Murder and the Death Penalty William C

Journal of Criminal Law and Criminology

Volume 65 | Issue 3 Article 8

1975 Murder and the Death Penalty William C. Bailey

Follow this and additional works at: https://scholarlycommons.law.northwestern.edu/jclc Part of the Criminal Law Commons, Criminology Commons, and the Criminology and Criminal Justice Commons

Recommended Citation William C. Bailey, Murder and the Death Penalty, 65 J. Crim. L. & Criminology 416 (1974)

This Criminal Law is brought to you for free and open access by Northwestern University School of Law Scholarly Commons. It has been accepted for inclusion in Journal of Criminal Law and Criminology by an authorized editor of Northwestern University School of Law Scholarly Commons. THE jouR AL or Canm;AL LAw & CR=OLOGY Vol. 65, No. 3 Copyright © 1974 by Northwestern University School of Law Printed in U.S.A.

RESEARCH NOTES

MURDER AND THE DEATH PENALTY

WILLIAM C. BAILEY*

A survey of the literature on homicide and capi- methodological shortcomings.n Before examining tal punishment reveals that the past decade has these shortcomings and the scope of the present produced no new research on this question. Appar- investigations, it is necessary to review the avail- ently, the early investigations by Bye,' Suther- able evidence. land,'2 Kirkpatrick,8 and Void 4 and later examma- tions by Sellin, 5 Schuessler6 and Savitz7 have Previous Research convinced most students of homicide that the The conclusion that capital punishment has no ineffectiveness of the death penalty as a deterrent deterrent effect on murder stems primarily from to murder has been demonstrated conclusively.8 three types of investigations: (1) comparative Not all remain convinced, however, of the con- analyses of homicide rates for states which differ clusiveness of the evidence. In a recent examina- in provisions for the death penalty; (2) longitudinal 9 tion of the question, Bedau argues that most investigations of homicide rates for states before criminologists skeptical of capital punishment have and after the abolition and/or restoration of the not come to this conclusion by a critical examina- death penalty; and (3) longitudinal examinations tion of the evidence, but rather because of their of homicide rates immediately preceding and adherence to a general theory of violent crimes immediately following publicity of executions.u that excludes the influence of the threat of punish- The most common approach to testing the de- ment.' Furthermore, careful examination of the terrent effect of the death penalty has been a com- literature reveals the evidence usually cited as parison of homicide rates of abolitionist and reten- questioning the death penalty to be less than con- tionist states 3 These investigations have generally clusive. With few exceptions, these investigations shown homicide rates in the latter states to be suffer from a number of serious theoretical and two to three times that of the former 4 This find- * Department of Sociology, Cleveland State Uni- ing is contrary to what deterrence theory would versity. predict. Such comparisons have usually been de- IR. BYE, CAPITAL PUNISHMNT IN THE UNITED clared invalid, however, for the two groupings of STATES (1919). 2 Sutherland, Murder and the Death Penalty 15 j states are not uniform with respect to other possible Am. INST. ConmAL LAW & CRI [NoLoGY 522 (1925). important etiological factors-population com- 8C. KIREPATIICK, CAPITAL PUNISHMENT, COM- position, social structure and cultural patterns.15 MITTEE ON PnITh Ropic LABOR Or PHIADELPHIA YEARLY MEETING or FRIENns (1925). h E. SuTnERLAND & D. CREssEY, PRINcIPLEs or 4 Vold, Can the Death PenaltyPrevent Crime, 3 PRISON Cm NOLOGY (7th ed. 1970); F. ZUIRING & G. HAw- 8 (1932). EINS, DETERRENCE: LEGAL TAREAD I. 5 IN CRIME CONTROL T. SELLIN, CAPITAL PUNISHMENT (1967); Testi- (1973); Bedau, Deterrence and the Death Penalty: A mony of Thorsten Sellin before Royal Commission on Reconsideration, 61 J. Can. L.C. & P.S. 539 (1970); Capital Punishment, ROYAL COme'N ON CAPITAL Gibbs, supra note 8; Erickson & Gibbs, The Deterrence PuNIssmoET, CmD. No. 8932, at 17 (1955). [Hereinafter Question: Some Alternative Methods of Analysis (1972) cited as Sellin, ROYAL Comme'N]. (unpublished article). 6 Schuessler, The Deterrent Influence of the Death "Sellin's examinations of police safety and prison Penalty,7 284 ANNALS 54 (1952). homicides and the death penalty are also often cited in Savitz, A Study of Capital Punishment, 49 J. Cpm. discussions of capital punishment as a deterrent to L.C. & P.S. 338 (1958). homicide. SEU.m, CAPITAL PUNISRTA NT (1967); s Gibbs, Crime, Punishmentand Deterrence,48 SOUTH- SELLwN, ROYAL Coma'N, supra note 5. wEsTERN SocIAL SCrENCE Q. 294 (1965). "E. SuTnmRmLa- & D. CRnssEY, supra note 11. 9 Bedau, Deterrence and the Death Penalty: A Recon- 14Id. See also Schuessler, supra note 6; Sutherland, sideration,61 J. Cam. L.C. & P.S. 539 (1970). supra1 note 2. 10In addition, McClellan points out that much of the 6 Schuessler, supra note 6; Sutherland, supra note 2. evidence on the deterrence issue is questionable for it Schuessler points out that this criticism indirectly would appear to have been collected for the sole purpose affirms that the relative occurrence of murder is the of disproving the value claimed for punishment. G. result of a combination of social circumstances of which MCCLELLAN, CAPITAL PUNISHMENT (1961). punishment is only one. 19741 THE DEATH PENALTY

To meet this objection, Schuessler16 and Sellin" study and clinical observationsO have brought compared homicide rates of abolitionist states most criminologists to what Selln has termed the with neighboring capital punishment jurisdictions. "inevitable conclusion," that the death penalty 5 These investigations have consistently led re- has no discernible effect as a deterrent to murder. searchers to one of two conclusions: abolitionist Deterrence theory suggests that if punishment states have slightly lower homicide rates than is to act as an effective deterrent to crime it must their death penalty neighbors"8 or that it is im- be: (1) severe enough to outweigh the potential possible to differentiate capital punishment from pleasures crime might bring; (2) administered with abolitionist states by solely examining homicide certainty; (3) administered promptly; (4) admin- rates. 9 Furthermore, examinations of the relation- istered publicly; and (5) applied with the proper ship between the risk of execution in retentionist judicial attitude.26 Typically, however, only one states and homicide rates have shown no discern- aspect of capital punishment-its severity-has ible correlation between these two factors. 0 been examined as a deterrent to murder. Little Comparative examinations of homicide rates attention has been paid to the certainty of the before and after abolition, and in some cases, the death penalty, with examinations of the remaining restoration of the death penalty, have also ques- three aspects of punishment being completely tioned the efficacy of capital punishment. These absent from the literature. In short, the question investigations reveal that states that have abol- of the death penalty as a deterrent to murder has ished the death penalty have generally experienced only been examined in the most narrow theoretical no unusual increase in homicide. Moreover, the sense. Deterrence theory has simply never been reintroduction of the death penalty (eleven states tried and given a "fair chance." 21As Jeffery states, have abolished the death penalty but later re- "The lesson to be learned from capital punishment stored it) has not been followed by a significant is not that punishment does not deter, but that decrease in homicide."1 the improper and sloppy use of punishment does Another source of evidence questioning the effec- not deter ...., 28 tiveness of capital punishment has come from Of methodological concern, each of the above investigations of the effect that publicity of execu- studies rests upon a number of assumptions, some tions has on homicide rates. Dann's early analysisn of which appear highly questionable.29 These pri- of homicide rates in Philadelphia sixty days pre- 24J. GnLsn, THE WIscoNsIN PRISONER (1946); L. ceding and following the mass execution of five LAWEs, TWENTY THOUSAND YEARS IN SING SING killers revealed no significant difference in rates (1932); K. MENNINGER, THE CRIME OF PUNISHMENT before and after this highly publicized event. Simi- (1968); S.PALmER, A STUmY or MUDER (1960); T. THoMAS, Tis LIFE WE TAEE (1965); Cuthbert, A larly, in a more recent investigation in Philadel- Portfolio of Murders, 116 BRIT. J. or PSYCHIATRY 1 phia, Savitz found no significant difference in the (1970). 25As Bedau notes, a review of Sellin's writings on rate of capital crimes eight weeks before and eight the death penalty, which span 1953-1967, reveal a weeks after the well publicized sentencing of four certain vacillation in the conclusions he draws. At men to death." times, he categorically denies the death penalty as a deterrent to homicide, and at other times he denies it In sum, the above investigations as well as case as a superior deterrent to life imprisonment-two quite "6Schuessler, supra note 6. different conclusions. Bedau, Deterrence and the Death 7 T. SE N, CAPITAL PUmsHENT (1967). Penalty: A Reconsideration, 61 J. CxRi. L.C. & P.S. 539 18E. SUTHERLAND & D. CREssEY, supra note 11; (1970). Sutherland, supra note 2. '6 C. BECCARIA, AN ESSAY ON CRIMES AND PUNISH- mENT '9W. CHAmBLiss, CRnIE AND THE LEGAL PROCESS (1809); J. BENTHEMu, PxN'ciPEs or PENAL LAW (1969); T. SELEIN, CArrrAL PuIs,s~mNT (1967); (1843); E. Ross, SOCIAL CONTROL (1901). Schuessler, supra note 6. 2E. PuT AmER, ADmmisTRATioN or CRaNrAL 20 H. BEDAU, THE DEATH PENALTY IN AMERIcA LAW (1953). (1967); T. SELLIN, CAPITAL PUN1SHENT (1967); 2 Jeffery, CriminalBehavior and the Learning Theory, Schuessler, supra note 6; Sellin, ROYAL COm'N, supra 56 J. Crxm. L.C. & P.S. 294, 299 (1965). note 5. "Bedau lists four such common assumptions: (1) "H. BEDAu, THE DEATH PEANLTY IN AmERICA homicides as measured by vital statistics are in a gen- (1967); H. MATTICK, THE UNEXPLAINED DEATH: AN erally constant ratio to criminal homicides; (2) the AwALYsis or CAPITAL Pums=uxNT (2d ed. 1966); T. years for which the evidence has been gathered are SELLIN, THE DEATH PENALTY (1959); Schuessler, supra representative and not atypical; (3) however much note 6. fluctuations in the homicide rate owe to other factors, 2Dann, The Deterrent Effect of Capital Punishment, there is a non-negligible proportion which is a function FImNDS SocIAL SERviCE SERIrS, Bulletin 29, Third of the severity of the penalty; and (4) the deterrent Month2 3 (1935). effect of a penalty is not significantly weakened by its Savitz, supra note 7. infrequent imposition. Bedau, Deterrence and the Death WILLIAM C. BAILEY [Vol. 65 marily concern the adequacy of using available necessary to accept the view of experienced crim- aggregate homicide statistics, issued by the Federal inologists35 that available homicide statistics per- Bureau of Investigation and the Public Health mit an adequate test of the effect of the death Service, as an index of murder in examining the penalty.'6 This is a regretable situation because so effect of the death penalty. 0 much of the deterrence debate over death penalty In the United States, generally only one type of turns on the validity of this assumption. 7 Clearly, homicide-murder in the first degree-is punish- additional research is needed in this area. able by death, with murder in the second degree and voluntary manslaughter usually being pun- The Present Investigation ished by imprisonment." Typically, however, investigations of the death penalty have opera- The research reported in this article is a further tionally defined premeditated murder as homicide, examination of the relationship between homicide a much more inclusive offense category. This and capital punishment. The approach is similar s 9 practice has been necessitated by the fact that no to that of Schuessler- and Sellin's" with one im- alternative statistics are currently available on a portant exception: the murder data examined here nationwide basis that break down homicide by permit a direct rather than indirect assessment of type and degree. As a result, investigators have the relationship between capital homicides and been forced to make a large and possible erroneous the death penalty. assumption whether they use police or mortality To avoid the above difficulties and obtain theo- statistics, that the proportion of first degree mur- retically appropriate data on first degree murder, ders to total homicides remains constant so that a survey was conducted of all State Bureaus of statistics on the latter provide a reasonably ade- Corrections throughout the United States. In- quate indicator of capital offenses. quiries were made to each agency requesting figures Most investigarors have been quick to accept on the number of convicted first degree murderers this assumption as a matter of faith.12 Some, how- referred to penal institutions in 1967 and 1968.40 ever, have attempted to justify this practice on Data were only requested for 1967 and 1968 be- empirical grounds. For example, Schuessler argues cause initial inquiries to corrections authorities that the high degree of correspondence between revealed that referral statistics for prior years police, prisoner and mortality statistics on homi- (before 1967) were unavailable in many cases. cide-not murder-dearly suggests its plausibil- Secondly, this investigation was initially launched 33 ity. late in 1970, and referral statistics in many cases The net effect is that no one has succeeded in had not yet been compiled for 1969. Consequently, accurately counting the number of capital offenses reasonably complete data could only be obtained hidden in the available homicide statistics in for these two years. In total, complete data were order to test this assumption. 4 Presently, it is received from 41 states, with Mississippi, Arkansas, Penalty: A Reconsideration, 61 J. Cans. L.C. & P.S. Georgia, South Carolina, Missouri, Pennsylvania, 539, 545 (1970). 30The homicide offense category used by the Federal Arizona and Alaska unwilling or unable to supply Bureau of Investigation is murder and non-negligent the needed figures." manslaughter. It is defined as "all willful felonious homicides as distinguished from deaths caused by M. WOLFGANG, PATTERNS IN CPIINAL HoMrE negligence." FEDERAL BUREAU oF INvEsTiGAiON, (1958). Nevertheless, there has been no progress in CIE IN =rn UNrIE STATES: UmNoUM CnIE RE- filling this void. PoR s at 61 (1967). The Public Health Service defines 3'T. SErIn', CAPITAL Pu is=Nr (1967); Schues- homicide as "a death resulting from an injury pur- sler, supra note 6; Sutherland, supra note 2. posely inflicted by another person." Intent to kill is not 36H. BEDAU, THE DEATH PENALTY IN AIERICA required to classify a death as a homicide. PUBLIC (1967). HEALTH SERVICE, HoAncmE IN =x UNITED STATES: "T. SELLIN, CAPrAL PmusnmirN (1967). 1950-1964 at 9 (1967). 8Schuessler, supra note 6. 31T. SELIIN, THE DEATH PEANLTY (1959). 3T. SErniN, CAPITAL PUNISSM N (1967); T. 2Id. See also Schuessler, supra note 6; Sutherland, SELLIN, THE DEATH PENALTY (1959). supra note 2. 40 For states with no central corrections authority, "Schuessler, supra note 6. But see Bailey, First and individual inquiries were made of each penal institution Second Degree Murder: Some Empirical Evidence in the state. (paper presented at the 1974 Alpha Kappa Delta Socio- 41 For the states of Virginia, New Jersey, Oregon, logical Research Symposium, Mid-Atlantic Region, Minnesota and Connecticut, figures were only available Richmond, Virginia). for the fiscal years 1967 and 1968. Further, statistics '4Studies of homicide have consistently recognized were only available for 1967 for New Jersey. These the absence of adequate statistics on capital homicide. cases were included in the analysis. 19741 THE DEATH PENALTY

First and Second Degree Murder offenses. These data may not be interpreted as reflecting the number of first and second degree Due to variations in homicide statutes across murders committed in each jurisdiction, the num- the country, a definition of murder in the first ber accused of first and second degree murder nor degree was provided with the inquiry to assure the number of persons tried for these two offenses. comparability of the data.12 Since it was impossible In addition, these data also do not completely and to break down homicide referrals by degree for accurately reflect the total number of first and Florida, this state was dropped from the analysis.4 second degree murder convictions in each state. In addition, prison officials were asked to report Undoubtedly, a few persons convicted of these admissions for murder in the second degree. Second offenses were referred to mental rather than penal degree murder, although usually not thought of as institutions. The number here is quite small, how- of theoretical importance in examining the death ever, and probably does not exceed 3 per cent of penalty, is considered here for two reasons. First, 48 convicted homicide offenders. it is well recognized that many offenders initially In sum, these data do reflect, although probably charged with first degree murder are later re- with slight error, the number of convicted first and charged with second degree murder in exchange second degree murders for the states and years for a guilty plea. As a result, many actual first surveyed. How well these figures reflect the actual degree murders are listed in court and prison 44 volume and distribution of first and second degree statistics as second degree murders. murder must remain a mystery, however, for as first degree murder is the only Second, although noted above there are no police or mortality figures capital homicide in most states, deterrence theory currently available on a nationwide basis for these suggests that the death penalty may also have a two offenses and the decision of whether a homicide deterrent effect for other forms of homicide as well. is a first or second degree murder is a matter of society so condemns murder that it The fact that court decision. Unfortunately, national figures are the life of the offender "helps to engender 47 demands currently unavailable on court dispositions. attitudes of dislike, contempt, disgust, and even horror for these acts, and thus contributes to the Comparison of Death Penalty and Abolition States development of personal forces hostile to crime." 41 Table I reports mean rates of first and second In fact, the subtle, unconscious effect of law and degree murder, total murder and homicide for the punishment, as opposed to the cool, conscious cal- states and years surveyed. Comparison of figures culation of punishment, was believed by Beccaria for abolition and capital punishment states reveals and Bentham to provide the major mechanism of that for both years, rates for each offense are sub- deterrence. stantially higher for death penalty states. For Limitations of the Data 1967, rates for all four offenses are at least twice as high as those for states without the death pen- to note that the first and second It is important alty. Similarly, mean rates for 1968 for death solely degree murder figures examined here refer penalty jurisdictions substantially exceed those to persons convicted and imprisoned for these two for abolition states and range from a high of 1.9 4 Murder in the first degree typically includes both times higher for second degree murder to a low of premeditation and malice aforethought, while murder 1.6 times higher for homicide. in the second degree lacks the element of premeditation. "Premeditation designates intent to violate the law A comparison of rates for death penalty and formulated prior to the activity," while "malice afore- abolition states with mean rates for all states sur- thought refers to the simple presence of intent to kill at the time of the act." D. GIBBONS, SocIETY, CR=I veyed further reflects the disparity between the AND CsnmNAL CaARs 346 (1968). two types of jurisdictions.18 For both years, aver- 4 This is an unfortunate loss since Florida reported a 46 total of 191 combined first and second degree murder M. WOIrGANG, supra note 34. convictions for 1967 and 256 combined convictions for 47Until 1945, the Federal Bureau of the Census 1968. gathered court statistics under the title JuDicIAL 4It should be also noted that it is a common practice CRnnNAr STATISrCS. in many prosecutor's offices to initially charge many 4s For 1967, mean rates for first and second degree homicide suspects with first degree murder, and "bar- murder, total murder and homicide are .31, .73, 1.15 gain down" thereafter. Whether these two practices and 4.85, respectively. For 1968, corresponding rates "balance out" one another in the statistics must remain are .50, 1.89,1.38 and 5.48. For the relationship between an open question in the absence of hard data. See rates of first and second degree murder and homicide, GIBBONS, supra note 42. see Bailey, supra note 33. It is of interest to note that 4"1R. CaLDWEI, CRamNoLocY 425-26 (1965). the correlation between rates of first degree murder and WILLIAM C. BAILEY [Vol. 65

TABLE I death penalty and abolition states as well as com- MEAN OFFENSE RATES FOR FiRST AND SECOND DEGREE parison of rates for each with the nation's average, MURDER, TOTAL MURDERS AND HOMICIDE FOR shows rates of all murders to be substantially DEATH PENALTY AND AEOLITION STATES, 1967 AND higher in capital punishment jurisdictions. These 1968a findings are consistent with those reported by Schuessler 9 and Sellin5 for the offense of homicide, Abolition Capital States Punishment All Statesb but quite contrary to what deterrence theory would States Offense predict. Some, however, have objected to compar- 1967 1968 1967 1968 1967 1968 ing average offense rates for death penalty and abolition states for such comparisons ignore other First Degree .18 .21 .47 .58 .31 .50 possibly important etiological factors.51 To meet Murder this objection, a comparison of otherwise similar Second Degree .30 .43 .92 1.03 .73 .89 capital punishment and abolition states would Murder seem warranted. TotalMurders .48 .64 1.38 1.59 1.15 1.38 Homiddee 2.72 3.09 5.90 6.04 4.85 5.48 Comparison of Contiguous Capital Punishment Offense rates are computed per 100,000 population. and Abolition States b Mean rates are only computed for the states sur- Table H reports rates of first and second degree veyed in this investigation. cSource: Federal Bureau of Investigation, Uniform murder, total murder and homicide for eight Crime Reports, 1968, (Washington, D.C.: 60-65). groupings of contiguous death penalty and abolition states for 1967 and 1968.52 These data reveal age rates for all four offenses are below the nation's a very similar picture to that reported above. average for abolition states, while rates for capital Inspection of the first grouping of states (Maine, punishment states exceed the national average for Vermont, New Hampshire) for first degree murder each offense. A state-by-state comparison of rates for 1967 reveals the rate for Maine, an abolitionist for each type of state with the average for the state, exceeds that for New Hampshire, a death country further reveals that for both years com- penalty state, whereas the opposite is true when bined 88 per cent of the abolition states have first rates for Vermont, also an abolitionist state, and degree murder rates below the nation's average, New Hampshire are compared. When such com- while only 52 per cent of the retentionist states parisons are repeated within all groupings of con- are below the mean. For second degree murder, tiguous states for 1967, 67 per cent of the compari- 91 per cent of the abolition states have rates below sons show death penalty states to have higher the mean while 52 per cent of the death penalty first degree murder rates than their abolitionist states are again below the national average. In neighbors, while the opposite is true for only 20 addition all states that have abolished the death per cent of the comparisons. In 13 per cent of the penalty have rates of total murder below the comparisons, rates for both types of states are country's average, while only 48 per cent of the the same. capital punishment states are below the average. For 1968, comparison of first degree murder For homicide, 83 per cent of the abolition states rates for the two types of states reveals a very have rates below the nation's average while 48 per similar picture to the former year. For this year, cent of death penalty states again fall into this 64 per cent of the comparisons within neighboring category. groups of states show rates to be higher in capital In sum, a comparison of rates both between punishment jurisdictions, while rates are higher in 2 homicide is r = .38 and r = .20 for 1967 and 1968, 49 Id. respectively. Corresponding correlations for these two 50T. SELLiN, CAPiTAI PuNism NT (1967); Sellin, ROYAL COMM'N supra note 6. years2 between second2 degree murder and homicide are r = .42 and r = .24. Although each coefficient is 51Schuessler, supra note 6. statistically significant at beyond the .01 level, for 52Maine, Vermont, Rhode Island, Michigan, Minne- neither year nor offense do police homicide figures per- sota, Wisconsin, Iowa, North Dakota, Oregon, West mit as much as 50 per cent explained variation rate. In Virginia, Hawaii, and New York are considered aboli- short, contrary to Schuessler and other's claims, police tion states. The death penalty may be prescribed, how- data do not appear to provide "a reliable index of ever, in Vermont, Rhode Island, North Dakota, and murder in general and first degree murder in particular" New York for certain offenses. See H. BEDAU, Tan as commonly assumed. Schuessler, supra note 6, at 55. DEATH PENALTY IN AMERICA (1967). 19741 THE DEATH PENALTY

TABLE II RATES o F FiRST AND SECOND DEGREE MURDER, TOTAL MuRDER AND HOmICDE ron EIGHT GROtJP.ITGS or CoNrIGuous DEATH PENALTY AN ABOnol STATES, 1967 AND 1968a

First Degree Murder Second Degree Murder Total Murder Homicideb State 1958 1967 1 1968 1 1967 1 1968 1967 1968 1967 1

Maine* .50 .30 .50 .80 .90 Vermont* .00 .75 .00 .75 .50 New Hampshire .14 .00 .43 .14 .57

Rhode Island* .00 .11 .67 .11 .67 Connecticut .28 1.21 1.10 1.49 1.33 Massachusetts .09 .20 .54 .29 .82

Michigan* .34 .76 .78 1.11 1.22 Indiana .16 .32 .35 .48 .64 Ohio .43 .66 .71 1.09 1.26

Minnesota* .08 .14 .19 .22 .24 Wisconsin* .26 .21 .60 .47 .89 Iowa* .07 .21 .21 .28 .25 Illinois .63 1.17 1.46 1.80 2.42

North Dakota* .00 .00 .00 .00 .00 South Dakota .00 2.00 1.29 2.00 1.58 Montana .00 .14 .43 .14 .43

Washington .09 .44 .36 .53 .42 Oregon* .30 .45 .50 .75 .65 Idaho .57 .43 1.14 1.00 2.14

West Virginia* .33 .61 .67 .94 1.00 Virginia 1.02 1.31 1.37 2.33 2.09

New York* .14 .11 C .25 0 New Jersey .24 .74 .66 .96 .93

* Abolition states. aOffense rates are computed per 100,000 population. b Source: Federal Bureau of Investigation, Uniform Crime Reports, 1968 (Washington, D.C.: 60-65). Murder I and II statistics were not available for New York for 1968. only 29 per cent of the cases for abolitionist states. compared, rates are higher in the former jurisdic- Seven per cent of the comparisons show rates of tions. In contrast, rates in abolition states exceed first degree murder to be the same for both types those in neighboring death penalty states in no of states. more than 40 per cent of the cases compared. Further inspection of Table II indicates a very These findings are consistent with earlier examina- similar pattern for the remaining three offenses. tions of homicide, but contrary to what deterrence Comparison of abolition and death penalty states theory would predict. for these offenses, as well as first degree murder, are summarized in Table III. Figures reported in Offense Rates and the Certainty of Punishment Table I for contiguous death penalty and aboli- Proponents of punishment argue that if legal tion jurisdictions for second degree murder and sanctions are to act as effective deterrents, they total murder and homicide (for both years) reveal must be "real." That is, if the probability of pun- that for at least 60 per cent or more of the states ishment is very slight or non-existent, it will not WILLIAM C. BAILEY [Vol. 65

deter no matter how severe. This point assumes TABLE I particular importance when examining past inves- SUMMARY OF COMPARISON OF OFFENSE RATES FOR tigations of the death penalty, for as Giggs notes, FIRST AND SECOND DEGREE MURDER, TOTAL MUR- much of the evidence on the inefficiency of the DER AND HOmnCmE :FOR CONTIGUOUS ABOLITION death penalty is based upon normative legal differ- AND DEATH PENALTY STATES' ences among political units (whether or not there Higher Rates is a statutory provision for the death penalty), and Rates for RateHigher for Tied for 5 Death Abolition Both not upon the actual use of capital punishment. " No Penalty States Types of Offense Year States States one would argue that the death penalty could be an effective deterrent if it is never used. Accordingly, No. % No. % No. % the important question would appear to be, how are differences in the use of the death penalty in re- First Degree 1967 10 67 3 20 2 13 tentionist states related to offense rates in these Murder 1968 9 64 4 29 1 7 jurisdictions? To examine this question, execution Second Degree 1967 9 60 6 40 0 0 rates were computed for each retentionist state Murder 1968 9 64 5 36 0 0 (operationally defined as the total number of Total Murder 1967 10 67 5 33 0 0 executions for homicide during the last five years 1968 11 79 3 21 0 0 per 1000 homicides for these years) and correlated Homicide 1967 10 67 4 27 1 7 1968 9 60 6 40 0 0 with rates of first and second degree murder, total murder and homicide for 1967 and 1968. Figures Figures in the number (No.) columns refer to the for homicide are used in the denominator of the total number of cases where comparison of contiguous execution index for figures for first degree murder- death penalty and abolition states show rates to be (a) the most appropriate offense-are not available higher in the former, (b) higher in the latter, or (c) the for these two years. In addition, a five year time same for both types of jurisdictions. period preceding 1967 and 1968 was used in com- puting average execution rates in order to provide TABLE IV greater stability in rate and to allow sufficient CORRELATION OF RATES OF FIRST AND SECOND DEGREE time for the presumed deterrent effect of executions MURDER, TOTAL MURDER AND HOMICIDE WITH EXECUTON RATES FOR 27 DEATH PENALTY STATES, to be realized. Results of this analysis are reported a in Table IV. 1967 AND 1968 Deterrence theory would suggest that the higher Offense 1967 1968 the execution rate the lower the rate of capital homicides in death penalty states. Figures in row First Degree Murder -. 137 -. 194 one of Table IV reveal only a slight inverse rela- Second Degree Murder -. 167 - .351 tionship between executions and rates of first de- Total Murder -. 180 - .302* gree murder. Although both coefficients are in the Homicide - .166 - .039 predicted direction, neither is statistically signifi- Coefficients are Pearson product moment correla- cant at the .05 level nor does either permit as tions. much as 4 per cent explained variation in rates of *p < .05. first degree murder. Further inspection of Table IV reveals a very similar pattern for the remaining three offenses. As with first degree murder, each tentionist states (two in 1967, one in 1966, seven of the coefficients is in the expected negative direc- in 1965, thirteen in 1964, twenty-one in 1963 and tion, but only the correlation for second degree forty-seven in 1962), thus restricting the range of for 1968 reaches statistical significance at the .05 the execution index. Accordingly, it might be level. Even here, however, only approximately argued that had the distribution of this variable 12 per cent of variation in offense rate can be ac- not been so restricted, the negative correlations between execution and offense counted for by executions. rates would have In interpreting these findings, it should be noted been larger. Although an attenuated distribution on an inde- that for the five year periods preceding 1967 and pendent variable would have this effect, this factor 1968 there were relatively few executions in re- is not of great importance since this study is con- r Gibbs, supra note 8. cerned with the relationshin between actual (not