
Indeed, we learn a great deal when we examine the variance in responses based on the structure or environment in which a question is asked. In 1990 and 1991, when the United States was leading up to its involvement in the Gulf War, for example, we quickly learned that a significantly higher percentage of Americans would support "going to war" if the question stressed that it was a joint effort of many countries under a U.N. mandate. We thus learned that Americans were hesitant about the United States instituting a war by itself, but that they found reassurance in their country joining with other like-minded nations.
Let’s get back to death penalty attitudes. The Timothy McVeigh case has put the public’s approach to the ultimate sanction very much in the news. (The issue has bubbled to the surface in recent years for other reasons, including the impact of DNA evidence in uncovering errors in the application of the death penalty, Illinois’ moratorium on the death penalty, and George W. Bush’s position as former chief executive in the state that leads the nation in its number of executions.) Gallup’s traditional question on the death penalty was first asked back in 1930s: "Are you in favor of the death penalty for a person convicted of murder?" Note that this question offers the respondent no alternative option for the death penalty, just yes or no. This is the question that Gallup has asked almost 30 times in over 60 years. It is quite labile, with a low point of 42% in 1966, and a high point of 80% in 1994. In the poll we finished on May 14, 65% responded "yes," a number consistent with five polls conducted over the last year.
This represents what I would call the basic response. Without much elaboration, without arguments for or against, without giving the respondent any alternative, the top-of-mind, quick response of about two thirds of Americans is to say yes to the death penalty.
Now consider another version of the death penalty question we have used in recent years: "If you could choose between the following two approaches, which do you think is the better penalty for murder -- [ROTATED]: the death penalty (or) life imprisonment, with absolutely no possibility of parole?" Here we find that the number choosing the death penalty alternative, as we noted above, drops to 52%, while 43% choose the life imprisonment alternative.
But that’s not all. As noted above, we get a higher affirmative response when we get down to specifics -- namely the Timothy McVeigh case. Here, we observe a jump of 16% points over the baseline 65% number, up to 81%.
The reasons behind this higher affirmative response are not a great mystery. The McVeigh case removes several of the objections that people ordinarily might have to the death penalty. First, there is little possibility that its application would kill an innocent man, since McVeigh has essentially confessed to committing the crime in interviews he has given to reporters. Second, McVeigh himself has indicated that he wants to die, and -- at least before the latest FBI mix-up -- waived his right to further appeals in order to speed the process. And, third, the crime McVeigh committed was of the most heinous variety -- murdering 168 innocent people, including women and children with whom he had no prior connection.
It is obvious, then, that some Americans who are opposed to the general legality of the death penalty are willing to make an exception in specific instances. In fact, Gallup polls suggest that about half of those who ordinarily express opposition to the death penalty end up favoring it in the case of McVeigh.
What we really have here is some very interesting social scientific insight.
If there were a national referendum on the death penalty, it appears that the public would vote "yes" in almost all situations, but our analysis of these responses suggests that the precise margin of the outcome might well depend on the nature of the referendum question, and on how alternatives to the death penalty were phrased in the course of the campaign.
We can go further. We know from our open-ended questioning that those who favor the death penalty give as their number one reason the fact that it provides a sense of fairness and justice -- an eye for an eye -- a fitting punishment. We can hypothesize further that many Americans may favor the death penalty out of a fear that the criminal may avoid paying an appropriate price -- that is, getting a parole or wiggling out of punishment in some other way. The "life in prison with no possibility of parole" option sounds like it eliminates some of this wiggle room, and thus apparently satisfies some people’s objections to getting rid of the death penalty. Furthermore, many Americans may oppose the death penalty because they feel that innocent people may be convicted and executed – their innocence discovered only when it’s much too late. There are also concerns that minorities and others who cannot afford fancy legal help are disproportionately placed on death row. The McVeigh case, as we have noted, removes many of these objections, and that is why we find a higher percent willing to go along with the death penalty in this particular case.
We thus end up with a sociological hypothesis: Americans’ support for the death penalty will increase to the degree that it is made clear that no mistake has occurred and that the death penalty is being applied to a truly guilty person. One corollary to this hypothesis: support may actually end up increasing in the years to come, rather than decreasing, as new DNA testing techniques become widely used. Why? Because, while this evidence may prove that some on death row are in reality innocent, it also may reduce any chance of a mistake in future cases.
All of this we learn from a careful analysis of different question wordings that, rather than confusing us, can often help us arrive at a better understanding of the phenomenon under investigation.