I thought it might be worthwhile to explain why this less definitive conclusion is drawn from the research.  This less definitive conclusion is drawn not only by those who support the death penalty but also by those who strongly oppose the death penalty.  The latter would clearly like to have a stronger conclusion, but governed by their reason, as they are, they feel obliged to limit themselves to what the argument supports.

Let me first put the material that I read to you before us.
 

Abolitionists attempt to disprove these hypotheses by amassing statistical evidence to demonstrate that there is no correlation between criminal activity and the existence or nonexistence of a capital sanction. . . . Thorsten Sellin, one of the leading authorities on capital punishment, has [amassed evidence which indicates that the propositions which would follow if the death penalty deters prospective murderers are not true.] . . .
    . . . This evidence has its problems, however. One is that there are no accurate figures for capital murders; there are only figures on homicides and they, of course, include noncapital killings. A second problem is that certain murders undoubtedly are misinterpreted as accidental deaths or suicides, and there is no way of estimating the number of such undetected crimes. A third problem is that not all homicides are reported. Despite these difficulties, most authorities have assumed that the proportion of capital murders in a State's or nation's homicide statistics remains reasonably constant, and that the homicide statistics are therefore useful.
   Sellin's statistics demonstrate that there is no correlation between the murder rate and the presence or absence of the capital sanction. He compares States that have similar characteristics and finds that irrespective of their position on capital punishment, they have similar murder rates. In the New England States, for example, there is no correlation between executions and homicide rates. The same is true for Midwestern States, and for all others studied. Both the United Nations and Great Britain have acknowledged the validity of Sellin's statistics.
    Sellin also concludes that abolition and/or reintroduction of the death penalty had no effect on the homicide rates of the various States involved. This conclusion is borne out by others who have made similar inquiries and by the experience of other countries.
    Statistics also show that the deterrent effect of capital punishment is no greater in those communities where executions take place than in other communities. In fact, there is some evidence that imposition of capital punishment may actually encourage crime, rather than deter it. And, while police and law enforcement officers are the strongest advocates of capital punishment, the evidence is overwhelming that police are no safer in communities that retain the sanction than in those that have abolished it.
    There is also a substantial body of data showing that the existence of the death penalty has virtually no effect on the homicide rate in prisons. Most of the persons sentenced to death are murderers and murderers tend to be model prisoners.
    In sum, the only support for the theory that capital punishment is an effective deterrent is found in the hypotheses with which we began and the occasional stories about a specific individual being deterred from doing a contemplated criminal act. These claims of specific deterrence are often spurious, however, and may be more than counterbalanced by the tendency of capital punishment to incite certain crimes.  (from CONSTITUTIONAL LAW: Civil Liberty and Individual Rights, by William Cohen and John Kaplan, Second edition.)

We might start by characterizing the logic of the argument implicit in the Sellin study in the following way:
 

1. If the death penalty is a significant deterrent, then those states that have the death penalty have a lower incidence of capital offences than those states without the death penalty.

2. Those states with the death penalty do not have a lower rate of capital offences than those states without the death penalty.
----
3. The death penalty is not a significant deterrent.

Premise 1 captures a relationship between the death penalty and an impact that some think the death penalty will have.  Premise 2 reflects the actual data (regarding this impact) that has been collected.  From just this much, one might draw the conclusion expressed in 3.

This argument can appear to be convincing, but it shouldn't be.  The argument sholdn't be convincing, as it stands, because, while it is valid (the conclusion follows from the premises), the truth of premise 1, as it stands, is questionable.  It is questionable in a way that can be made clear by considering the following clause that should be added to the premise:
 

‘unless there is some other factor that is driving the incidence of capital offences up in spite of the death penalty’.

1.1. If the death penalty is a significant deterrent, then, unless there is some other factor that is driving the incidence of capital offences up in spite of the death penalty, those states that have the death penalty have a lower incidence of capital offences than those states without the death penalty.

From this premise and premise 2 all we can conclude is

3. Either the death penalty is not a significant deterrent or there is some other factor that is driving the incidence of capital offences up in spite of the death penalty.

In order to conclude that the death penalty is not a significant deterrent an additional premise is needed:

4. There is no other factor driving the incidence of capital offences up in spite of the death penalty.

To establish premise 4, one would have to decide (a) ‘there is no such factor’ and (b) ‘There is such a factor, but we have not found it yet’.  The difficulty in ruling out (b) with respect to the complex social environments under scrutiny is enormous. The difficulties here are part of what forces the conclusion that there is no evidence that the death penaly deters rather than the conclusion that the death penalty does not deter.

This is not all there is to the difficulites.  We have considered only one possible factor that should be expressed in a qualifying clause of the sort that changes premise 1 to 1.1.  There are other factors that must be considered.  Among these is the accuracy of the statistics reporting the incidence of capital offences.  Unless we are sure that we have an accurate count of the capital offences in the states under comparison, we don't know that the rates of capital offences are the same. There are numerous complications as a result of different interests at work in the reporting efforts.  (I read recently that the alleged drop in crimes of certain sorts in New York city is wholly the result fo creative reclassification and repriting and not the result of any increased success in the crime-fighting activities in the streets.  Is this true?  I don't know.  But it is relevant)

And so on,....

Because of the many, many other factors that make such research enormously complex and, hence, definitive conclusions all but impossible.  So, where reason is the order of the day, the more tentative conclusion is all that is warranted.  Keep in mind that the more tentative conclusion is not nothing.  To the extent that the supporters of the Death Penalty bear the burden of proof, the absence of support for greater deterrence hurts their efforts.