.11 1 ‘ Iwzx’i‘fifi. 5"‘7 ' L»..« .,V\-" “51' cg‘zsifm unriflfi ‘ ‘ '1‘ «NIPI: ‘2‘].17‘ . ‘ ‘ ) qu“ La ‘ . s _ $ ‘ ga 1,] .. 'r’r 1‘ - r .3, ~ \-E:L..‘7.i7‘-L"_v-. egj‘ufri‘f’f“ g‘C‘L.:5.‘;.ixd"3‘9;\n-3‘~ ‘ 1;. a - .4 . , 5.. klmt’.‘ . .5. " -' - - ~"--~ . , J‘ * my” now An A: w“? $7 ‘5‘“:‘1‘1’56‘ kw ”. Rum- {Kai-‘13)?“ b ‘ 3‘12.- film .. .,..- w. 9‘ rem. 5'.“ 2:5“ ’: .4 "m“ V 3‘ ~' ‘ ' 33% ,4; ‘3 - * . r .«a “1;; w‘m |~~ A ‘ ' — -. 55:3?» ”1 _ "£3 “15$? ' .1‘1‘W3 ‘ ‘ K» fir 4‘ X‘ $31.? :7 Js. W “~74. ‘ .«-. ”.e s,..,‘4V' V..._‘.....‘.‘..‘_.‘E§2 ; .L‘” _- ., . . ~ ‘ I n. V , yak-"5.1“: . .~ «$1? ‘ . w‘r'" ‘ ‘jSuzé; 3" <., . m |---z‘-imr1‘l-h.':.'.... :J.- M -— .1... ..-. :55“ «NF; ‘3 ¢§- \ ‘24; . 1‘ A a . o. I: q a ,u \ mk ,w, Fur}. I . ’. r'., ‘ vj ;-; ..- .‘q..u‘ .54; '.' _ “Pl-.3315 MICHIGAN STATE IVERSITY HA m u mun Ill/mmmill/ill fl 3 1293 009001607 I' This is to certify that the thesis entitled INTRAFAMILY HOMICIDE: THE GENEALOGICAL AND AFFINAL RELATIONAL DISTANCE BETWEEN VICTIM AND OFFENDER presented by DIANNE DETRIO NADERI has been accepted towards fulfillment of the requirements for M. s . degree in CRIMINAL JUSTICE dawn/A Majoan/ofessor / / Dateinafidl gilqg/ 0.7639 MS U is an Affirmative Action/Equal Opportunity Institution F LIBRARY Michigan State University K J PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE OCT 2 31993? ‘ l l ll—_ll l MSU In An Affirmative Action/Equal Opponmlly Institution GWMI INTRAFAMILY HOMICIDE: THE GENEALOGICAL AND AFFINAL RELATIONAL DISTANCE BETWEEN VICTIM AND OFFENDER BY Dianne Detrio Naderi A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE School of Criminal Justice 1991 .‘I 1 ‘ . If} '1'): I C' 5‘ ‘ ‘ ”4’ ." ., , . - -‘ I fife” I. ‘ I. P "'{t - ' '1“ i ’ h i: ' t j . ’2', ‘ I s I: “ "J ' 1“ I" \ I A K E1 ' .t a 0‘ l'lgmdll m . .. ‘,r[' . ,_ I) [.i 4 . .-‘p .d I ‘_. ‘ v,“ "7 w“)! . V! - I (!-,I .‘ ’4 9 1 " 4 _ I v' .1 4'. . 2 I ‘3 " '1'. , l \‘r withi AI relal A . t .1; _"l :'fl."? ’_- 13:.‘r: and Chic: for ‘ anal: indi‘ With: brotl exhil and ' 2 4 9 .‘ l ' i ‘- ‘37? ' +. .‘ am rue ‘ I ’* ta .hie'al’= ‘rt I'I'I'JI‘I ‘ . ‘rw f-Ql?‘ ! .a‘f‘4 .u,‘ ‘-‘ "‘:’"::“ h ABSTRACT INTRAFAMILY HOMICIDE: THE GENEALOGICAL AND AFFINAL RELATIONAL DISTANCE BETWEEN VICTIM AND OFFENDER BY DIANNE DETRIO NADERI This thesis explores the question of who is killing whom within families. The genetic relational distance or non- relatedness between offender/victim categories is measured and analyzed for all family homicides which occurred in Chicago 1965-81. The distribution or incidence of homicide for 50 relational categories (first offender to victim) is analyzed. The research hypothesis that non-blood related individuals are at greater risk to be killed by each other within families is supported. With the exception of brother/ brother homicides and parent/son homicides. a pattern was exhibited wherein less closely related, affinally related. and unrelated family members were killed more often by each other. Marital homicides account for 65.9% of the total family homicides. Husbands and wives killed each other in almost equal numbers. It was concluded that family homicide is a non-random event and that genetic relatedness and other predictions derived from selection theory and coevolution may help explain the dynamics. Daly and Wilson's study of homicide and natural selection theory and Durham's 'coevolutionary" theory wherein biology and culture is an interwoven DIOC888 are 618008896. '4 l) C i. . . . b e lo A c I. . o in. J . q ' 1 , o . i. . s. i _ l . .7 . x . . . v . . . , e ' ‘ . . . l I | i v‘ .I O . T . 0.. f. I i . t w. . a . r . ll .6: .. rt . Chapter TABLE OF CONTENTS Page LIST OF FIGURES ............................... iii LIST OF TABLES ................................. iv ACKNOWLEDGMENTS ................................. v INTRODUCTION TO THE PROBLEM: INTRAFAMILY HOHICIDE ........... ...... .......... 1 Purpose for Study ....... . .......... ...........3 Hypotheses and Theoretical Perspective ........ 8 Coevolutioanry Theory and Family Homicide....l4 REVIEWOF THELIIERATURBOOOOOOOOOOOOOOOO000000021 Part I. Homicide: Major Theories.... ........ 21 Family Violence Studied in Cross- Cultural Perspective... ........ .....22 Intrafamily Homicide ................ 24 Marital Homicide. . .............. ...37 Child Victims of Homicide ........... 39 Juvenile Homicide Offenders ......... 41 Part II, Homicide: Genealogical and Affinal Distance Between Victim/Offender....43 An Ecological/Evolutionary Perspective .......... . ........... 53 Discussion of Previous Research .............. 54 RESEARCH DESIGN OF STUDY ....................... 55 The Samples.. ................................ 55 Methods and Hypotheses ....................... 57 ANALYSIS OF RESULTS.... ........................ 70 Presentation of Results ...... ...... ........ ..70 Discussion of Results ........................ 82 Summary of Results ........................... 87 SUMMARY AND CONCLUSIONS ....................... 88 Summary ................................... ...89 Conclusions ................................ .90 Discussion ................................... 91 Implications for Future Research .......... ..102 LIST OF REFERENCES ............................ 104 ops‘l ‘1 fl ' I ‘i a o r i . C 1‘! 't I l a TC I: I f . ,i b" x"’ . O u‘. .“w- »4 ‘- O a e O U 0 e e o e a a a i \9 1-3 .ziTIDH 11 f 'C‘ :4 !)I ilinffi I’fi'. 3 .. .2 D. 0 e on I m”... ".l I“ T, l l';‘.’ iwl a o e e I ‘l s , ,- .n‘“ J 'v i'fl-‘I‘ -' 3 '3 :3 Mill if)?! i I‘frl.’ I .. 1"C’Q.I1r: b i I ' . 3’9. 2'1“ ."_"”."'I'A.,3'7‘!"T.I EHT V o 4 I .' v I 'I'J- '{.'3.D.~'.' I . . .Y‘J‘ITB "-m .Td'il T‘ll W 4' ‘0'.. Y if QM 'aj‘y‘: . 5.". - .11 "4 10-: 'nt -H \r I \ :fI".'1 in» r r: H .."I)[F;.‘.‘:.;I enutag e. . a .‘ 't 40 '1 7853' 7.2“) 305“" .83 IHAT '40 Tilll .£-;'3 wen; 2* 02mm in .' 'Mi “INT UT If"i1'l"i:!r_fn-.}T;,H ‘(J ”46117 g 'M 1 ‘.I:,"'(. ‘3' .c OT' “7"" 5H” \i".:c:..,1j l ‘;.I‘.I'.Irg.‘I‘ 1 .r _'l 1'06"] 38"~'.'!It'7' II“? A a ,4! LI’HJ- ( ', .,".\l""-“‘ " ‘9‘”16. =- «V'- R A ' ‘ s' c": C _I \ armiv'e'rq i. r ‘ i .7 i4?) ‘ ”.1153 3".- V". ‘7"! 5‘ I ._ '11"? I ' o - meme“ INT"? Ft‘!‘ ‘L 'z _:;--‘r m :32 am :i’ry 1; L‘s . 2» 4 I 3‘0: CT 1") 37].? I h ‘ I e .' " I H 2T", (":3 ,- ‘I" . rti .- ' 19-..‘."‘ J V v v: I :2“: 1' gift hi,“ ."y {le -, '-:’ ,_‘ p. “17' I ff: Ir”: " \ II; I~ c' I ‘ l r "1" J» .. . .. I 'i f. g 4 ' I, r ;-j.t' ,I.7'.,i>,;',yi 3‘! .h ‘0" I ‘I Figure LIST 9: IGURES Page Risk of homicide by relationship. considering only those cases where victim and offender co-reside ..................... ....47 Frequencies and percentages of family. acquaintance. stranger. and mystery/unknown homicide, which occurred in Chicago. 1965-81...71 Number frequencies of homicide for 57 relationships of first offender to victim. Chicaqu 1965-810000000000 eeeee eeeeeeeee eeeeeee 73 Relative percentage frequencies of homicide for the relational distances of 0. .01. .125. .25. .5 with plot of logarithmic curve. Chicago 1965-81 ......... . ....... .. ......... ....78 Comparison of consanguineal and affinal family homicides for Chicago (1965-81). Detroit (1972) and Miami (1980) ........................ 82 Model of one mode of relationship between genes and culture (genetic mediation) as seen in the homicide of a relative...... ..... .... ..... .94 Model showing how coevolution may operate to produce behavioral strategies explained by and/or derived from both ultimate and proximate pressures .............. .. ............ 95 iii LIST OF TAB ES Table Page 4.0 Frequency and relative percentage frequency for certain categories of family homicide. Chicago 1965-81 ....... . ........ ........ ....... 77 4.2 Comparison of family homicide for Chicago (1965-81). Detroit (1972) and Miami (1980)....81 iv ACKNOWLEDGENENTS I wish to thank my master thesis committee members: Dr. Vincent Hoffman (Chairman). Dr. Timothy Bynum. Dr. Dennis Payne. all from the School of Criminal Justice. and Dr. John Mullins. from the College of Integrative Studies: Natural Sciences. It is acknowledged that the Chicago Homicide data utilized in this thesis were made available (in part) by the Inter—university Consortium for Political and Social Research. The data for HOMICIDES IN CHICAGO. 1965-1981 were originally collected by Carolyn Rebecca Block. Neither the collector of the original data nor the Consortium bear any responsibility for the analyses or interpretations presented here. I would like to thank Harriet Dhanak. at Michigan State University for helping me access the Chicago homicide data file by SPSS-X. And. I would also like to thank Dr. H. Rahber for his help in the statistical research design and interpretations. I: .' 1 ‘ ' 1’ ' . .1 C a (I 0 I b CHAPTER 1 INTRODUCTION 19 THE PROBLEM: INTRAFAMILY HOMICIDE The killing of another individual is a complex behavior. which universally has occurred even among our earliest ancestors. In fact. in pro-agricultural societies the killing of outsiders occurred and was not perceived to be wrong or a crime. unless it brought wrath upon the entire group. "Lethal retribution is an ancient and cross-culturally universal recourse of those subjected to abuse" (Daly and Wilson. 1988:226). The practice of killing one's child, shortly after birth. infanticide. has persisted in many human societies over the centuries; and was never considered a crime. Since the beginning of agriculture and more sedentary patterns about 10.000 years ago. murder has increased steadily. with lethal retribution in some instances rising to the point of being a sacred obligation for the individual(s) who had a family member or some other killed. without due cause. Contemporary society has witnessed homicide perpetrated en masse and serially. There are multitudes of causes and circumstances which may culminate in a homicide in modern society. There is also great variation in the definition and practice regarding the wrongfulness of homicide. such as excusable and justifiable homicide. and the government's use of capital punishment (Nettler. 1982). Homicide within a family. where one kills another who is related by blood or marriage. is a category of homicide which some have studied in detail and with precision. Many have looked at the relationship between offender and victim since Wolfgang (1958) first suggested and delineated such an inquiry. However, few have elaborated beyond descriptions of rates. proportions. gender. race. and age pertaining to the broad categories of family. acquaintance. and stranger homicides. Most of those who have studied family homicides have looked at relatively few numbers of individual cases (usually psychological and psycho- social studies) and bear no concern toward the larger picture of the dynamics involved. The intent of this study is to examine specifically who is killing whom within families. and to also explore the larger picture which may help explain the dynamics of lethal violence within families. On a very preliminary level. coevolution theory: the joint evolution of biology and culture (Durham: 1979. 1983) will be discussed because it may be a potential component of the larger picture and regarding its potential application for researching the etiology of intrafamily homicide. The definition of homicide shall include those assaultive and other acts directed against another person that occur outside the context of warfare. and that result in death. Criminal and "justifiable” or ”excusable” homicides are of interest because we seek to examine all homicidal behavior not just those cases of homicide remaining after prosecutor decisions and convictions. The few falsely accused innocent cases will most likely be insignificant; whereas an exclusion of ”justifiable” or "excusable" cases are. in fact. often significant in number. What is not of interest is an accidental homicide such as vehicular manslaughter (Daly and Wilson: 1988a. 14-15). Purpose for Study The purpose of this study includes the following. 1.) '_I‘_o__ increase our understanding of family homicide. Richard Gelles. one of the most prominent investigators of American family violence. implores that we "need research which has potential to generate theories. if not test.” and which may state what the data mean in terms of a particular theory (1979: 178). Authors Silverman and Kennedy (1987) uphold that homicide research should consider and control for relational distance. age relationship. and gender relationship of the offender and victim. An analysis of the complex social relationships of roles within a family and also outside of a family (wife. husband. step-parent. daughter. etc.) may elucidate how certain social relationships are related to homicides which share dissimilar origins (Parker and Smith. 1979). Accordingly. the need for theoretical research which specifically addresses the victim/offender relationship within family homicide is evident. 2.) Family homicide is s serious and great problem which requires more in-depth research. The issue of family violence demands greater understanding. Riedel and Zahn (1985) found that family homicide was second in frequency in the United States during 1976-1978. after acquaintance homicide. with stranger homicide third. Other studies from other countries and time periods vary little from this statistical configuration. The incidence of acquaintance or family homicide usually predominate firstly or secondly in number. Therefore. the vast majority of homicides involve individuals whose relationships are crucial to understanding the violence (Silverman and Mukherjee: 1987). The literature about family homicide in the 0.5. generally uphold the following conclusions. Family homicide is rarely associated with the commission of another felony. unlike stranger homicide which very often is associated with another felony. Family homicide displays a large representation of females as victims. This is in contrast to the other categories of homicide where males are overrepresented both as victims and offenders. It is observed that family homicide seems to be the only type of crime where males as victims are not overwhelmingly represented. Minority ethnic groups. especially blacks. usually are overrepresented as participants in family homicide. All of these issues require investigation in order to understand why family' homicide (displays distinct patterns. 3.) Research and theories which include the svolutionssy and adaptive significance sf behavior sgs important Lg; gaining knowledge sbout the universal occurrence sf family homicide. Some research suggests that the rates of family homicide vary little among nations. Where there is an overall low rate of homicide in a particular country. that country may show higher proportions of family homicide in comparison to a country like the United States where the overall homicide rate is extremely high. The implication is that family homicide is a category of homicide in which the rate or number (not the proportion) is similar cross-culturally. Therefore. this category of homicide where stability and persistence are apparent (i.e. some evidence of universal similarity of family homicide rates). notwithstanding diversity. is particularly appropriate for this thesis. This is because the perspective searches for the adaptive significance of a persistent but diverse human behavior. Consequently. an analysis of the victim/offender relationship for intrafamily homicide is especially applicable in order to understand why so much lethal violence occurs within families the world over. Or. as Zimring. et. a1. (1983: 925) contrarily view the situation. we should examine why there are so few intimate homicides "within the turbulence of marital intimacy." As Wolfgang states. homicide is a very personalized type of crime. Looking at the relational distance of the perpetrator and victim would seem to be a compelling route toward understanding the personalized aspect of family homicide. Other routes would probably include exacting' analyses of the situation. psychological backgrounds. motives. and all pertinent features of the ”family" such as lifestyle. socioeconomic station. and many other factors relating to the victim and offender. many of which have been studied by homicide researchers. 4.) Potential for practical uses which msy _b__e_ derived from findings. Another crucial and ethical reason why we need to understand the dynamics of family homicide is the potential utility of the findings for planning prevention and crisis intervention. Social service agencies. the police. and the courts have in the past given little attention to domestic violence. probably because of the notions that "a. man's house is his castle." and that family matters are private. For instance. United States child protection laws were not passed until the 1960's. and the Child Abuse and Neglect Act until 1977. Maney and Kedem (1982) concluded that the rate of child homicide (filicide) in the District of Columbia was decreased after the enactment of the more stringent legislation in 1977. In the past. parents have considered their children as property to be abused. sold off. and disposed of. The behavior of neglecting and abusive parents was not even perceived to be behavior which could be treated or prevented until the recent rise of social casework (Maney and Kedem. 1982). Others have empirically demonstrated that :many domestic fatalities occur subsequent to police calls for domestic disturbance and after long histories of assaults. and suggest that improved intervention may help to reduce fatalities (Wolfgang. 1978a. 1978b; Straus. 1986). Wolfgang notes that even children from a subculture of violence may be resocialized towards nonviolence with specific activities promoted at home and school. Rewards and affection for not harming others would help children to learn alternative behaviors. Mulford (1983) in a critical evaluation of programs found preventive efforts to be almost non-existent; and that new approaches for dealing with violent behavior patterns and providing services and supports which stabilize and strengthen family life may help towards prevention of violence. Also. the Attorney General's Task Force on Family Violence in their final report recommend treating assaults within the family as seriously as assaults between strangers (1984). If the relational distance between individuals can be demonstrated to be a significant factor. or to predict risk for homicide. then, needless to say. such predictions may be useful for planning intervention and prevention strategies. Unfortunately. official statistics and most previous authors have failed to distinguish between various relational distances based upon blood or marriage. Family homicides are usually lumped together without regard to the specific relational distance of the victim and offender except for spousal homicides. where it may be usually assumed that the individuals are genealogically unrelated. Thus. the severe obstacle for such research is tedious first-hand data-gathering from police documents and :medical examiner reports. until such time that. for example. the FBI Uniform Crime Reports undertake a revision. H othe as an heoret ca Pers active The intent of this study. consequently. is to cast some light upon homicides which occur among family members. The research questions are: 1) Is the frequency of homicide less common among genealogical or consanguineal (blood) relatives? 2) Is the frequency of intrafamily homicide expected to be greater among non-genealogically related individuals and/or where the relationship is by affinity (marriage) only? 3) When comparing the actual frequency of homicide to some ‘hypothetical predicted frequencies. based. on greater or lesser interaction time of the offender and victim. are the actual and predicted frequencies different? 4) Will three geographical urban areas during different time periods show rates and patterns of family homicide reflecting more homicide occurring among non- genetically related family members? The present study will attempt to elaborate upon Daly and Wilson's (1988a. 1988b) finding that the risk of homicide for some categories of ”intimates" is indeed significantly greater even when the category of homicide between spouses is removed from analysis. The assumption in researching the present questions is that an adequate explanation of homicide requires an interdisciplinary exploration of the entire milieu. including obscure elements. Behavior. including the very personalized act of homicide within families. results from proximate and ultimate causes. In agreement with Daly and Wilson (1988a). a theory of 10 ultimate causation which might apply to the study of homicide is the general theory of evolution: natural selection theory. However, it is questioned whether the paradigm of natural selection theory is firmly established regarding homicide. as Daly and Wilson assert. The possibility of a more complex explanation is presented. The tenets of Durham’s Coevolutionary theory will be presented and one mode of relationship between genes and culture will be used to generate hypotheses regarding family homicide. Natural selection is the main process whereby human evolution occurs. It is a condition in which. on the average. individuals with certain genotypes (the DNA pattern present) produce more offspring that survive long enough to reproduce than do individuals with other genotypes. This "differential reproductive success” means that individuals better adapted to their physical and social environment on average will have higher fitness. "Kin selection” is one process by which genetic fitness or ”inclusive fitness" takes place. It includes marriage and mating behavior. aggression and any behavior or genetically encoded behavior which enhances the survival of the individual's genes and/or the individual's gene forms also carried by relatives. Other evolutionary concepts such as "parental investment." ”parental certainty." and paternity 11 certainty are important in the study of family violence. Natural selection operates on an individual level. with very few exceptions. All human beings are complex adaptive systems. This is a problem of understanding how genes and the environment interact to produce adaptive and maladaptive behavior. Montague states "that aggressiveness. under certain conditions. is a quality of considerable survival value. a form of adjustment” (1974: 371). Although current legal systems unequivocally consider nonaccidental homicide as deviant behavior. outside of the legal norms. evolutionists and behavioral biologists uphold the belief that persistent human behavior patterns occurring throughout the millennia will most often contain some measure of adaptability or perhaps neutrality. This is not to infer that homicide may be an enhancing adaptation. but that we would expect a pattern of homicide where genetic relatives are least likely to conflict; and that the patterns would show frequencies in the direction of preserving reproductive success and "inclusive fitness” for the perpetrators of homicide. The addition of an evolutionary perspective to understanding human social behavior enables the researcher to expect that most forms of behavior will be either biologically adaptive or will be expressions of 12 evolved tendencies that were adaptive in the past. A behavior‘s adaptiveness is always determined by the environment in which it occurs. The ability to describe that a particular behavior is adaptive to a particular environment is a statement about its effect on survival and reproduction. nothing more (Irons. 1979: 38). Given that the predominant behavioral inclination of human beings is cooperative and nonviolent. especially’ within families. the study’ of .fatal interaction within families may enlighten the behavioral scientist's understanding of the dynamics of both violent and nonviolent behavior. Durham (1979. 1983) proposes that human social behavior results from a coevolutionary process of biology and culture that maximizes the individual's reproductive and survival success (see also: Dobzhansky. 1962; Blum. 1963. 1967. 1968; Cavalli-Sforza. 1971. Cavalli-Sforza and Feldman. 1981. and others). Further. he and others agree that culture is a “proximate” mechanism which operates among individuals to selectively retain adaptive traits. He posits that culture would selectively retain those traits which are adaptive in the "biological sense.” Significant progress in both anthropology and evolutionary biology has prompted Durham to believe that the proposed theory may help to end the nature/nurture debate. He argues that more progress may be made by a 13 theory which makes no prior assumptions. e.g. about the genetic or cultural evolutionary development of a particular form of social behavior. Durham discusses one recent development in Anthropology which reorganizes the earlier view of culture as behavior or as ”adaptive systems" to a view of culture as conceptual phenomena in the minds of human beings. The conceptual phenomena (ideas. values. beliefs) are distinct from action and are the "blueprints" or "programs” which shape human behavior and adaptations (Geertz. 1965). He sees this as leading to the recognition that the human possesses two major information systems: one genetic. and one cultural. Both may transmit across space and time; both shape the behavior; and.'both. are simultaneously' co-resident in every living human. In addition. our human ancestors were very much cultural beings during the last 2.5 million years. two to three times earlier than previously believed. according to the archaeological record. Thus. the interaction and interdependence of genetic evolution and cultural evolution operated during this period. Also. "the human capacity for culture itself evolved by genetic evolution" and rapidly evolved as time went along. one indicator being the increasing brain size (1983: 5.13). Evolutionary biology and the study of animal social 14 behavior have also made undisputed scientific strides and has given Durham direction. W.D. Hamilton's (1964) "kin selection" theory broke theoretical barriers and many other models followed. Durham remains critical of the application of the sole use of the genetical theory of natural selection for explaining human behavior. Critics would disagree with Durham. but coevolution is purported to be an undeterministic process and in fact the cultural mechanism would frequently override the biological mechanism. Coevolutionary Theory ssg Family Homicide This thesis will test several hypotheses which have been generated from a coevolutionary perspective and thinking. Durham's coevolutionary theory reaches beyond sociobiology and evolutionary psychology. It recognizes the two principle mechanisms which transmit all attributes: the cultural mechanism and the biological mechanism . and that these mechanisms operate inextricably intertwined. The theory posits five modes of relationship between genes and culture: (1) genetic mediation; (2) cultural mediation; and (the self-selective potential of culture) which may result in an (3) enhancing; (4) opposing; or (5) neutral relationship of the two inheritance systems. The cultural mechanism spreads an attribute by learning and maintains it by tradition. It is also proposed that the consistencies between 15 biological theory and human behavior are best explained by the joint evolution of biology and culture. "The apparent consistencies between biological theory and human behavior suggest not that there is necessarily an underlying biological basis that guides. steers. controls. programs. predisposes. or inclines every human activity. but rather that the traditions and customs produced by cultural processes are often adaptive in the "biological sense." (1979: 41) The joint evolution of biology and culture may be insightful in understanding the persistent pattern of homicide among humans. A coevolutionary perspective is suggested to be especially appropriate for analysis of a given phenomenon where one can investigate the differing reproductive benefits and when the "costs” of the phenomenon can be factored out in terms of time. energy. and resources (1979:57). Homicide of individuals closely related. more distantly related. and unrelated have differing reproductive benefits for the perpetrator. An attempt at homicide is an extremely high risk behavior as the individual risks his/her life being taken by the victim. although this is very rare. Other "costs" associated with the action are arrest and/or imprisonment. another form of court sentence. and loss of status. all of which may impinge upon the individual's future reproductive success. An attempt to analyze and factor out the costs that an individual incurs subsequent to a homicide should be a worthwhile endeavor. and is the appropriate course for the 16 researcher embracing a coevolutionary perspective. Daly and Wilson. in their book Homicide (1988a) which addresses homicide from a "selectionist" perspective. admit that they make no attempt to test the "meta” theory of natural selection. Nonetheless. they make no mention of the interwoven process of culture and biology (i.e. natural selection) which may contribute to the presence of homicidal behavior. It is my contention that a theoretical perspective incorporating both biology (genes) and culture (the brain) is the optimal route toward. understanding homicide. It is hoped that this thesis will preliminarily look at Coevolution's potential. Just one of five modes of relationship ‘between genes and culture will be addressed in this thesis. that of genetic mediation. Any further attempt to dissect coevolutionary theory regarding family homicide shall be put aside because the present samples of data and research design do not warrant it. Therefore. the one mode of relationship between genes and culture that will be used to generate hypotheses may more simply be referred to as natural selection. Such a general theory of human behavior is lacking in criminology and may at the very least raise new questions and generate fresh directions in homicide research and in criminology. A number of criminologists are attempting to delineate a ”meta” theory. such as 17 Williams' theory which describes eight elements of orienting strategy for looking at criminal behavior (1988). He contends that "meta” theories reflect what concepts should be in the research and provide guidance in how one designs research; and that theories should be dynamic. not static. Some criminologists boldly argue that certain patterns are best explained by biological disposition. Nettler (1983) states in a chapter about the social locations of homicide and the commonly held "indicators of difference" that aid our understanding of behavior: age. sex. work. and wealth that "Biology is translated into action only through social sieves. Biology limits capacity and creates disposition. but social relations channel the expression of disposition. Males are disposed to greater violence than females. but social situations variously facilitate or inhibit manifestations of this biological given. This means that females in one culture may be more murderous than males in another” (1983: 17). A pattern of homicide that is consistent with natural selection theory (and with the genetic mode of relationship between genes and culture) includes killing genealogically unrelated individuals with greater frequency. It is apparently maladaptive for an individual to kill his or her kin because this reduces the individual's gene representation in succeeding generations. Evolutionary models of social motives predict that genetic relationship will be associated with less conflict. 18 According to the same deduction from a selectionist perspective. an obviously maladaptive manifestation is the action of killing oneself. with the possible exception of a suicide occurring in very old age after the individual is unable to provide further child bearing. providing. and any emotional and financial supports for closely related descendents or relatives. Maladaptive behaviors often occur in evolutionarily novel environments. Some research suggests that suicide and homicide are parts of a complex social process (Brown. 1986; Palmer. 1975; Palmer and Humphrey. 1979. 1982). Such findings alert one to the possibility that a search for the etiology of homicide may entail very broad investigations of other forms of deviant and maladaptive behavior. such as suicide and the linkages and social meanings thereof. not to nmmtion greater social forces of the sort identified by Durkheim and others. Brown amd others (Douglas 1968; Atkinson 1978) contend that the social meaning of suicide should be addressed because. critical of Durkheimian. approaches. the situations purported to cause suicide may just as well produce other kinds of deviant behavior» 'The same argument may be true of homicide. However. it should be remembered evolutionary theory holds that natural selection and coevolution operate among individuals without their necessarily 19 consciously or unconsciously attaching meaning to behavioral alternatives (and their consequences for survival and reproduction). only that the net effect enhances survival and reproduction. Importantly. Durham cautions that positive evidence of correlation of a human behavior to natural selection theory in no way implies causation; as the behavior may result from the complementary process of cultural inheritance and cultural evolution. As Simpson (1972) states. ”the more powerful means of human adaptation" is culture. Thus. the paradigm of coevolution and "selectionist thinking” in homicide research may only be insightful as a perspective for portraying the adaptive significance of the behavior within the complex dynamics of interpersonal lethal violence. Description can be an insightful endeavor. Complex investigations of case studies which seek to elucidate and disentangle patterns consistent with the two mechanisms of inheritance may be possible depending upon whether or not the data can be accessed. Durham has done so in the study of polyandrous marriage patterns. adult lactose absorption. and the incest taboo. to name a few. In view of Alexander's (1979: 78) assessment of how to obtain true answers. one should ask: 1. to what extent are cultural patterns actually independent of predictions from natural selection and 2. how could patterns of cultural behavior be consistent with natural selection that do not do violence to our knowledge of the extent and 20 nature of learning. Thus. in exploring the problem of intrafamily homicide. such questions should eventually be addressed. The hypotheses to be tested are small preliminary links in a chain for further complex investigations. In Chapter 2. the literature discussing intrafamily homicide will be highlighted” The chapter will be divided into two parts. The content of Part I will briefly describe: homicide: major theories; family violence from a cross-cultural perspective; intrafamily homicide; marital homicide; child victims of homicide; and juvenile homicide offenders. Part II will cover literature which addresses homicide from the perspective of natural selection and examines the relational distance between perpetrator and victinn Also. an ecological/ evolutionary theory regarding family violence will be summarized. At the end of Chapter 2. the previous research will be very briefly discussed. The content of Chapter' 3 will discuss the research design and methods of the study. and the samples used in the study. In Chapter 4. the results will be presented. analyzed. and summarized. And, in the concluding chapter. Chapter 5. final conclusions and implications for future research will be discussed. CHAPTER 2 REVIEW 0 q E?! E H H ERAIDRE l 1 l l I Part I Homicide: Major Theories Theoretical writing on the subject of homicide is scarce. ”Sociological" explanations of homicide include: the cultural-subcultural; the structural; and the interactional. Theorists and others writing about behavioral science continue to conceptualize ”divisions" such as sociological and genetic explanations; however. the emerging and quite accepted view is that all behavior does result from both nature and nurture. not one or the other. Subcultural theorists argue that subgroups exhibit higher rates of homicide because they share norms and values that encourage the use of violence in interpersonal relations and that these values are learned and transmitted from one generation to the next (Wolfgang. 1958; Wolfgang and Ferrucutti. 1967). Structuralists argue that broad-scale social forces such as lack of opportunity. institutional racism. persistent poverty. and population density determine homicide rates. Cloward and Ohlin (1960) related opportunity and control structures to conflict in gangs. Van Den Berghe (1974) contends that resource competition leads to aggression. that when population 21 22 needs increase. competition for resources increase. and unless regulated. homicide rates increase. None of the above theorists specify how structural forces may affect the rates of family. friend. or stranger* homicides. Parker and Smith (1979) suggest that poverty is a more telling variable for explaining family and friend homicide. but not for other types of homicide. Interactionists focus on the interaction process of the participants and how the conflicting interaction escalates to the point of homicide. Each approach provides insight into the study of homicide. Some have attempted to blend explanations along with studying the impact that weapons possession. alcohol consumption. and drug use have upon the commission of a homicide. Yam 1 violence Studied in Cross-cultural gerspective Levinson (1989) studied a subsample of 90 cultural groups in the Human Relations Area Files. from 60 distinct regions of the world. Each culture group shares the same speech. and a very few are from modern nation-states. He did not look at homicide; however. his quantitative and qualitative analyses suggested seven conclusions about family violence: 1. Wife beating. physical punishment of children. and fighting between siblings are the few forms of family violence which are evident in many societies. and which occur frequently in 23 a great many. Infanticide. patricide. and killing of the aged are rare within all cultures. 2. Adult women are the most likely victims. while adult men are likely perpertrators and least likely victims. 3. The majority of people in the world have experienced or witnessed violence within their families. Physical punishment of children by caretakers is apparent in 74% of societies; sibling fighting is seen in 44% of societies; and wife beating is seen in 84% of societies. and is more common during early years of marriage. when young children are in the household. 4. Wife beating and physical punishment of children are the most common and frequent forms of family violence cross-culturally. 5. Greater frequency of wife beating is associated with societies where husbands have the most power economically and for decision-making. 6. More complex societies witness greater use of physical punishment of children. 7. Family violence is not universal. as 16 societies showed little or no violence. The culture groups displaying little or no violence come from all seven world regions. and represent these economic systems: hunter-gatherers. horticulturists. herders. and. agriculturists. Generally» all of these societies share the following traits: husbands and wives share in domestic decision making. wives have some control over the fruits of family labor. wives can divorce their husbands as easily as their husbands can divorce them. marriage is monogamous. there is no premarital sex double standard. divorce is relatively “—5. 24 infrequent. husbands and wives sleep together. men resolve disputes with other men peacefully. and intervention in wife beating incidents tends to be immediate. (p. 103) Levinson states that many of these findings are the opposite of that which were found regarding the causes of family violence. He also looks at several theories regarding family violence and how the data support or do not support these theories. For example. Culture of Violence Theory is partially supported for one model (general culture of violence). but finds little or no support for the model of a violence spillover effect. nor for a strong link between male aggressiveness and wife beating. Social Learning Theory is indirectly supported. general Systems Theory. specifically. Straus's positive feedback model regarding wife beating. is supported with limitations. Intrafamily Homicide Most studies on family violence have focused on the microsocial or psychological level of analysis. Gelles (1979) agrees with Straus's assessment that most violence which takes place in American families may be traced to the nature of the family and the larger society. not from individual aberrations. Another major concern has been to measure the rate of family violence and searching for patterns. The latter has resulted in two opposing views. There are those who argue that violence cannot be explained by social factors and those who uphold that violence is restricted to one social 25 group (e.g. poor. blacks). Gelles and Straus (1979) have identified. eleven factors which describe why the family is apt to engender violence. Moreover. the same characteristics have been pointed out to have the ability to also contribute to a loving environment. These include: time at risk; range of activities and interests; intensity of involvement. impinging activities; right to influence; age and sex differences; ascribed roles; privacy; involuntary membership; stress; and extensive knowledge of social biographies. Gelles asserts that the significant factor which allows violence within families is that they exist in a cultural context which tolerates. accepts. and even mandates violence outside and within the home. Wolfgang's classic study' of the victim/offender relationship in 550 criminal homicides in Philadelphia. 1948-1952. found that women were most frequently killed by relatives (usually a husband) and that males were most frequently killed. by acquaintances. A. ”family relationship" included relationships by consanguinity or legal affinity. but were in almost all cases very close (husband. wife. parent. child. sibling). Family homicides accounted for 25% of the known 550 homicides. Blacks and whites displayed similar victim/offender distributions with the exception that whites were three times as likely to be killed by a stranger. A significant difference was found in that females were 26 more often (51.9%) from the category of family relationship than were males (16.4%). The majority of criminal homicides occurred between similarly aged persons. which is expected. The 136 victims who were related to their killers included: 100 husbands or wives. 9 sons. 8 daughters. 3 mothers. 3 brothers. 2 fathers. 1 sister. and 10 other types of associations. Significantly. wives killed by a husband constituted 41% of all female homicides; whereas. husbands killed by wives constituted only 11% of all male homicides. An excessive use of violence was found to be associated with mate homicides. The majority of motives in husband/wife homicides were identified as a family quarrel or jealousy. The vast majority of mate slayings occur in the home and with greater use of weapons such as a knife. Wolfgang introduced the concept of "victim-precipitated" homicide to refer to cases where a victim contributes to his victimization. i.e. is the first to use physical force in a homicide incident. These factors were significantly associated with the 150 victim-precipitated homicides: black victims and offenders; male victims; female offenders; stabbings; victim/offender relationships involving male victims and female offenders; mate slayings; husbands who were victims in mate slayings; presence of alcohol; alcohol 27 in the victim; and victims with a previous arrest record. particularly. assault. Boudouris (1971) analyzed 6.389 Detroit homicides during 1926-1968. He describes his classification system ‘based on social interaction. with family (or domestic) relations at the top of the list of twelve. He found that the proportion of ”family relations" was 24.8% for the years 1948-52. remarkably similar to Wolfgang's result of 24.7%. He stresses that comparing same time periods is important. The proportion of family members involved in homicides was 29.5% (1603 of 5443). The lowest proportion was 13% for the years 1926-29. the highest was 28.1% from 1955-59. In an analysis of the legally married. common-law. co-resident. estranged. (n: divorced subdivisions. very little variation was found throughout the forty-three years. Each category showed similar risk for homicide over time. Ihl addition. a homicide was more likely to occur before an estrangement. rather than afterwards. Other domestic relations. which excludes married or divorced couples showed proportions ranging from 19.9% during 1955-59. to 29% during 1926-29. Much of later research findings agree with Wolfgang's findings. Zahn and Riedel (1985) again find that females are represented in greater proportion in family homicides. as victims and offenders. They looked at the period 1976-1978 for both the nation and eight 28 cities from four regions (Northeastern. North Central. Southern. and Western). It is only since 1976 that the FBI Unifornl Crime Reports record the ‘victim/offender relationship based on incident. The information identifying the relationship. however. is inferior for research such as Daly and ‘Wilson's and the present study. Approximately one fifth of all homicides. 1976-1978. were between family members. The authors state that this proportion has held steady through the early 1980's. Victims and offenders involved in family homicide share similar demographic characteristics. Victims and offenders had median ages of 33 in the nation and 32 in the cities. older than for other types of homicides. Male family homicide victims were a 57% majority in the nation; a larger proportion was apparent in the cities. Male offenders were a 65% majority. Several cities showed equal or near equal percentages of male and female offending. In the nation. whites accounted for 55% of family homicide victims. and 53% of female homicide offenders. Blacks accounted for 44% of both victims and offenders. Data for the cities reverses this pattern. showing black predominance as victims and offenders. At any rate. blacks are overrepresented for all homicides. The weapon most often used in family homicide was a 29 handgun. In the nation and cities very few family homicides were associated with a felony. Most family homicides occurred within the home. ranging from 60% to over 90% taking place in the home. Straus (1987) found that lower homicide rates are associated with higher percentages of intrafamily homicide. This suggests that rates of intrafamily homicide are relatively stable from one country to the next. Using nine data sets from the United States. Canada. and Denmark. a "primary group lag theory" was supported. Primary groups. especially a family. have certain characteristics which induce a minimal rate of violence and restrict serious violence. Flewelling and Browne (1987) found that 87% of women-perpetrated intrafamily homicides for the years 1980-1984 were against spouses and children. Blacks were represented by a greater proportion than in the population for all categories of intrafamily homicide. In over half of the homicides in which women killed their own children. the children were less than one year old. compared to 30% of males who killed children. In examining intrafamily homicide rates for nine regions of the country for 1976—1984. they concluded that the degree of variation increased as the relational distance moved from primary to more distant family ties. Silverman and Kennedy (1987) analyzed trends of homicide for 22 years in Canada. controlling for both 30 relational distance and gender relationships. They categorized relational distance from more intimate (spouse/lover) as the closest to other family relationships to friends and acquaintances to the least intimate of strangers. They found a recent proportional rise of homicides involving more distant relationships than the family. The rate of spouse/lover homicide was quite stable with other forms rising. Female victims of males in spouse/lover situations declined. while male victims of females remained reasonably stable. It is suggested that future research of gender relationship should place relational distance as the independent variable. Family homicides involve more younger victims. The number of female victims is significantly higher than that found in the U.S. The above study concurred with Plass and Straus in finding an overall decrease of intrafamily homicide. in the United States. for the period 1976-1984. Another study of homicide in North Carolina contradicts this decrease. Palmer and Humphrey (1982) found that the proportion of intrafamily homicides increased from 1972 to 1977 in an analysis of 985 criminal homicides. Their data were derived from incarcerated cases of first and second degree homicide and first degree manslaughter; therefore. caution should be taken in comparing to other studies which usually look at all 31 recorded homicides. For the years 1972. 1976. and 1977. 22.5% of homicides involved offenders and victims who were members of the same family. There were 12.5% marital homicides; 3.1% parents killing children; 1.3% children killing parents; and 5.3% of other familial relationships. Family homicide offenders are generally older than those who kill unrelated victims. Whites killed members of their families more often (26%); blacks next (21%); and Indian offenders. the least often (13%). Females killed family members two-thirds of the time they killed. males only one—fifth. A significant increase was apparent in cross-sex family homicides. with a 60% rise of females victimizing males and 33% rise of males victimizing males. Palmer and Humphrey (1982) note that suicide is much more often associated with family homicide. as has also been evidenced elsewhere. They propose that self- directed violence increases as the subculture of self- directed violence emerges as a greater force. a variation of the subculture of outward-directed violence. as envisioned by Wolfgang and Ferracuti. They state that since North Carolina's suicide rate is increasing and the overall criminal homicide rate is decreasing. then. perhaps. this direction is supported by a growing subculture of self-directed violence. That family homicide is also increasing at the same 32 time as suicide and the fact that homicide followed by suicide is almost always associated with family homicide warrants an investigation of whether family homicide is a pattern resembling self-directed violence and also influenced by a subculture of self-directed violence. By the logic of natural selection. family homicide directed against genetically related persons is maladaptive as is suicide. thus a connection might be discerned. Zimring et. a1. (1983) studied 151 intersexual homicides which occurred in Chicago in 1981. Victims who were killed by the opposite gender numbered 81 female and 70 male. Blacks were overrepresented and Hispanic surnamed victims were underrepresented. in proportion to their numbers in the population. White non-hispanic offenders committed family homicides in greater proportion than they committed other kinds of homicide. Over half of all intersexual homicides involved prior intimate or familial relationship. Only 25 of the homicides were between legally married individuals; many more were between boyfriends and girlfriends. The trend of intimate homicide held quite steady from 1962-1980. while the general homicide rate skyrocketed. Females used weapons more than males did. suggesting that male physical strength does not require the use of lethal weapons. Black females used knives 33 more than did white females. suggesting that cultural patterns help to determine the use or restraint of lethal weapons. They also tentatively hypothesize a "chivalry" or male restraint component among black males as black males participate in intersexual homicides only 3.4 times the white rate. rather than the tenfold greater rate for all homicides. Spouse homicide rates are held to be extremely low. They calculate a ratio of 1 homicide to 4715 violent acts for 1980. based upon a recent national survey. The authors speculate on the apparent ”taboo" for using lethal violence against women by Hispanic males who in general have very high rates of committing homicide. Finally. they discuss the connection between intimate and other lethal violence. They lament that the literature fails to address the social values that determine patterns of violence and their complexities across and within subcultures. It is argued that attitudes toward both family and. gender' are crucial determinants in intimate homicide; and that a general "culture of violence” does not explain high rates of intersexual violence. Silverman and Mukherjee (1987) studied 126 intersexual homicide incidents which occurred in three cities in Canada over a five year period. 1976-1982. Analysis of victim/offender relationship showed that 75% involved some degree of intimacy; 40.8% involved a 34 marital arrangement or prior marital bond; father killing child (4%); mother killing child (1.6%); child killing mother (0.8%); and other family relationship (5.6%). The major finding is that unstable (separated and divorced) and stable (defined as intact legal and common law marriages) relationships directly affect the rate of homicide for relational categories. The rate of homicide in the domestic-unstable category (0.17%) was three times that of the domestic—stable category (0.05%). Different weapons were dominant in the unstable group. with shooting being the most popular. Also. the gender ratio was different. as only males perpetrated murder in the unstable group. The authors claim that many previous studies missed the finding that marital stability is an important variable because they included divorced and separated individuals in one category of married individuals. However. Boudouris's evidence suggested the opposite: that stable marital relationships were more likely to end in homicide than unstable. Several authors find support for a stress factor associated with the incidence of homicide. Backman-Prehn et. al. (1988) found, a strong relationship between the State Stress Index and. the incidence of family and acquaintance homicide. The relationship held even when they allowed for the effects 35 of other major variables such as cultural support for violence. weak social control. economic deprivation. urbanization. and a large minority population. Kratcoski (1988) found support for stress. exchange/social control. and social learning theories of family homicide. Like many other authors have found: family homicides involve :more female and older offenders; more female and juvenile victims; majority of victims were married and killed by a spouse. Family homicides were more often spontaneous. often occurring during or after a quarrel. Daniel and Holcomb (1985) compared 44 domestic homicide offenders with 169 nondomestic offenders. They found that domestic offenders were more likely to be older. married. unemployed at the time of the offense. Childhood behavioral problems were more common among domestic offenders. A significant number of both groups came from disrupted families. Domestic murderers attempted suicide after the offense 17% of the time. and exhibited bizarre or psychotic behaviors immediately after the offense 55% of the time compared to only 25% of nondomestic murderers. In addition. 34.09% of domestic offenders received a pretrial psychotic diagnosis as compared to 11.67% for nondomestic offenders. Duncan et. a1. (1978) and Totman (1978) both describe that a family murder is tied to the 36 perpetrator's inability to escape from an unbearable relationship except through violence. Totman. in a psychosocial study of 50 incarcerated female homicide offenders. summarizes that she kills her mate or children at the following times: 1) when the relationship is felt to be destructive to them or their sense of identity; (2) when they feel they cannot share their concerns and thus get adequate help from significant others or the community; (3) when they have exhausted all other perceived alternative courses of action; and (4) when they have redefined their negative situation in a manner that calls for action not previously considered possible. Hagaman et. a1. (1987) developed a psychological profile of family killers based upon an interdisciplinary analysis of three cases (including a check of four other cases). Obviously. the tiny sample size prevents importing any real credence regarding the findings; nonetheless the findings are interesting. It was found that in all three cases. the perpetrator was a member of a cultural or social group that placed great emphasis on status and saving face and had made previous threats of suicide. All involved a failed love relationship and deep depression and all used drugs or alcohol at the time of the crime. Wolfgang (1978) and Straus (1986) found that most family homicides occur after a long history of assaults. and. after police disturbance calls. Wolfgang’ finds support for the subculture of violence model in that 37 homicide within the family and violent assaults are more common in lower socioeconomic groups. Both argue that improved family crisis intervention by the police and other professionals would help to reduce family homicides. Wolfgang suggests that even children. from a subculture of violence. may be socialized in the home and school toward nonviolence. given affection. recognition. rewards for being unharming. among others. Marital Homicide Howard (1986) describes that although husbands and wives kill their spouses in approximately similar proportions. husbands usually provoke the attack and wives often kill in self-defense. Browne (1988) compared 42 battered women charged with murder or attempted murder with 205 battered women who had not killed or attempted murder of their spouses. It was found that most women in the homicide group had no previous history of violence; and that the women resorted to violence after attempts to survive with an increasingly violent and unpredictable mate. Chimbos (1978) in a study of 34 Canadian spouse killers. found that most had little education. held unskilled or semi-skilled jobs, and that common-law married couples more often committed a homicide earlier during the course of the marriage than did the legally married couples. The author found these circumstances to be common in the scenario leading to a homicide: 38 quarrels over extra—marital affairs or sexual refusals. temporary separations. previous physical fights. and heavy alcohol consumption by both offender and victim. In addition. those apparently prone to marital violence very often had unsatisfactory and frustrating relationships with parents and had learned violent reactions during their childhood. Rod (1980) also found that men are often the aggressors in marital violence in Australia. Precipitating causes cited include the marriage relationship itself (the tension from fights and obligations); disputes over money matters. personal deportment; and. mercy killings. A. significant difference was found in that more American women kill their husbands than do Australian women. The battered wife syndrome as a legal defense is discussed by Buda and Butler (1985). Guerin (1983) discusses the battered husband defense and states that the battered spouse defense is not a license to kill their abusers. but appropriate in circumstances involving fear of death at the hands of the abuser. Fields (1978) found that police and prosecutors fail to protect abused wives. A policy of nonarrest only serves to enforce a husband's belief that he can act without fear of serious consequences. 39 Child Victims of Homicide Children probably experience higher rates of violence. from both parents and siblings. than is experienced among married couples (Straus.1986). One national study showed that 16% of marriages incur some kind of violence during the year. Several authors find that child victims are very often illegitimate preschoolers (Riehl. et. al.. 1978. and Kaplun. 1976). It has been found that many abusers come from abusive backgrounds themselves. and repeat the pattern. Steele (1978). describes that the child abuser may have the unique and primary power to transmit violence within the culture. Also. it is stressed is that no two abusers are alike; but that several psychological elements and a socioeconomic factor often precipitate incidences of abuse. Fein (1979). in agreement with others. found that families with a child homicide. unlike other abusing families. abuse more than one child. This particular finding is a tangible one for planning preventive intervention. In looking at 45 recidivist child abusing families. the family situation was chaotic. with 109 father figures sporadically visiting the families over a period of seven years. Mothers were the perpetrators in 80% of the fatalities and colluded with the father or paramour in 13% of the fatalities. Results indicated 40 child homicide is most likely in an abusing family when and where: the mother is the only continuing adult in the home; more than one child is abused; a chaotic family situation; the mother is the perpetrator of the abuse; and where there are many stress factors present. Zumwalt and Hirsch (1987) found that 15-20% of child abuse fatalities are part of the battered baby syndrome. Nonaccidental fatalities. equal in number. often result from a single or isolated occurrence and too often escape detection. Chrisoffel (1988) describes that homicide is the only leading cause of death for children under fifteen that has increased in incidence in the last thirty years. Children under eighteen have a cause of death as homicide 5% of the time. whereas. persons over eighteen only 1% of the time. Research shows that homicide is very commonly perpetrated against infants and that some adults punish their young children in a dangerously severe manner because of age-inappropriate expectations. I would like to point out that a certain amount of homicide of infants by mothers may be attributed to post-partum depression and/or psychosis. The literature did not mention this as a possible cause. However. this disorder has been documented for over 100 years. Women. within several weeks to several months after delivery of a child experience hormonal changes in their bodies. which sometimes. give rise to a temporary psychosis. 41 leading to murder of their child. In England. the legal statutes prohibit prosecuting a mother for murder of a child less than 12 months old. because of the commonness of this disorder. This disorder has been used as a defense by mothers in many places. Juvenile Homicide_0ffender§ Offenders younger than age fourteen commit murder at a neglible rate. Mones (1985) found. in an analysis of seven juvenile killers of their parents. that all offenders had been physically or emotionally abused by their parents. Several authors find a common factor of offenders bearing a relationship with an over-dominant mother (Walshe-Brennan. 1977. 1974; Russel. 1984; Haizlip. et. al.. 1984). Corder (et. a1. 1976) found that adolescents charged. with 'parricide significantly' differ from adolescent murderers of friends or strangers. These differences include: fewer typical adolescent sexual outlets and social relationships; fewer indications of aggressive response; more physical abuse by parents; more evidence of atypical sexual stimulation by parents; greater frequency of abuse of a passive mother by the father; family patterns of absent fathers; and more instances of amnesia of the murder act. Crucial environmental and sociological factors in 42 adolescent homicide include early‘ deprivation (of parental love. parental abuse or seduction. and exposure to violence. 43 Part II Hgmicide: Genealogical and Affinal Distance Between Victim=and Offender Probably the first to examine kinship relationships as revealed in conflict or crisis situations were Chagnon and Bugos (1979). They began with the assumptions that natural selection. inclusive fitness. and altruimn may supplement other traditional analyses of kinship behavior. and cite Sahlins (1968) observation that closeness of kinship is an important mediator of interpersonal relationships of many kinds in human societies. Kin selection theory generally predicts that: if you sustain the same "costs" to your potential fitness in helping a (1) close relative. (2) a distant relative. or (3) a non-relative. your inclusive fitness would be better served by aiding the relative that is genetically most related to you. In addition. ...as the costs to your own potential fitness increase. you would increasingly favor more closely related over more distantly related or unrelated individuals. providing that the benefits to the recipients of your aid remain approximately the same (pp. 221-222). They' measured. the actual genetic relatedness of individuals who participated in an ax fight in a Yanomamo village. Villagers dealt each other insults and fought with the blunt and of axes. with several injuries resulting. but no death. The fight began because a woman (with close ties to the local headman) had been beaten with a stick by a visiting ex-villager. 44 after she refused to give him a share of her plantains. The authors describe that the data were acquired over a period of twelve years. that the ax fight was filmed which provided the only accurate method to identify participants; and that the data were gathered long before an intention to analyze with an evolutionary biological approach. thus removing any suspected bias. It was found that supporters on either side of the ax fight were not random sets of individuals from the village. and that close genealogical relationship was apparent among those that supported the leader. Mohesiwa. They conclude that closeness of relationship measured genealogically was a mediator of interpersonal behavior in this conflict situation. In addition. they recognize that relatedness alone cannot provide an adequate explanation for the building of coalitions. and that affinity or alliance are determinants. Daly and Wilson suggest that homicide lends itself to the analysis of conflict because of a minimal reporting bias and because it is a ‘rare and final manifestation of interpersonal violence. They cite the growing evidence which confirms that evolutionary theories help to explain social motives and behavior of both animals and humans. They propose that: (i) genetic relationship is associated with the mitigation of conflict and violence in people. as in other creatures; and (ii) that evolutionary models predict and explain patterns of differential risk of family violence (1988a:519). -/_l 45 The broad category of ”relatives" killing each other in U.S. homicide samples have never exceeded a proportion of one—third. There are two studies where genealogical and marital relatives were distinguished. Marie Wilt's study of 1972 Detroit homicides showed that 19% were victims related to their killers by marriage compared to 6% by blood (Daly and Wison. 1988b; Wilt. 1974). A study of Miami 1980 homicides showed that 10% of the victims were marital relatives compared to 1.8% blood relatives (Wilbanks. 1984). The U.S. data do suggest that blood kin may be more immune from fatal interaction than non-blood kin. Other societies may reveal higher proportions of blood relatives. They note that certain agricultural societies where the family farm is not easily divided and one's brother is a rival. show significant numbers of fratricides (1988b). Of the 690 ”nonaccidental" homicides in Detroit during 1972. 508 were closed and a victim/offender relationship was known (Daly and Wilson: 1988a). Unrelated acquaintances numbered 243 (47.8%); strangers. 138 (27.2%); and ”relatives”. 127 (25.0%). Of the 127 victims related to their killers there were: 32 genealogical relatives (8 children of killers. 11 parents of killers. 9 brothers. 1 sister. 1 cousin. and 2 nephews). The great majority were not blood related. numbering 95: 80 spouses (36 wives and 44 husbands); 10 46 in-laws; and 5 step-relations. See Figure 2.0. The authors sought to compare a pool of potential victims of various relationship categories available to potential homicide offenders. The approach they chose confines analysis to cases where both victim and offender live in the same household. They compared the expected rate of homicide to the observed rate according to the average household composition by U.S. census. for Detroit in 1972. The resulting analysis showed an elevenfold greater risk of homicide for co-residing genealogically unrelated kin compared to genealogically related kin (Daly and Wilson. 1982. 1988a). They also looked at the distribution of relationships for co—accused homicide offenders compared to the distribution of relationships for victims and offenders. For homicide collaborators in Miami. 29.6% were blood relatives compared to only 1.8% of victims and killers. The finding has been that collaborative killers show a much greater degree of relatedness than do victims and offenders. in such samples as tribal horticulturists. medieval Englishman. Mayan villagers. and Urban Americans. Therefore. it appears that homicide co-offenders come together not from random opportunity. but also in the direction which would preserve their gene pool. 47 0.8 spouses 0.1 relatives 0.0 “off-pun." 0.4 ”parents” 1.0 other rdative A=observed B=exp C=risk (obs /exp) - Series A \\\\\\‘ Series a Series c FIGURR 2.0 Risk of homicide by relationships. considering only those cases where victim and offender co-reside. for Detroit. 1972. The Average Detroiter 14 years old and older lived with 3.0 people. ”Other relatives” are mostly siblings. but census data does not allow separating these from more distant relatives. Offspring. parents. and other relatives include step and in-law as well as natural. (From Table 2.1. Daly and Wilson. 1988a: 23) 1.3." fi.fl.3fleflu1fl.fifl.;flsiugflpflu3 .4... . tn 5. .... FIFTFIM Fm...” t...“ 1...... w, .._ .lnIuiu .1!"an . ... .. r flint Ewe a 5%, . a '1. ‘1. .‘ IF' 48 relatives include step and in-law as well as natural. (From Table 2.1. Daly and Wilson. 1988a: 23) relation to the fact that no genetic posterity’ may accrue from these children. Differential rates of violence have been found in homicide of natural children compared to step-children. The evidence points to a much greater risk of homicide for step-children. In addition. child abuse is prevalent among step-families. and seems to be independent of factors such as low socioeconomic status. maternal youth. family size. or personality characteristics of the abusers. Abusive step-parents abuse step-children at the same time as showing no abuse to their genetic children. An English sample of "fatal battered baby cases" showed 15 killed by stepfathers and 14 by genetic fathers. although fewer than 1% of same-age babies lived with step-fathers. U.S. data for 1976. showed that a child living with one or more substitute parents was 100 times more likely to be killed than a same-age child living with genetic parents. Canadian and Australian samples attest to similar statistical results. Daly and Wilson observe that exclusively monogamous couples retain genetic interests greater than that of blood relatives. The problems which may arise to disrupt this harmony include: the possibility of extra-pair reproduction; and the partners nepotistic interests in their distinct kin. 49 Daly and Wilson and others have recognized that the "majority of lethal violence occurs between unrelated men and involves competition for material. social. and sexual resources that were directly related to reproductive success in the environments of human evolution" (1988b: 524). Males compete more dangerously in pursuit of their reproductive interests. Female reproduction is resource-limited. whereas. the reproductive capacities of females are the limiting ”resource” for males. They argue that men take a proprietary view of women and their reproductive capacity. universally. as for example the widespread provision of sexually asymmetrical adultery laws. The data suggests that the prime motives in marital violence and homicide are "jealousy.” concern for the wife's fidelity. and her intent to leave. Estranged wives appear to be especially at risk for being killed. The authors also discern that age of the couple may be an important variable. Youthful wives and then again much older wives appear to be at greater risk for homicide. Men may perceive youthful women to be more attractive to rival men and shorter marriages may be less stable. Even when controlling for the age of the man. young females married to older men are still at greater risk to be killed. thus removing a bias for the younger homicidal male group. 50 They state that analysis is needed which unconfounds the variables of the two parties' ages and marital and reproductive histories in order to elucidate their separate contributions to risk. .Australian and Canadian data show elevated risks among common-law unions compared to legal unions. with a probable similarity in the U.S.; however. U.S. Census data fail to distinguish the two types of unions. Trivers provided a new view regarding parent - offspring conflict (1974). This view sees the parent and the offspring in a conflictual situation because a parent's efforts and allocation of resources which maximize genetic posterity infrequently match the child's. Such "maladaptive" behavior as weaning conflict and developmental regression by a child upon a birth of a sibling are explained by this theory. Infanticide may also be a behavior that parents would incline toward when the following are present: there is doubt that the child is a true offspring; poor offspring quality; all those extrinsic circumstances. for example. lack of food and/or social support. and overburdening demands from older siblings. All of these factors would have impinged upon the child's chance to survive during the evolutionary past. The ethnographic record supports the presence of most of these in nonindustrial societies where infanticide occurs. The factor of post-partum. psychosis leading to 51 homicide of an infant should also be kept in mind; as this disorder is directly related to hormonal changes in mothers after delivery.v Therefore. a certain number of homicides of infants may be directly attributed to this disorder. which arises from body chemistry changes. Also. since fathers do not experience this disorder. a comparison of rates of homicide should take this into consideration. Evolutionary considerations suggest that the age of the mother and of the child are important to her willingness to commit homicide. As a woman ages she would disincline toward infanticide. because of her diminishing reproductive capacity. and the data supports this. An older child would also be more valued after passing the high mortality stage of infancy. Canadian data support the prediction that. mothers ‘would. kill offspring at the earliest because of her shorter reproductive life-span than a father; that fathers would kill more evenly across the child's ages. one reason being. that later phenotypic and paternity questions may be aroused. Parricide may also be predicted by age-specific rates. An offspring would be expected to disvalue an elderly parent more than a younger one. The data supports this with the finding that parental age at the child's birth have opposite effects on the rate of homicide against each other. 52 Daly and Wilson propose that the application of an evolutionary model to the study of violence helps to explain predictions of homicide and their adaptiveness. on average. not specifically. They insist that although specific acts may be maladaptive. particularly in novel environments. ”selection has shaped the social motives. emotions. and cognitive processes underlying them" (1988b: 523). As mentioned earlier. the main problem with Daly and Wilson's line of reasoning is that they fail to appreciate that selective pressures do not emerge solely through a biological process. They claim that the paradigm. is :neither direct nor simple; however. the argument appears simplistic. Evidence of adaptive significance for a particular behavior is not evidence of causation. as the behavior may be due to cultural evolution or inheritance. as Durham has argued. The use of evolutionary models in describing diverse and predicted patterns of homicide is nonetheless insightful. It should help to forge new ideas in criminology. but. the approach should also recognize the complexity of the dynamics involved. They conclude that evolutionary models show differing risks of homicide. as a function of the parties' ages. circumstances. and specific relationship to one another» Genetic relationship does seem to predict a decrease of violent interaction. They also 53 suggest that further research should attempt to test similar hypotheses with regard to less extreme forms of conflict and with positive measures of harmony. An Ecological/Evolutionary Perspective Burgess and Draper (1989) present an explanation of family violence in which they argue that marital violence and child abuse (both of which are present in society universally) may have been adaptive behaviors in certain past environments. The evolutionary concepts of paternity and parental investment are purported to help explain the presence of discord between spouses and between parents and offspring. It is suggested that the probability of marital violence increases as ecological instability increases. For example. underemployment. financial pressures. anxiety. and alcohol abuse are thought to be "markers of ecological instability or its consequences" (1989: 59). They also present an explanation of how child maltreatment may have occurred in unstable ecological circumstances to optimize the reproductive fitness of parents. Mating and parenting behavior are said to be very sensitive to unstable ecological factors. Burgess and Draper do not present any empirical data to support their explanation and make no mention of homicide. Nonetheless. they do present a convincing explanation of how ecological instabiltiy. biological. behavioral. and cultural selection may operate in the 54 dynamics of family violence. Their explanation is very similar to Daly and Wilson's and to Durham's Coevolution. but does in fact add the concept of ecological instabilty as being important to the probability for family violence to occur. either now or in certain past environments. Indicators of ecological instabilty have been looked at by other authors. however not in combination with an evolutionary perspective. Discussion_pf Previous Research Levinson's findings suggest that family violence is not universal. but that more complex societies exhibit greater amounts of family violence. such as the punishment of children. The general findings regarding family homicide are that economic and other stressors are associated with the occurrence of family homicide. In addition. various relational distances. age. and gender appear to predict risk for homicide. Particularly. women are more at risk for being victims in family homicide and a far greater number of homicides occur between spouses and genetically unrelated family members. CHAPTER 3 BE EARCH DESIGN QE 8 up: -- ._..___. The main obstacle to research of intrafamily homicide is the lack of detailed data. As Loftin. et. a1. (1987) argue there is also a need for a standardized method for measuring relational distances. They presented their ”attribute coding form" which entails simple binary decisions and fourteen consanguineal categories. thirteen affinal categories. and eight miscellaneous categories. This would be used by the researcher who must meticulously hand gather data from police reports. medical examination reports. etc. There were a few errors in Loftin's list of relational distances. for example the ”half sibling” category should be in the consanguineal list. not the affinal list. *1 The Samples The data set for all the family homicides which occurred in Detroit. during 1972 and the data set for all family homicides which occurred in Miami. during 1980. both previously gathered by Wilt (1974) and Wilbanks (1984) and described in Daly and Wilson's book will be utilized. A data set of! family homicides from the city of Chicago that will be used to retrieve a sample was also obtained. It is stored at the Consortium for Political 55 56 and Social Research. and is titled "Homicides in Chicago. 1965-1981.” by Carolyn Block and the Ilinois Criminal Justice Information .Authority. It contains information on every homicide. in the Chicago Police Department murder analysis files for the years 1965- 1981. (12.875 cases). It meets the definition of homicide cases that I am interested in. i.e. prior to prosecutor decisions. .It was found in the Spring 1989 catalog entitled "National Archive of criminal Justice Data." The data set contains 54 variables and includes 91 specific relationships of the first offender to the victim. which do specify. for the most part. the genealogical or affinal relational distance. The drawbacks are that the data set does not contain any socio-economic variables regarding each case of homicide and that it is a Class IV data set. meaning that it is in the condition as received by the Consortium. prior to it being checked or corrected» The data set does contain variables such as: age. race/ethnicity. and gender of victim and offender. motive. previous offense. victim participation. child abuse. drugs involved. location where body was found. police area. police district. and address of homicide occurrence. The code book also contains 87 frequency tables pertaining to the variables and a bibliography of 18 references which have utilized the Chicago homicide data 57 set. Methods and Hypotheses Durham's coevolutionary theory predicts that five modes of relationships between genes and culture are major patterns in cultural and behavioral change. These include: 1. genetic mediation; 2. enhancing relationship; 3. opposing relationship; 4. neutral relationship; and 5. cultural mediation. One mode or pattern which will be focused upon in this thesis is the "opposing" relationship which is part of the genetic mediation mode. An opposing relationship is displayed and implied by the pattern of homicide wherein ones' consanguineal relatives are killed at a higher rate. The other offender/victim relationships whiCh bear IN) genetic relatedness may display a "neutral" relationship because the homicide of a non- consanguineal relative does not directly select against the individual's genes or relative's genes. This analysis is rather simplistic; however. for the present preliminary type of research design and data samples available -- will be pursued thus. Three of the four hypotheses below are derived from the prediction of the "opposing relationship" which may also be referred to as natural selection theory. The hypotheses to be tested include: *For the Chicago data set of 1965-1981 family homicides: 1) The frequency of intrafamily homicide is less 58 common among genealogical or consanguineal (blood) relatives. 2) The frequency of intrafamily homicide is expected to be greater among non-genealogically related individuals and/or where the relationship is by affinity (marriage) only. 3) When comparing the actual frequency of homicide to the expected frequency based on greater or lesser interaction time of the offender and victim. the actual frequency is predicted to be different. This concept is discussed below. 4) Three geographical urban areas during different time periods are predicted to show rates and patterns of family homicide reflecting more homicide occurring among non-genetically related family nembers [Chicago (1965- 81). Detroit (1972). and Miami (1980)]. Operationalizing and testing of the hypotheses are difficult endeavors. The dynamics involved in the incidence of an intrafamily homicide are complex. Nonetheless. an attempt will be made to test the hypotheses by including a recognition of those complex variables. Firstly. I recognize that the incidence of homicide between various victim/offender relationships very probably does not occur with equal probability due to many factors. The one factor that will be recognized herein is the great variation of interaction time spent 59 between those two individuals. The homicide of a relative is not an independent event as there are many variables involved in the dynamics of this behavior. Therefore. merely comparing the incidence of genealogically related verses affinally related offender/victims in intrafamily homicide is not an adequate method to test the hypothesis. Daly and Wilson (1988a) did take into account the potential victims living in the same household for homicide cases where both victim and offender live in the same household. However. such a nwmhod does have objectionable features because Census data is inadequate and one loses data regarding homicides of non-coresiding victims and offenders. The interaction time spent between family members can be understood as being a specific measure for greater risk for conflict and/or for harmony. identified by Straus and Gelles (1979) as ”time at risk.” Interaction time between individuals and within the diverse lifestyles of families varies greatly. For example: 1. Spouses will generally spend more time together than other family members. in fact for a lifetime. if a divorce or separation does not occur. 2. Children generally spend greater amounts of time at school and with friends. and very often will move out of the home in their late teens or early twenties. 3. Some families may see their grandparent or uncle or aunt once a week. Whereas. another family. may see them five times a year. or perhaps five years may separate visits among relatives separated by miles or separated by discord (e.g. not socializing). 60 4. Some families live in typical nuclear family situations. while others live at some time or another with a grandmother. or nephew. (Often. grandparents (k) raise grandchildren. after a daughter has died or happens to be drug addicted. In some countries. the typical arrangement is the extended family occupying a single household. 5. Since divorce is so common in the U.S. but not in some other countries. it may be expected that the pattern of family homicides would differ cross- culturally. and over time within the U.S. (since the time divorce became more prevalent). Interaction time is affected because many children are raised in female headed households and with step-parents. and have parents who marry again and have children (half-siblings). These step and half siblings generally will not spend as much time together as would full siblings both because of there being a shorter time span from when the relationship began and/or they do not live in the same house. In addition. the "bond" that may develop between full siblings may in the instance of half and step-siblings co—vary with the amount of conflict (which is commonly associated with step-families) and with the age at which the individuals became step-siblings. 6. The economic well-being of a family may also affect the interaction time. An impoverished family may live in very close quarters and may see other relatives more often because of their needs. In fact. they may live as an extended family in one household. 7. The occupation of the parents may affect whether they have to move (e.g. a military family will generally not see their relatives as much as a family where all live in the same city). It may be true that more middle class and upper class educated people are mobile from one part of the country to another. This mobility and its influence upon the interaction time spent within families would also impact the occurrence of family homicide. There are many other variables which might affect the "time at risk." In order to include this variable into the analysis of the problem. I will assign five levels of probability for ”time at risk" to five groups of victim/offender relationships. Percentages from 80% down to 20% will be assigned based on greater 61 interaction time to lesser interaction time. This will be based upon a very general and hypothetical concept of how typical American families live and interact. including divorced and step-families. In a diverse city such as Chicago. it may not even be useful to attempt to use census data to obtain percentages of female-headed households. children living with step-parents. and so (XL The data set does not provide income level or household composition on a case by case basis. nor even income for the various sections of the city. Individuals will be arbitrarily assigned percentages indicating their greater or lesser time at risk based on the continuum of physical proximity. i.e. greater interaction time to lesser interaction time. Those individuals in the immediate family (which will be referred to as Group A) will be assigned the higher probabilities for "time at risk": husband/wife. ex-husband/wife. parents/children (including adopted) will be assigned: p = 80%. Spouses. very obviously. usually do spend more time together. except for exceptional cases. Ex-spouses will be included in this category because one has no way of knowing the length of the marriage or the time that has elapsed since a divorce. Parents and children spend greater amounts of time together. generally. especially' mother and children. 62 Full siblings will be assigned: p = 70% (Group B). During their childhoods. siblings spend a great deal of time together; and upon growing up. they move out of the house. get jobs at various locations. sometimes far from their parents. and will generally begin to spend more of their time with others. including their own boyfriends/ girlfriends. spouses and/or children. Step-parents/step-children and half and step- siblings will be assigned: p = 60% (Group C). These individuals will generally spend less time together than the previous categories. both in quantity and quality. One reason is that step-relationships begin at a later date than birth as with biological parents/offspring. Also see point # 5 of how families members interaction times vary. Other individuals will be assigned lesser probabilities for "time at risk”: all other genealogical relatives. affinal relatives. and legal guardian/ward: p 50%. It is difficult to generalize regarding this category; however. in a continuum for "time at risk”. they wouLd probably spend a.little less time together than the previous category of individuals. Foster parent/child and any other unspecified ”family" member will be assigned p = 40%. To summarize. these groups consist of: Grgup A (immediate family): p = 80% Group B (full siblings): p = 70% Grgup g_(step-parents and step-children): p = 60% Group D (other genealogical and affinal 63 relatives): p = 50% Group E (foster and any other unspecified): p = 40% Each category' of relational distance connotes a number [i.e. r' (the coefficient of relatedness). r = the fraction of genotype shared or zero]. We are interested in looking at the incidence or frequency of homicide. We will not call this a dependent variable. since causation is not to be presumed between the variables. The other variables are the various relational distances between victim and offender. and the factor of interaction time spent between the victim and offender. It should be mentioned that the data set does not show an offender/victim relationship between both first and second. or third offender. The data set does contain information (n1 the second and third offenders. but not specifying the victim relationship. Therefore. the findings in this thesis only pertain to first offender and victim. Those cases where multiple offenders were involved will not be known. The variable which shows the relationship between offender and victim also shows the incidence of homicide. This variable has been recoded into 3 variables. The first variable consists of 57 different categories of relationship. ordered from unrelated and affinal to genealogical. Those relationship categories where degree of 64 relatedness was not clarified as to whether blood relationship was present or not present have been presumed (for the purposes of analysis) to have been half blood-related and half not related. These cases will not be excluded because of the high number present. For example. aunt/uncle with niece/nephew. who may be blood-related or merely related by marriage. were split into two groups. (i.e. one half of the occurrences were put into the genetically related (r = .125). and the other half were put into the unrelated (r = 0.0). Categories wherein an uneven number occurred have created slight tnases. In another example. 28 cousin homicides. who were not specified as to whether they were first. second. or third cousins. were split into two groups: half (14) to a group with r = .125 and the other 14 to a group with r = .01. The second variable to be analyzed consists of the same 57 categories collapsed down to five groups wherein each group shares the same fraction of genotype of .5. .25. .125. .01. and .0. The third variable which will aid the analysis is a variable which describes the interaction time spent typically by family' members. .All 50 categories of relationship are again recoded and collapsed down to five groups from high interaction time to low. Since the frequency tables show that the data do not approximate a normal distribution or bell curve. but 65 show that the great majority of homicides occurred toward one end of the scale of order. parametric or interval level tests are deemed inappropriate (Siegel. 1956: 19). One test that will be used is the Kolmogorov- Smirnov or K-S. one sample nonparametric test for ordinal data (Siegel and SPSS-Xfi User's Guide. 1988: 733). This is a test of goodness of fit to see whether there is agreement between the distribution of the observed values with some specified theoretical distribution. Each of the following two variables will be tested as described: 1.) Variable (Relatord) H° = there is IN) difference in the distribution or incidence of homicide for the 50 categories of relationship. for the entire sample of 17 years. Chicago. January 1965 - December 1981. H1 = It is predicted that genealogical relatives sustain less risk for homicide compared to affinal relatives or unrelated individuals. This is not a statement of causality but one of description and positive or negative association for greater incidence of homicide. 2.) Variable (Fractr) H° == There is no difference in the distribution or incidence of homicide for the five groups wherein each shares the same fraction of "r” genotype. H‘ - It is predicted that the groups with greater fraction of genotype shared display infrequent homicides compared to those sharing 66 no relatedness. The level of significance that will be used is .01. These hypotheses will be tested upon the entire data set of intrafamily homicides which took place in Chicago. from 1965-1981 (N a 1.977 / 12.875). N actually should be 1.982. as a discrepancy with the SPSS—X output was discovered. This problem is discussed below. Rather than take a subsample from the data set which covers 17 years of homicides. it was decided to use the entire sample. which gives us a broader picture of the various offender/victim relationships in family homicide. The population we are interested in is actually all the family homicides which have occurred in the U.S. and other countries. In the checking of the data from the K-S and frequency tables SPSS-X computer runs it was found that the frequencies were not matching the frequency tables in the "Codebook" for the data set. Differences up to 10 and zero compared to the numbers in my output were detected. The total number of family homicides in my computer runs showed 1.977 homicides. compared to 1.982 homicides in the codebook. The consortium was contacted. who checked which frequencies were accurate. my computer runs. or the codebook. They told me that the codebook frequency values were correct. Rather than attempt to use SPSS-X again with a data set which was not producing output accurately. and not 67 knowing exactly' what the problenl was. the remaining analyses will use descriptive statistics. manual calculations. and/or other software statistical programs. where the correct values may be input. Several computer programs. which can be run on a personal computer‘ ‘were obtained. titled "Mystat" (Systat. Inc.. Evanston. IL.); ”PwrStar” by HyperstuffWare and "STAR" (Statistical Analysis for the Researcher by Antonio Gino. Ph.D.). the latter two obtained from Public (software) Library. Houston. Texas. Also. any other hypotheses that I wanted to study. such as an analysis of the incidence of homicide over time. from 1965 to 1981. or an analysis of ethnicity. could not be carried out; as the output that I obtained showing the incidence of family homicide for each year. is not reliably' accurate. The: codebook's frequency tables did not show family homicides over time. nor any other statistics for variables regarding family homicides. It has come to my attention by personal communication with the author of the Chicago data set. Carolyn Block. that ethnicity is an important variable as her latest studies are showing that for example. Hispanic spouses and boyfriend/ girlfriend homicide rates seem to be lower than that for other ethnic groups. However. this variable will not be considered in this thesis. The following analyses will be done in 68 order to arrive at conclusions regarding the hypotheses: Hypotheses_ £_ 1. i- 2. and £_ 3. Descriptive statistics. of frequencies and relative percentage frequencies for the 57 relationship categories. and for the five categories of genetic distance will be shown. Percentages and frequencies for the five groups A. B. C. D. and E. with high interaction time to low interaction time. will be compared to the results for the categories based upon genetic distance. Chi Square Test will be run to see if the null hypothesis for the dual hypotheses 1 and 2 (that the homicide frequencies are equally distributed) should be accepted or rejected. A Friedman Test. a non-parametric analysis of variance. will also be run on the variables. wherein all the values are separated into 5 variables. each under one of the ”r" categories. A Chi Square Test will be run to see if the null hypothesis # 3 (that the frequencies are distributed. according to interaction time) should. be accepted or rejected. An attempt will be made to see if linear regression and polynomial regression analyses help to define the pattern of homicide. For 57 cases. the X value will be the r values (0. .01. .125. .25. .5). the Y value will be the incidence of homicide or relative percentage frequency. Y = N/1.982 total family homicides. These pairs will then be utilized. to run linear regression and polynomial regression. Hyporhesis fl 4. Descriptive statistics. pie chart. 69 and Friedman Test (a nonparametric test for ordinal scale data) of the three samples from Chicago. Detroit. and Miami. The values to be compared will be the five relational distance categories. because detailed information regarding the offender to victim homicide is lacking. especially. for Miami. In addition. since Detroit and. Miami data are derived from. one year's incidence of homicide. quite a few of the offender to victim relationships are not represented as they are in Chicago's data set for 17 years. BQTAILQE *1 This is because half-siblings share 0.25 fraction of genotype with their half-siblings. the same closeness as grandparent with grandchild and aunt/uncle with a blood-related niece/nephew. To summarize Sewall Wright's index of consanguinity. r = relatedness; persons A and B may be related by a certain fraction. of genotype. See Sewall Wright's "Coefficients of inbreeding and relationship” (1922) and "Killing Kinfolk" by Daly and Wilson (1988a: 34- 35). CHAPTER 4 AHALX§1§ 9E EE§QLI§ Presentation of Results Family homicides which occurred in Chicago 1965— 1981 numbered 1,982 out of a total 12.875 homicides. Acquaintance homicides predominated with 46%; second in frequency' were stranger homicides. and thirdly. were family homicides. with 15%. See Figure 4.0. Chicago homicides increased rapidly from 1965-74 and from 1977- 81. However. it appears that most family homicide remained relatively stable throughout the 17 years (Blockzl986). Results_pertaining_to Hypotheses # 1 and # 2 It is predicted that the frequency of homicide is less among genealogical or consanguineal (blood) relatives; likewise. the frequency of homicide is expected to be greater among non-genealogically related individuals and/or where the relationship is by affinity (marriage) only (for example: spouses. step- relationships. in-laws. etc.). The results from the SPSS-X computer runs will be briefly described. Although the frequencies of homicide for the various relational distances were not accurate (matching the numbers in the codebook) in the SPSS-X computer runs. the derived result most probably varies 70 71 sundown tam 440 a dfiultdmu Luz mystery/mime 1.997 : stranger 2.938 7 I / /. /’ ./////’ acquaint-nos 6.067 722/2 % of homicides # of homicides FIGURE 4.0 Frequencies and percentages of family. acquaintance. stranger. and mystery/unknown homicide. which occurred in Chicago. 1965—1981. mu X311 . 72 little from what would have been derived from a ”clean" or reliable data set. This is because the numbers varied very little. and the pattern of distribution was very much the same. The Kolmogorov-Smirnov test compared a uniform distribution to the data set's homicide frequency distribution for 50 relational categories. The result showed that the null hypothesis. stating that all the relational categories have equal distributions of homicides. should be rejected. Therefore. the research hypothesis. stating that various relationship categories exhibit different frequencies of homicide. is supported. K-S tests upon the variable of five groups of relational distance (0. .01. .125. .25. .5) and the variable wherein the Groups A through B were ordered from low interaction time to high -- also resulted in the null being rejected. An analysis of the accurate numbers from the data set include the following results. There is as expected a higher frequency of homicides among unrelated individuals and those related by marriage in comparison to the frequency among genealogically related individuals. Figure 4.1 presents a bar chart of the number of family homicides for each of the 57 offender/ victim categories. Each category of relationship shown on the chart had at least one homicide during the 17 years. The offender/victims are ordered from unrelated. 73 - # FREQ. or HOMICIDE # Frequency of Homicide 400- 300 200 100 0" . '7 IIIHW—F. ABGDEFG-ll WW1234567891111111111222222222233 .......p 0123456789012345678901 lst Offender to Victim _ FIQLEE 4.1 Number frequencies of homicide for 57 relationships of first offender to victim. Chicago. 1965-81. *assumed to be non-blood related. 5m A=FOSTER CHILD/PARENT. BaLEGAL GUARDIAN WARD. £8 BROTHER IN-LAW/SISTER IN-LAN. Q-SISTERS IN-LAW. 3' BROTHERS IN-LAW. §=SISTER IN-LAW/BROTHER IN-LAW. Q. FATHER IN-LAW/DAUGHTER IN-LAW. fl-SON IN-LAW/FATHER IN- LAW. l-FATHER IN-LAW/SON IN-LAW. £=DAUGHTER 'IN-LAW/ MOTHER IN-LAW. _K_=SON IN-LAW/MOTHER IN-LAW. L=MOTHER IN- LAW/SON IN-LAW. M=STEPBROTHERS. NBSTEPSON/ STEP-FATHER. Q-STEPFATHER/STEPDAUGHTER. £=STEPMOTHERI STEPSON. Q'- XWIFE/ XHUSBAND. R=XHUSBANDl XWIFE. §=COMMON LAW WIFE/ HUSBAND. I=COMMON LAW HUSBAND/WIFE. U=WIFE/ HUSBAND. IE- HUSBAND/ WIFE. H3 NIECE/AUNT*. }_{=AUNT/NIECE*. IaNEPHEW/ AUNT“. Z=AUNTINEPHEW*. JENEPHEW/AUNT". 2=UNCLElNEPHEW*. g-UNCLE/ NIECE“. i=COUSINS. §=NIECEIAUNT. §=AUNT/NIECE. ZINEPHEW/AUNT. §=NEPHEWIUNCLE. 2=UNCLElNEPHEW. _1_9_=NIECE/ UNCLE. 1_1_=UNCLE/NIECE. QIICOUSINS. L3=GRANDDAUGHTERI MOTHER. LEIGRANDSON/ MOTHER. _L§=GRANDSON/ GRANDFATHER. 1__6_=GRANDFATHER/ GRANDSON. _]J}HALF BROTHER/SISTER. L3! HALF SISTER /BROTHER. 1_9_=H.ALF BROTHERS. 2_Q=SISTERS. ll- BROTHER/SISTER. AZ=SISTERlBROTHER. 2_3=BROTHERS. ZA=SONl FATHER. fiFATl-IER/ SON. 268DAUGHTER/FATHER. 27-FATHER/ DAUGHTER. 28=SONlMOTHER. _2_9_-MOTHER/ SON. fl'DAUGHTER/ . MOTHER. fl-MOTHER/DAUGHTER. l 74 on the left to increasing genetic distance or shared r. on the right. Several possible relationships are missing from the Chicago data set's occurrence of homicide. to name a few. grandmother killing grandchild. step-sister homicides. and. step-son killing step- mother. Homicide between spouses. in particular. have the highest number of homicides during the 17 years in Chicago. The marital homicides account for 65.9% of the total family homicides. The total marital homicides number 1.308 homicides including these offenders/ victims: husband/ wife. wife/husband. husband/ wife common-law. wife/husband common-law. ex-husband/ex-wife. and ex-wife/ex-husband. There appears to be a difference between the common-law married couples rate of homicide compared to legally married couples. In both cases legally married couples killed each other more often. which would be expected. since. there are probably more legally married couples in society. The rate of victimization by gender was almost equal: 641 husbands killed and 638 wives killed. The interesting observation is that there were 122 more wives killed by husbands than common-law wives killed by common-law husbands. As for the wives killing husbands. the difference was only' 31 more than the common—law couples. Divorced spouses killed each other at a very low rate. numbering 29. and both ex-husbands 75 and wives victimized the other in almost equal numbers. 13 and 16. The only other categories of relationship which approach the high numbers evidenced among spouses are the 80 brother/ brother homicides. 71 brother-in- law/brother-in-law homicides. 63 father/son homicides. and. 61 mother/son homicides. Table 4.0 shows that there were 31 suicides associated with family homicides. This is a 46% proportion of the total number of suicides which occurred associated with a homicide in Chicago. The table also shows the number of homicides for each of the relational distances. parent/child homicides. step- parent/child homicides. and. a breakdown of homicides perpetrated by males. by females. and male against male. Mothers and fathers seem to have killed offspring with the same frequency. however. both did kill almost 50% more sons than daughters. Males committed homicide a little more than did females. 16% more. Step—parent/ child and step-child/parent homicides were quite low in number. 35. compared to parent/child and child/parents. which numbered 210. This may or may not be significant. We would have to know how many step-families there were compared to other families. Three-quarters of the murders of young children began as child abuse. As one ‘may see from ‘Table 4.0. the relational distances of 0. .01. .125. .25. .5. show a greater 76 number of homicides in the zero relatedness category. The next greater number is among those individuals sharing 50% of their genes. Figure 4.2 presents a barchart/scatterplot of this variable. The logarithmic curve for the 5 degrees of relatedness is plotted. This curve describes an increasing to decreasing. or positive to negative relationship. from zero relatedness (on the left) to the highest degree of relatedness, .5. (on the right). A Chi Square Test was run in an attempt to test the null hypothesis that the frequency of homicide is equally distributed. No result was obtained. only a warning that significance tests are suspect when more than one fifth of fitted cells are sparse (less than 5). A Friedman Test was run to compare the frequencies for the separate values under each relational distance (0. .01. .125. .25, .5). Each of the relational distances was made into one variable. and then these five variables were analyzed. There was a probability of .000. assuming a chi—square distribution with 4 degrees of freedom. Thus. the null hypothesis that there was no difference may be rejected. Linear regression was run for the variable of 57 categories of relationship. Since the scatterplot is extremely non-linear. the values were split into two groups. A linear regression was run for the zero relational distance values separately from the 77 TABLE 4.0 Frequency and relative percentage frequency for certain categories of family homicide. Chicago 1965-81. RELATEDNESE i 9E _QMl§l_ES WEL_IIEE § EBEQ of lst offender/victim in FAMILY HOMICIDES r = 1.0 (suicide) 31 r = 0.5 395 19.92% r = 0.25 17 .85% r = 0.125 43 2.16% r = 0.01 14 .70% r = 0.0 1.513 76.33% father/child 94 4.74% mother/child 98 4.94% child/parent 118 5.95% step-parent/child 12 .60% step—son/father 23 1.16% male perpetrated 1.134 58.0% female perpetrated 816 42.0% male/male 383 19.0% 78 lineunwelfieqwuuwn 8090* 6090* 4090* 2090* 0’ JJJ'llllllJlqulllilLJll11111111111 92222932933i;ii3.133éiéiiéii'ééié‘iiiiiiiié‘ Degree of Relatedness — Regression (.09. Flat * Freq. of Homicide FIGURE 4.2 Relative percentage frequencies of homicide for the relational distances of 0. .01. .125. .25. .5 with plot of logarithmic curve. Chicago 1965-1981. 79 consanguineal relational distance values. However. the results were very poor; linear equations fitting 80% of the values (adjusted squared nmltiple regression) were not produced. nor even were they close to 80%. .An attempt (x) do multiple regression analysis for the 57 relational distance categories and the interaction time variable resulted in a message that the problem was too big. Evidently. the program could not process it or my computer did not have the memory to process it. Attempts to (h) polynomial regression analysis or "least squares fit" also were not successful. The highest correlation coefficient was .455 for the variable of 57 relational categories plotted against the relational distances (0. .01. .125. .25. .5). at the 7th order. In consulting with a professor of statistics. H. Rahbar. about the above problems. he suggested that the reason that the regression analysis was yielding poor results was that the x variable (0. .01. .125. .25. .5) was too non-continuous. It was also suggested that the variable could be made more continuous; however. this was not possible because the exact values could not be changed without losing their meaning. Result§_Pertaining_to Hypothesis fl 3 It is predicted that the expected frequencies of homicide as derived from a hypothetical analysis of interaction time spent between family members in fact 80 are dissimilar to the actual frequencies. In an analysis of the variable of low interaction time to high interaction time. it is found that this variable does show a higher number of homicides toward the group with very high interaction time as follows: Group A (very high interaction). 1.588 Group B (high). 115 Group C (medium). 46 Group D (low). 204 Group E (very low). 29 An exception is Group D. who committed more homicide than Groups B and C. Group D consists of in-law relation-ships. step-brothers. aunt/uncle. and niece/nephew relationships. A Chi Square Test which compared the actual frequency of homicide for the 57 cases to the expected frequency according to interaction time (not an equal distribution) resulted in a .007 probability. suggesting that this null hypothesis should be rejected. Again. a warning was noted next to the results. that fitted cells were sparse. Results_pertaining_to Hypothesis # 4 For the three cities. Chicago (1965-81). Detroit (1972). and Miami (1980). it is predicted that all three (different geographical urban areas. during' the three time periods display a pattern of family homicide where consanguineal relatives are less at risk for homicide than are individuals related by affinity. The frequency table for the variable of homicides 81 occurring between first offender and victim. in Chicago (1965-81) reveal similar patterns compared to Miami (1980) and Detroit (1972). See Table 4.2 below. TABLE 4.2 Comparison of family homicide for Chicago (1965-81). Detroit (1972) and Miami (1980). RELATIVE % FREQ. OF TOTAL FAMILY HOMICIDES Chipaqq Detrgit Miami r = N=1.982/12.875 N=127/690 N=59/572 .0 76.337% 75.590% 84.745% .01 .706% .000% 1.694% .125 2.169% .787% 3.389% .25 .857% .000% 5.084% .50 19.920% 22.834% 5.084% In all cases, the majority of homicides occurred among family members who were related by marriage. The city of Miami exhibited a homicide rate much lower for an r of .50 than did Chicago or Detroit. Miami's family homicide distribution deviated from the pattern where .5 relationships were second in frequency. as with Chicago and. Detroit» .Also see Figure 4.3 which jpresents a barchart of the percentages of geneticallly related vs. affinally-related homicides for the three cities. The descriptive statistics do appear to agree with the research hypothesis that the three cities display similar patterns of family homicide. The Friedman Test result comparing the three cities is not in accord with 82 XIHMNXIKNHNDEIOFIUWM. \ 15 -;"° .......-............ . V. -(\\\‘ cmcaoo ' DETROIT mam 1965-1981 1972 1980 I - ammonia: scrum “\1 beoly norm I r;ggag_1ra Comparison of consanguineal and affinal family homicides for Chicago (1965-81). Detroit (1972). and Miami (1980). the give 0159 MD; Non. dis hyp gen The rep is fre rel CO! su. Fr fr Se hy Ch 0C re 51 83 the descriptive results. A probability of .165 was given. therefore. the null cannot be rejected. Discussion of Results The descriptive statistics do support the research hypothesis for the dual hypotheses 1 and 2. Nonetheless. upon closer analysis of the data one may dispute the directional statement of the research hypothesis. that the frequency of homicide decreases as genetic distance between the offender/ victim increases. The fact that the closest relatives are well represented. second in frequency. indicates that there is not a clear linear pattern in which homicide frequency decreases with closer consanguineal relationship. and increases with lesser or no consanguineal relationship. The K - Smirnov tests and contingency table do suggest that the null hypothesis may be rejected. The Friedman Test upon the variable where all the frequencies under the degrees of relatedness were separated. also yielded support for the research hypothesis. What is very clear from the descriptive statistics. Chi Square. K — S tests. and Friedman test is that the occurrence of homicide is not a random event. Certain relationships exhibit higher numbers of homicide. and also certain genders are victimized more often. 84 It has been demonstrated that non-blood related individuals do commit homicide with greater frequency. Whether the pattern may be attributed to some underlying predisposition which mitigates against killing one's genetically closer relatives. cannot be determined at this juncture. The paradigm of natural selection and Durham's genetic mode of relationship between genes and culture may have something to do with the dynamics of family homicide. The questions and predictions generated have been shown to be valuable. However. the present study's research design is very limited and does not address this issue reliably. What is needed in order to address the issue are analyses of the pool of potential victims. of ages of the victims/offenders. of precipitating motives. and many other factors. The data indicate that homicide did occur in much greater numbers between certain individuals. Marital homicide predominates the distribution of family homicide. This is in agreement with other studies of family homicide. The other relationships with 'high numbers of fatal conflict are brothers. brothers-in-law. fathers/sons. and mothers/ sons. This tells us that these relationships experience conflict which very often culminated in homicide. It is the case of brothers-in- law that we would not expect such a high number because they usually do not have a great amount of interaction time. Nonetheless. it is not surprising. since most homicides if A compa fluid homici not have a 1965-1981. '. for Canadia than 70 tin ummlly a 1 Chicago san families we Sons 5 often than to this pa thim daugi Dauthers Other ham mOthers. Also, than were may be ti "iVES (iif “live the and argum. that hqui freQUency Droprietar to Common. 85 homicides in general involve males killing males. A comparison of the parent/child and step-parent/ child homicides suggests that step—children probably did not have a greater risk for homicide. in Chicago. from 1965-1981. This contradicts Daly and Wilson's finding for Canadian data. in 1984. Step—children were more than 70 times as likely to be killed by a step-parent. usually a father. in the Canadian sample. Unlike the Chicago sample. their study did have data of how many families were step-families. Sons seem to have been killed by parents much more often than daughters were by parents. What may have led to this pattern is unknown. Sons killed parents more than daughters killed parents. a difference of 28. Daughters killed 17 more mothers than fathers. On the other hand. sons killed only 9 more fathers than mothers. Also. wives were killed more frequently by husbands than were common-law wives. Speculation on this pattern may be that legally' married ‘husbands perceive their wives differently than do common-law husbands. One motive that is frequent in marital homicide is jealousy and arguments over extra-marital affairs. It could be that husbands act upon their jealous rage with greater frequency simply because they have an elevated raroprietary attitude toward their wives. in comparison to common-law married husbands. In the above cases. gender Brother much g1 M2 perpet: homici homici compaz family famil resea membe 0i t) reQa this Who tag. exc rel he: fa: Th 57 Cc 1: 86 gender is an important factor for predicting risk. Brothers and brothers—in-laws also seem to have had a much greater amount of fatal interaction. Male perpetrated homicide outnumbered female perpetrated homicide by 16%. This suggests that family homicide is clearly distinguished from other types of homicide. in which males commit almost all the homicide compared to females. Females committed 42% of the family homicides. a very high number. The suicide rate was very often associated with a family homicide. which is in agreement with previous research. It may be that individuals who kill a family member. will often kill themselves also. as an extension of that killing of someone close to them. The results of the analysis. hypothesis # 3. regarding the interaction time variable indicates that this variable does describe a pattern where individuals lflH) we would expect to spend more time at risk. do in fact. experience higher numbers of homicide. The exceptions to this pattern indicate that in-law relationships. step—brothers. aunt/uncle. and niece/ nephews -- for some reason have a greater amount of fatal conflict. although they spend less time together. The Null Hypothesis. that the actual frequencies for the 57 categories of relationship show 1“) difference in comparison to the expected frequency according to interaction time is rejected. Therefore. this indicates 87 that there are multiple factors involved in producing the actual distribution pattern. Interaction time may be a an important variable. but it does not by itself describe the actual pattern. The results concerning Hypothesis # 4. indicate that Chicago. Detroit. and Miami all had similar distributions of family homicide. The similarity of the patterns is a strong indication that this pattern is not random and that certain individuals in all three cities experienced varying risks for homicide. Although the Friedman Test failed to reject the null. the descriptive statistics display enough similarity so as to concur with the research statement. It may be that the Friedman test was not the appropriate test. SummarLoLResults Hypotheses # l and 2. that the frequency of homicide increases as the genetic distance between the offender and victim decreases is supported; however. some clarification is warranted since the pattern is not linear. The indicated pattern is that certain relationships and genders have a greater risk for homicide. It appears that non-blood related individuals. marital homicides. are prominent with almost 66% of the family homicides. Second in frequency are the closest relatives who share 50% of their genes. brothers with 4% of the family homicides. Brothers-in- law and sons killed by parents also appear to be at greater ri: Hypoti is differt 'time at 1 descriptiv this vari family he tautology. HYDO‘ Detroit. homicide, 88 greater risk for homicide. Hypothesis # 3. that the actual pattern of homicide is different from the expected. pattern according’ to "time at risk" or interaction time. is supported. The descriptive statistics also support the assertion that this variable helps to describe the (distribution. of family homicide. However. since this analysis is a tautology. the validity of these findings are suspect. Hypothesis # 4. that the three cities of Chicago. Detroit. and Miami exhibit similar patterns of family homicide. is also supported. SWMBLX Family been found of the Chi< Previi in few cas between t Dronosed Conflict Introductj Coevolut1< WhiCh max theory re b1010qy produCe I argued t‘ acCUIate] rather ( consiSte. natural : The 9530c1at relation CHAPTER 5 SUMMARY AND CONCLUSIQNS Summary Family homicide is a complex phenomenon. as has been found in the literature review and in the analyses of the Chicago homicide data. Previous studies of family homicide were sparse and in few cases looked at the specific relational distance between the victim and offender. Daly and Wilson proposed that natural selection operates to initigate conflict among related genetic relatives. In the Introduction to this thesis the emerging theory of Coevolution has been discussed as a possible theory which may bring insight to homicide research. This theory recognizes that there are intertwined dynamics of biology (natural selection) and culture operating to produce persistent behaviors. such as homicide. It is argued that this general theory of behavior may more accurately explain a behavior such as family homicide. rather than simply attributing patterns of behavior consistent with natural selection to the mechanism of natural selection. as Daly and Wilson have proposed. The literature review found that family homicide is associated with economic stressors. that the relationship between victim and offender is important. 89 as are age hmMcide r To my kno smmne ava. question c Problems prevented 1n ti patterns hypothese appears among th ho; 90 as are age and gender. The obstacles to doing family homicide research is that data must be hand-gathered. To my knowledge. this Chicago data set is the largest sample available which has the detail necessary for the question of who is killing whom in family homicide. Problems were encountered with the data set which prevented some intended analyses. In the analysis of Chicago family homicide. several patterns were described and interpreted. Three hypotheses were tested. The main result is that there appears to be a non—random distribution of homicide among the many offender/ victim relationships in family homicide. Conclusions Several conclusions may be derived from this thesis. as follows: 1) The incidence of homicide is a non-random event; affinally related individuals. especially spouses. are at a much greater risk to be involved in a homicide. 2) The degree or ”r” of shared genes between individuals appears to predict the frequency of homicide. With the exception of brother/brother homicides and parent/son homicides. in all cases of first offender to victim homicides which took place in Chicago. 1965-81. a pattern emerges of greater ”r" -- less homicide; zero relatedness -— more homicide. This pa‘ter noweve the me 3) Gende homici at qr dauqht risk t 4) In ma other famil homic often 58%.' 5) AnalY homiC a ver relat by ea 6) There occur have "time dista Discussic The is killi 91 pattern is consistent with natural selection theory; however. natural selection theory is not proven to be the mechanism operating to produce this pattern. 3) Gender is an important factor in the pattern of some homicide. The Chicago data set showed that sons are at greater risk to be killed by parents than daughters and that common-law wives are at lesser risk to be killed by husbands. 4) In marital homicide. husbands and wives killed each other in almost equal numbers. which suggests that family homicide is unique. as all other types of homicide are dominated by male victims and offenders. Males do predominate as perpetrators. 58%. compared to female-perpetrated homicides. 42%. 5) Analysis of the available samples of family homicide in Detroit. Miami. and Chicago. results in a very similar pattern of family homicide. Affinal relatives appear to have been killed much more often by each other than were consanguineal relatives. 6) There are many variables involved in the occurrence of a family homicide. Just three that have been demonstrated to be important include: "time at risk"; shared "r” (the genetic relational distance) between the offender/victim; and gender. Discussion The intent of this thesis has been to explore who is killing who within families. The results suggest 92 that spouses are killing the most. much more than any other family members. Evidently. marriage is a relationship where volatile conflict arises. Spouses also spend a great deal of time together. so their "time at risk” is greater than others. A question that may be explored regarding marital homicide is whether cross-cousin and cousin marriages experience less homicides. There are certain cultures and religions that permit first cousin marriages. It would be expected, from selectionist thinking. for blood-related spouses. to kill each other less often. The finding that the closer genetic distance between individuals predicts less homicide. and zero genetic distance predicts more homicide is consistent with what one would expect from natural selection theory. It would be expected that individuals. if they are going to kill other individuals. it is in their best interest. not in) kill other blood-related individuals. Natural selection theory recognizes that many behaviors may have been selected during our evolutionary past but may persist today. in totally novel and nmdern environments. Those behaviors which hold no benefits today would in all likelihood slowly or abruptly disappear. One's best interests. as developed over the millennia is to preserve oneself and one's genes which are possessed by close relatives. One's best interests include minimizing conflict with one's closest 93 relatives. The fact that this pattern of family homicide is consistent with selection theory. however. does not indicate that natural selection theory causes it to occur nor explains how or why it occurs. Figure 5.0 presents a preliminary model of dual inheritance theory. coevolutionary theory. which may aid our understanding of the complex etiology of a behavior. such as family homicide. The person possesses two information systems. one cultural and one genetic. These two intertwined systems select and produce behavior. The action of homicide is influenced by both cultural and genetic mechanisms. For example. a person grows up in a family where parents "teach” their children that violent conflict is an appropriate way to solve problems both within the home and outside. In this way the child who grows up has learned that violent interaction is a preferred way to interact. At the same time. there may be an underlying predisposition not to harm one's genetic relatives. or a cultural taboo against harming one's blood relatives may be a learned attitude by the child. Cognition of the benefits and risks is not an issue. In some cases. both may be present. Figure 5.1 presents a model which shows how behavior may occur due to both ultimate causation factors and proximate causation factors. For example. 94 PERSON iiiNfimOMfiflAN €W5TBM6: __- 5 Moves 0F mammal? BETwGE'lU Gauss («(01.7qu '- enunncws emmusmr (9601611; MEDiATloN ' OPPOSING eemnolsmpi‘? ® comm Mammy NEUTRAL (summer!) EIQURE 5.0 Model of one mode of relationship between genes and culture (genetic mediation) as seen in the homicide of a relative. 4t1‘ll“ lull“ 4‘ 'IIIIIII‘III‘ I‘l 95 PESE< in d from both oduce behavioral strategies :32! (En hammwa $\()\g¢§>OL ”Np—.332. ¢ < - .933... :96 +3.: . 4.38 H z 2 :25 3.09 83°23<-282@ i < . .3155. .6 Sflwbw embemwfi sfifivfiw u m we “fiaw hzssae. mum lessees; oSSQSwQ a 3.. mm. §O® .3533 «wfiyiyi Hr 452.12 a no flaw)”.— $ x§§$ ~ - l§§§.$§§3§€. b§f0§§§§§3§§x nae kfi 33.139 .‘é~ 3‘3 wuwfizy Mm>o Wh.