The Death Penalty: Questionable Evidence for Deterrence

The death penalty is about as divisive an issue as abortion or gun control. Studies are offered by proponents and opponents to both support and condemn both sides of each issue. It is therefore imperative to assess and critique the quality, reliability and validity of the studies.
This post was published on the now-closed HuffPost Contributor platform. Contributors control their own work and posted freely to our site. If you need to flag this entry as abusive, send us an email.

The death penalty has become an issue in national politics again with GOP candidate Rick Perry's position on its use, and in recent news with the Troy Davis case in Georgia where there is reasonable doubt of the accused and as such, the increasingly likely prospects of executing innocent persons.

The death penalty is about as divisive an issue as abortion or gun control. Studies are offered by proponents and opponents to both support and condemn both sides of each issue. It is therefore imperative to assess and critique the quality, reliability and validity of the studies.

Evidence for Deterrence?

There are studies that don't find a relationship between the death penalty and murder rates. These are largely done by criminologists and sociologists. Research can't prove a negative (we can't prove a null hypothesis; we reject a null hypothesis); research can technically only find no relationship. So the burden of proof is not on proving the death penalty doesn't deter, the burden of proof is on finding that the death penalty does deter.

Pro-death penalty activists argue that there are several studies that find deterrence in the death penalty. However, some researchers have reviewed these studies and there are serious problems with the assumptions and/or methods of these studies.

For example:
1) The sample sets are too small to conclude any statistical power, and there are no control groups (or no meaningful control groups). Also, since the sample sets are too small to draw any conclusions, any findings are considered statistical artifact. Not only are the numbers very small, the effect is very small, too, suggesting that just a few people might be saved by the execution of one person. (This is reminiscent of a dialogue that goes something like - a proponent said: "If this intervention saves just one person, then the intervention was worth it"... to which the pragmatist said: "Statistically speaking, if just one person was saved it was due to chance, not the intervention.") This doesn't pass the sniff test.

2) One of the studies that shows deterrence even dismisses all other studies indicating that the sample sets are too small to draw conclusions about deterrence. While this study undermines/dismisses all other studies as having too small sample sets to draw any conclusions from (a point ignored by pro-death penalty activists), this study is itself undermined by the fact that it did not rule out all third variables and several other concerns.

3) They are largely regression analyses, which is like correlation on steroids. Correlation isn't causation. While regression is a stronger test than correlation, if regression doesn't consider the right variables the findings are useless. In the case of each of these studies none of them show that the death penalty is the only reason someone would be deterred from killing another person. This is a serious and major methodological problem in that the presence of a third variable completely undermines the conclusion that the death penalty deters. Eliminating "all" confounding variables is a rule of regression analyses that must be followed to infer causality. It is a rule that has been violated in every study.

4) Most of the studies are done by economists who are notorious for incorrectly making assumptions about human behavior, especially criminal behavior. The economic models used to find deterrence assume rational behavior. Furthermore, economists are not trained in criminology - this is an argumentum ad verecundiam (an appeal to an unqualified authority). It makes about as much sense for an economist to do crime analysis as it does for a criminologist to do economic analysis: both are out of their field of expertise. This is not to say that just because they are an economist they did their analysis wrong, but the fact that these economists are out of their field of expertise is supported by the notion that these studies were largely published in economics journals (which have editors who don't understand the nuances of crime analysis or criminal behavior), or second tier criminal justice journals.

5) The best way to determine deterrence is through a series of randomized controlled experiments designed to test the effect of the death penalty on deterrence. Such a study doesn't exist and won't exist for obvious legal, political and ethical reasons. Regression has serious limitations to drawing causality: it must account for 1) all third variables-these studies don't; 2) time order-which these studies can control for, and 3) spuriousness-which these studies don't account for... they can't due to the lack of small numbers.

One of the great things about published research is that it can be reviewed and accepted or dismissed. Just because something is published, it doesn't mean that the findings are fact. The scientific method and published research is ever evolving. Evidence is alive and changes based on new methods and new research.

Considering that there are so many problems with the death penalty and research surrounding the death penalty, the prudence of the death penalty is easily called into question.

Third Variables?

It is illogical to assume that anyone is deterred by 'only' the death penalty. Clearly, finding such a person would probably be impossible.

For example, who would stand up and say: "the only reason I am not going to kill someone in a heinous way is because I fear the death penalty?" Ask this hypothetical person if that is the only reason and they will say 'no'; they could also say that they don't want to lose their freedom, that they don't want to tarnish their reputation, or that they don't want to upset or disappoint their family or friends, or that they don't believe in killing another person, or any number of other reasons.

It strains credulity to believe that the death penalty is the only reason someone isn't going to commit a heinous murder ultimately rendering the use of the death penalty as a deterrent meaningless.

Philosophical Assumptions?

One of the worst arguments that states that the death penalty deters is: All prospects of a negative outcome deter some. It is a truism. The death penalty, the most severe of criminal sanctions, is the least likely of all criminal sanctions to violate that truism.

This line of reasoning is not based on any evidence or research; it is a philosophical assumption - an ipse-dixitism (a dogmatic assertion). It does not prove anything so it is also somewhat of an argumentum ad ignorantian (which incorrectly asserts that a proposition is true because it has not been proven false (or vice versa). This argument is also circulus in probando, i.e., the death penalty is severe because it deters and it deters because it is severe (which is circular reasoning). This reasoning is also petitio principii (a proposition which requires proof is assumed without proof). In short, this argument makes logical fallacies.

Furthermore, research shows that certainty of punishment is more important than severity or celerity. It doesn't matter how severe the punishment is to the person who doesn't think s/he will get caught.

The above argument also assumes that the death penalty is the most severe sanction to everyone, which is an assumption. For some people, there may be worse sanctions including life without the possibility of parole; including the shame and disappointment that their own life has become; etc. Considering that almost 25% of inmates who committed suicide in prison had a life sentence, the death penalty seems to have been the preferred sentence. For example, Harold Shipman was a convicted murder sentenced to serve a life sentence. He committed suicide after about serving 4 years of his life without parole sentence. Just as there is no single best reward for all people, there is no one single worst sanction to all people; people have individualized and differing preferences. Do people appeal the death penalty? Yes, it is a matter of legal action, not necessarily fear. It is rare to hear of people screaming and yelling in fear as they are being executed, or every day while on death row. Admittedly, sometimes inmates do scream and yell but for reasons unrelated to their sentence.

This line of thought also assumes that the only reason that a person won't kill is because they would receive the death penalty. This is absurd. There is no reason to believe that someone won't kill only because they might get the death penalty. As noted above, anyone who fears the death penalty is most likely going to fear other, less harsh sanctions (rendering the death penalty irrelevant in terms of deterrence) because the have a stake in conformity. It could be: their position in society; their family; their better life; the realization that killing someone just isn't worth it, etc.


There is also the argument that the death penalty is justice because it is legal. This is circulus in probando. This says that something is legal because it is right and something is right because it is legal. It is also petitio principii.

This claim of justice means that what is justice in one state is not justice in another. By this account, justice is subjective. I am not saying the death penalty is or isn't justice served; I am just pointing out that just because something is permitted it doesn't necessarily mean it is justice.

Texas Governor Rick Perry was right when he said on 7 September 20011 that we should err on the said of caution when human life is at risk... too bad he said this in response to providing HPV vaccines for 12 year old girls, and not the death penalty since there is good evidence to show that at least one innocent person was executed on his watch as governor. Not everyone finds that the dealth penalty is justice. Arguments that claim that the families of victims support the death penalty is not necessarily true.

In Conclusion

One of the most important elements of crime prevention is legitimacy in the justice system. The legitimacy of the justice system is already in question for many people because of the use of the death penalty. This is made worse by the idea that an innocent person would be executed in America has a perverse effect in undermining the legitimacy of the justice system and consequently, undermining crime prevention. The execution of Troy Davis will probably result in a perceived loss of legitimacy in the criminal justice system, and possibly less safety and security for citizens.

Like gun control or abortion, the death penalty issue likely won't ever be solved with data or by point-counterpoint. The death penalty is a highly emotional issue; a political issue; and a moral issue. Each of these three points - emotion, politics, and morality - are subject to the human experience and so they vary from person to person, culture to culture. Ramming data or flaws in data at someone may sway moderates or people who haven't thought much about the issue, but it won't sway people who are very passionate about one side or another.

PAUL HEROUX previously worked in a prison and in jail. He holds a Master's in Criminology from the University of Pennsylvania and a Master's in Public Administration from the Harvard University JFK School of Government. Paul can be reached at

An earlier version of this post incorrectly stated that Harold Shipman was sentenced to serve in Pennsylvania.

Go To Homepage

Before You Go

Popular in the Community