Home → Publications → Reports → Other reports → Violence: directions for Australia - Part 1: patterns and trends in Australian violence
Publications
Violence: directions for Australia
Part One - The state of violence in Australia
Patterns and trends in Australian violence
Homicide
Homicide is defined here as unlawfully and intentionally causing the death of another person. Because of the severity of the offence and the fact that it is of concern to a number of different agencies (police, coronial authorities and health) a great deal of information is available on homicide and its victims. It is some consolation to know that the recorded incidence of homicide in Australia today is relatively low by contemporary world standards, and in historical terms as well (Grabosky 1983; Mukherjee 1981; Mukherjee & Dagger forthcoming). An earlier publication of the National Committee on Violence reaffirms these findings (Chappell 1989, Violence Today No. 1).
There have been significant changes in homicide rates over the course of Australian history. Based on the best available evidence (nineteenth century statistics were by no means comprehensive) one may conclude that the rate of homicide in the colonial period was much higher than the homicide rates which prevail today. To be sure, this was especially the case during the convict era; homicide rates began a long-term decline which continued throughout the nineteenth and early twentieth centuries and lasted until the Second World War. With the end of the war, homicide rates increased, fluctuating around an upward trend, and continued to do so until the late 1960s. Since then, they have remained relatively stable, remaining at the level of those which prevailed at the turn of the century ( see Figure 1).
Figure 1: Number of homicides per 100,000 population 1915 to 1987
Figure 2*: Reported homicide rate per 100,000 population, 1972 to 1987
* The reported homicide rates from West Germany are not comparable
with those from other nations, as the German data is based on a broader definition of homicide.
Figure 3: Deaths by homicide - selected nations, rate per 100,000 population
Australia's record of homicide, in terms both of overall levels and of recent trends, compares favourably with those of other Western industrial societies, and is quite low compared to homicide rates in some of the world's poorer nations (see Figures 2 and 3).
Rates of homicide are not uniform across Australia. As Figure 4 illustrates, the homicide rate for the Northern Territory has been consistently more than five times greater than the comparable rates for the other States and Territories. This may be explained in part by the unique demographic composition of the Territory. These include the relative youth of the population, the high proportion of males, many of them transient, and the high Aboriginal population.
Because of the seriousness of the circumstances, information about homicide tends to be more comprehensive than does data relating to other types of violent crime in Australia. A variety of studies suggest that the risk of becoming the victim of homicide is very unevenly distributed across the Australian population.
Najman's study of the 517 death certificates relating to homicides recorded throughout Australia in 1965-67 revealed that the risk of homicide victimisation varied inversely with occupational status. Those from the most prestigious occupations had the lowest homicide death rates. The risk of becoming the victim of homicide is considerably greater within the lowest occupational categories (Najman 1980, pp. 275-6).
Other studies have shown that men are at a much greater risk of suffering a violent death, from whatever cause, than are women. Men are half again as likely to become victims of homicide (Bonney 1987, p. 7). One should note, however, that women are much more likely to be the victims rather than the perpetrators of violence. Data from three States reveal that slightly more than one-third of homicide victims are female; it has been estimated that females commit fewer than 15 per cent of recorded homicides (South Australia 1981, p. 51; Bonney 1987, p. 9; Kapardis & Cole 1988; Polk & Ranson 1989). Table 2 illustrates gender-based differences in rates of death by homicide and other selected causes. Table 3 illustrates age specific death rates by cause.
No striking variations appear to characterise age-based homicide mortality rates. Infants are at slightly greater risk, no doubt due to their fragility, their dependency, and the stress and frustration which may accompany their upbringing. Preadolescents face the lowest risk. Age-based rates increase until middle age, and then gradually decline (see Table 4).
In general, police statistics relating to homicide are regarded as more accurate than those which pertain to non-fatal incidents. The most significant analysis of homicide in Australia to date was conducted by Wallace (1986) with the support of the Criminology Research Council. This study analysed all homicides coming to the attention of the New South Wales Police between 1968 and 1981. The study was updated by Bonney (1987) to include homicides coming to police attention through 1986.
Whilst nearly one-third of all Australian homicides occurring during the period took place in New South Wales and were thus embraced by the Wallace and Bonney studies, they are not necessarily representative of homicides Australia-wide. A different homicide profile would appear to obtain in northern and western Australia, where Australians of Aboriginal descent comprise a greater proportion of the general population.
Among the most important findings from the Wallace-Bonney studies was the fact that most homicides were committed by family members (43 per cent) and friends or acquaintances (20 per cent) of the victim. Only a minority (18 per cent) involved attacks by strangers. Approximately one homicide in every nine was followed by the suicide or the attempted suicide of the alleged offender. Homicide rarely resulted from random attacks by deranged, pathological individuals. Indeed, the incidence of mental illness in homicide offenders was no greater than in the general population.
Females, who comprised slightly more than one-third (36 per cent) of victims, made up only 15 per cent of suspects. Approximately 56 per cent of all homicides were male-on-male, and 30 per cent male-on-female. Eleven per cent involved female assailants and male victims; in only 3 per cent were both victim and offender female.
Homicide rates were not significantly greater in urban areas than in rural areas, although the proportion of homicides involving family members was higher in the country, and the rate of stranger homicides was higher in urban areas.
Aboriginal victims and suspects were overrepresented compared to census enumeration of their proportionate numbers in the general population.
Approximately ten per cent of homicide victims were children under ten years of age. The overwhelming majority (96 per cent) of child victims were killed by parents or by other relatives. Infants (up to one-year-old) were at greatest risk.
Of those homicides which occurred outside the family, more than three out of four resulted from interpersonal disputes around hotels and clubs, or between neighbours, or between lovers and rivals. Only a small minority occurred in the course of other crimes such as sexual assault, robbery, or the attempt to resist arrest.
A survey of homicides coming to police attention in South Australia over a three-year period revealed patterns generally similar to those evident in New South Wales. The proportion of stranger homicides in South Australia, 9 per cent of all recorded incidents, was half the proportion noted in New South Wales. The South Australian study reported 25 per cent of victims killed by spouses, 15 per cent by other kin, 13 per cent by lovers or close friends, and 28 per cent by acquaintances (South Australia 1981).
A survey of homicide incidents coming to police attention in Victoria during the period January 1984-june 1988 was undertaken by Kapardis and Cole (1988). Among other findings, they report that 15 per cent of all homicides were followed by the suicide of the offender. This percentage was slightly higher than the 11 per cent noted for New South Wales.
Polk and Ranson (1989) studied the files of the Victorian State Coroner relating to 117 homicides occurring in that State during the years 1985-86. Over half of these, a total of sixty cases, were what could be described as "intimate" homicides, that is between lovers, relatives, or former friends. A further 25 cases (approximately 20 per cent of the total) were described as "confrontational" homicides, usually arising from a dispute between males in or around drinking places. An additional 21 cases were described as "homicides occurring in the course of another crime" such as armed robbery. In five of these cases, the homicide victim was the alleged perpetrator of the original crime in question.
The authors identified a dominant theme of masculinity in many of the homicides they analysed:
The vast majority of homicides which take place within the framework of relationships which are of a sexually intimate character appear to result of possessiveness, often jealous possessiveness on the part of males. The confrontational homicides are exclusively male, and imply the existence of a set of norms regarding "honour" or "face" which are acted upon in such a way as to lead down an interactional path to homicide (Polk & Ranson 1989, p. 54).
Another theme which the authors identified was the disproportionate vulnerability of what they describe as "marginal" members of society, that is the undereducated and unemployed, whose lifestyles place them at greater risk of becoming victims, and offenders.
When large proportions of persons are relegated to positions of permanent marginality, as a significant number of young people in contemporary society are ..., some of these are likely to drift into the highly risky activity that carries them into homicide. Often this is not because they wish the homicide to happen, but they drift into the high-risk behaviour that flows on outward into lethal violence (Polk & Ranson 1989, p. 54).
Serious assault
Serious assault is defined here as the unlawful and intentional infliction of bodily injury, and includes such offences as grievous bodily harm, malicious wounding, assault occasioning actual bodily harm and aggravated assault.
Statistics relating to non-fatal crimes of violence reflect quite a different picture from the stable rates of homicide in recent years, however. Steep increases in rates of reported serious assault, lesser assault, sexual assault, and robbery have occurred throughout the 1980s. Because of the aforementioned lack of equivalence in definition and counting practices, the Committee saw fit not to attempt any interstate comparisons.
As Figure 5 shows, the most dramatic increase in the reported incidence of serious violent crime in Australia in recent years has occurred in the category of reported serious assault. Whilst the Committee is convinced that increases in reported rates of non-fatal assault do in fact reflect a real and substantial increase in violence, the magnitude of this increase is by no means certain. It seems quite likely that in addition to an actual increase in serious assault, these statistics reflect a number of other factors, including changing social attitudes as to what is acceptable social behaviour, an increase in the inclination of victims to report an assault, a broadening of the definition by police of what constitutes "serious" assault, and more rigorous recordkeeping by Australian police authorities.
Robb (1988) and his colleagues analysed serious assaults recorded by police in New South Wales. Both the incidence and the rate of reported assault in that jurisdiction trebled in the decade after 1978. The authors concluded that there has been a real increase in the rate of non-fatal serious assaults since the late 1970s, to a degree which cannot be explained away by changes in the inclination of victims to report an incident to the police.
The New South Wales survey was based on police records of aggravated assaults, that is, reported incidents resulting in a record of some bodily harm to the victim. The periods covered in the study were the years 1982 and 1986-87. A random sample of 1060 incident reports provided the basis for the analysis.
Figure 4: reported homicide rate per 100,000 population 1981-82 to 1987-88
Source: Mukherjee, S. K. (Forthcoming) 2nd ed. The size of the crime problem in Australia, Australian Institute of Criminology.
Figure 5: Australia - reported violent crime rate per 100,000 population 1973-74 to 1987-88
Source: Mukherjee, S. K. (Forthcoming) 2nd ed. The size of the crime problem in Australia, Australian Institute of Criminology.
Figure 6: Reported robbery rate per 100,000 population 1972 to l987
Source: Mukherjee, S. K. (Forthcoming) 2nd ed. The size of the crime problem in Australia, Australian Institute of Criminology.
A number of consistent patterns were apparent across the two periods under review:
- Reported incidents occurred most commonly on weekends, between 10 p.m. and 2 a.m. on Fridays and Saturdays.
- The location of reported incidents varied little between the two years selected for analysis: 27 per cent of incidents occurred in the victim's home, and a further 27 per cent took place in the street. Just under 20 per cent occurred in the vicinity of licensed premises, and 8 per cent occurred in open places, parks or bushland.
- Nearly half (48 per cent) of the incidents arose from various altercations, and a further 20 per cent entailed domestic assaults.
- Slightly over one-third of all reported attacks involved the use of weapons, and weapon use in serious assault declined during the period under review. Most of the weapons in question consisted of sticks, clubs, or other instruments opportunistically at hand. The use of knives or firearms was rare, and decreased significantly between 1982 and 1987. In the latter year, only 9 per cent of the incidents involved the use of a knife, and 3.5 per cent the use of a gun.
- There was some evidence of an increase in the number and proportion of incidents arising from gang disputes; in 1986-87, 11 per cent of the reported incidents were of this nature.
- Most (67 per cent) of the assaults in question involved male victims and offenders. Overall, at least 90 per cent of suspects and 75 per cent of victims were males.
- The most common age group of both victims and suspects was 20-29 years.
A more recent study of serious assault, based on incidents reported to the Victoria Police, was published in 1989 (Ministry for Police and Emergency Services 1989). Because of differences in definition and recording practices, the Victorian findings are not necessarily comparable to those of New South Wales. Indeed, the Victorian report itself advises that it includes cases which would not have been defined as serious assault had they occurred in New South Wales.
This study was limited to reported assaults on civilians in which some physical injury was recorded; it excluded assaults on police officers. The study was based on a total of 1600 incidents occurring during five selected months, March and September 1987 and 1988, and March, 1989. The periods were selected to encompass the two years (1987-88 and 1988-89) which saw increases of 19 per cent and 21 per cent respectively in the incidence of serious assault.
As with the New South Wales data, a number of consistent patterns emerged:
- One-third of Victorian cases occurred in residential settings; just under 20 per cent took place in or around licensed premises; 27 per cent took place on the street or other public places; and 5.3 per cent occurred on public transport or at such facilities as stations or bus stops.
- Just over 30 per cent of Victorian assaults involved the use of a weapon. Only 7 per cent involved knives, and 2 per cent firearms. The proportion of assaults involving guns and knives declined over the periods sampled.
- Weekend evenings were the most common time of occurrence. Almost half of all incidents occurred on Saturdays and Sundays, with the highest proportion of incidents taking place between 10 and 11 p.m.
- Only 15 per cent of the cases involved violence between known intimates (family/partner). Over half involved persons not known to the victim.
- 75 per cent of the victims, and at least 83 per cent of suspects, were male. The proportion of male suspects was probably greater, as in a number of cases the suspect's gender was not recorded.
- Most suspects were aged in their twenties, with the mean age for suspects being 27 years.
The authors' categorisation of incident type and year of occurrence is as shown in Table 5.
A significant increase in the proportion of incidents occurring in or around licensed premises was observed from 1987 to 1989. A decrease was observed in the proportion of incidents relating to public transport.
The authors of the report commented that the increase in the proportion of incidents occurring in or around licensed premises appears attributable to "incidents that could be described as lying at the lesser end of the serious assault continuum" (that is, involving less severe injuries). The largest increase was not apparent until one year after the proclamation of the Liquor Control Act, which relaxed restrictions on the sale and consumption of alcoholic beverages. Overall, about 21 per cent of the incidents occurring in and around licensed premises involved bouncers, mainly as alleged perpetrators of the assault.
The authors further commented that the increased prevalence of reported domestic violence incidents may reflect changes in legislation to deal with family violence, and increased police familiarity with powers to intervene in cases of criminal assault in the home.
Informative though they are, analyses such as the above present only an incomplete perspective on serious assault. They shed no light on assaults (serious or minor) which do not reach the attention of the police. They appear vulnerable to changes over time and to differences between Jurisdictions in the definition of what, precisely, constitutes serious assault.
Domestic violence in particular tends not to be accorded full coverage. Traditionally, both victims and police were disinclined to invoke the criminal process in cases of criminal assault in the home. Only the more extreme incidents of violence between intimates were likely to be reported to police and defined as crimes. Findings such as those emerging from the New South Wales and Victoria surveys should thus be interpreted with caution. They are likely to be more indicative of violence in public places than in domestic settings.
Robbery
As figure 6 indicates, Australia's experience with the rate of reported robbery compares favourably with that of other Western nations.
The offence of robbery entails the theft of money or property from a person, accompanied by the threat or use of force. As Figure 6 indicates, Australia's experience As Figure 6 indicates, Australia's experience with the rate of reported robbery with the rate of reported robbery compares favourably with that of other Western nations. A number of jurisdictions have undertaken analyses of reported robberies. The most detailed of these studies was published by the New South Wales Bureau of Crime Statistics and Research (1987). This study, based on a sample of robbery incidents coming to the attention of the New South Wales Police during 1983, concluded that the incidence and rate of robbery increased in all jurisdictions, except Tasmania, during the 1980s, and that the increase in New South Wales was particularly pronounced.
The study drew a broad distinction between two types of robbery: commercial and personal.
Robberies of commercial enterprises:
- involve greater amounts of money and/or more valuable property;
- require some degree of planning, however rudimentary, on the part of the offenders;
- require offenders to negotiate particular physical and spatial conditions;
- entail specialist attention to security measures aimed at frustrating or deterring robberies. (New South Wales Bureau of Crime Statistics and Research 1987, p. 5.)
By contrast, the personal robberies in the sample tended to:
- involve much smaller amounts of money:
- result in more injuries to victims, and in injuries of a more serious nature;
- involve less use of firearms;
- generally bear the hallmarks of less planned and more impulsive offending. (New South Wales Bureau of Crime Statistics and Research 1987, p. 5.)
The most common targets for commercial robbery identified in the New South Wales study were banks and related financial institutions, and petrol stations. Banks provide lucrative targets, while petrol stations tend to be accessible and open late at night. Firearms were used in 47 per cent of the commercial robberies under review.
Personal robberies (which include bag snatching), on the other hand, tend to occur on the street or in parks, most commonly at night. Only 5 per cent of personal robberies were committed with a firearm. Approximately two-thirds of victims were male, with persons in the 25-44 age group at greatest risk. The incidence of personal robbery was disproportionately concentrated amongst persons of lower occupational status, although to a lesser extent than was reflected in the 1983 victims survey undertaken by the Australian Bureau of Statistics. It was suggested that this inconsistency may reflect a greater propensity on the part of persons with relatively high incomes to report incidents to police.
Those robbery offenders who were identified were overwhelmingly male. Most were unemployed, with limited formal education, and had substantial prior criminal records involving other crimes against property. It was estimated from police records that approximately half of the incidents in question were drug related. Far fewer perpetrators of personal robberies (9 per cent) than of commercial robberies (68 per cent) were found to be illicit drug users. On the other hand 40 per cent of alleged perpetrators of personal robberies were found to be alcohol affected, compared to only 3 per cent of commercial robbers. These findings are consistent with those of the study of convicted armed robbers conducted by the Australian Institute of Criminology on behalf of the Australian Bankers Association, which are discussed in more detail in Part Two of this Report.
Sexual assault
The definition of rape differs from jurisdiction to jurisdiction in Australia, and has been amended significantly in a number of jurisdictions during the past two decades. In general, it has come to include various forms of sexual activity without the consent of one party.
In New South Wales, for example, the Crimes (Sexual Assault) Amendment Act 1981, abolished the common law offences of rape and attempted rape, and replaced them with three categories of sexual assault of differing seriousness, depending on the degree of bodily harm inflicted. The amended Act expanded the definition of sexual intercourse to include the penetration of the vagina or anus of any person by any part of another person's body. Foreign objects inserted into the anus or vagina, except for proper medical purposes, are also defined as sexual intercourse under the Act, as would the introduction of any part of the penis into another person's mouth.
Despite the seriousness of the offence, few systematic studies of sexual assault have been conducted in Australia. This may in part be explained by the notorious difficulty of measuring the incidence and prevalence of sexual assault, whether by means of victim surveys or through records of cases coming to the attention of the police. Because of the nature of the offence, many victims are reluctant to disclose the fact that they have been assaulted. The "dark figure" of sexual assault is thus substantial and our knowledge of both its victims and its perpetrators is fragmentary.
One study of sexual assault which provides some evidence on these matters was conducted by the New South Wales Bureau of Crime Statistics and Research as part of its task in monitoring reforms to the law of sexual assault which were introduced in that State in 1981 (Bonney 1985).
The New South Wales study analysed cases of reported sexual assault giving rise to committal hearings during two periods: the first, 14 January 1979-13 July 1980, and the second, 14 July 1981-13 January 1983. The study excludes incidents not brought to the attention of police or incidents which were reported to police but which, for whatever reason, did not lead to committal proceedings against a suspect.
The following points emerged from the data:
- Only 25 per cent of the incidents under study involved an alleged assault by a person totally unknown to the complainant at the time. In nearly half the cases, the relationship between the complainant and the defendant was classified as "family", friends/acquaintances, or neighbours.
- Only 20 per cent of the incidents in question involved the use of a weapon. The most common weapon employed was a knife; firearms were used in only 2 per cent of the incidents.
- Serious physical injury tended not to characterise the incidents under analysis. About half of the complainants suffered negligible injury or none at all. One-third recorded only bruises, scratches or abrasions requiring medical attention. Only 2 per cent required hospitalisation for the injuries sustained. No data on psychological injury was recorded for. purposes of the study. As the Committee's publication Victims of Violence (Grabosky 1989) noted, the psychological consequences for the victim of sexual assault can be severe.
- There was a general similarity in the social background of complainants and defendants in the cases. "Cases which come before courts generally, but not exclusively, involve both an accuser and an accused who are poor, semi or unskilled blue collar workers, of whom many are unemployed" (Bonney 1985, p. 26).
- The typical complainant was a female aged between 18 and 30; after the 1981 amendments which redefined sexual assault to include offences committed against men, 7.5 per cent of complainants were male. The study did not seek to determine whether this difference reflected the actual victimisation of males relative to females, or whether there existed a gender-based difference in willingness to report sexual assault.
- The typical defendant was a male between the ages of 18 and 30, an unskilled or semi-skilled labourer. Nearly half were unemployed at the time of the incident in question. It was determined that in at least 70 per cent of the incidents, the defendant had consumed alcohol in the six hours preceding the alleged offence.
The South Australian Office of Crime Statistics (1983) published an analysis of 450 alleged sexual assault offenders apprehended by the South Australian police during 1980 and 1981. The charges in question arose from a wide range of alleged offences, from relatively minor acts of indecency to brutal rapes. Because the South Australian study was limited to offences coming to police attention and resulting in the apprehension of a suspect, the summarised findings cannot be generalised to all sexual assaults. In addition, because the study embraced a wider range of offences than those included in the New South Wales study, the two are not comparable.
The following points emerged from these data:
- The overwhelming majority of the South Australian offenders were male, and tended to come from relatively disadvantaged backgrounds: 40 per cent were identified as unemployed or pensioners and 24 per cent as labourers; only 5 per cent were identified as professional.
- 42 per cent of the alleged assailants were unknown to the complainant; 27 per cent were described as acquaintances and 31 per cent were described as relatives, close friends or neighbours.
- Only 7 per cent of the alleged offences involved the use of a weapon.
The Committee heard from a number of organisations providing support and counselling for victims of sexual assault. Jo Spangaro from the Sexual Assault Education Unit in New South Wales, told the Committee's Sydney community forum that sexual assault has "one of the lowest reporting rates of any crime ... [and] ... one of the lowest conviction rates among crimes tried". The submission from the Sexual Assault Referral Centre at Queen Elizabeth Hospital in Adelaide stated that the crime of sexual assault continues unabated and "it is our clear impression that most violent sexual assaults occasioning severe physical injury occur in the context of marital or de facto relationships in which violence of all sorts has been a common feature. Such sexual assaults are least likely to be proceeded with through the courts by the victim". This finding confirms those of Salimann & Chappell (1982) in their study of marital rape.
Originally published:
Violence: directions for Australia / National Committee on Violence.
ISBN 0 642 14975 5
Canberra: Australian institute of Criminology, 1990; pp 17-30