Australian Institute of Criminology

Skip to content

Knife crime: Recent data on carriage and use

Trends & Issues in Crime and Criminal Justice no. 417

Lorana Bartels
ISSN 0817-8542
Canberra: Australian Institute of Criminology, May 2011

Foreword | In this paper, an overview is presented of recent data on the carriage and use of knives. Analysis of the data indicated an increase in the use of knives as a proportion of all homicides, although the number of homicides remained relatively constant. The proportionate use of knives in robberies, by contrast, remained fairly constant, while the number of robberies decreased dramatically.

Responses to the Drug Use Monitoring in Australia program for 2005–09 indicated that only a small proportion of respondents reported owning or using a knife as a weapon in the previous 12 months. The most common justification given for having a knife was self-defence.

These findings have implications for legislation on the sale and carriage of knives, as well as for crime prevention measures. In particular, a better understanding of how and why knives are used in the commission of crime is vital when developing policy responses such as knife amnesties and education campaigns, and legislative measures such as stop and search powers and increased prison sentences.

The limited information available on the nature, extent, cause, motivation and possible growth of knife carriage highlights the need for improved data collection, along with the development of clearer evidence for what works to reduce knife carrying and knife offences.

Adam Tomison

Knife crime attracts a great deal of community concern, however, as the Australian Institute of Criminology (AIC) has previously noted, there is a general lack of academic research in Australia on the carriage and use of knives (AIC 2009). More recently, research published by the AIC (Bartels 2011) has considered the incidence and aetiology of knife crime, together with a detailed analysis of the legislative and policy responses to knife crime in each Australian jurisdiction. This paper draws on that research, with particular reference to data from the National Homicide Monitoring Program (NHMP) and the National Armed Robbery Monitoring Program (NARMP). Previously unpublished data on the use and carriage of knives from the AIC's Drug Use Monitoring in Australia (DUMA) program are also presented to further develop the understanding of, and consequent responses to, the issue of 'knife crime'.

In using the term 'knife crime', it is recognised that

'Knife crime' has become an expression commonly used by politicians and the media, but it is not always entirely clear what it actually is or what they actually mean when they use the term. 'Knife crime' potentially encompasses a very broad range of offences and thus causes problems in both the definition and determination of its prevalence...Whatever the meaning, the public and political debate about 'knife crime' would benefit from both an attempt to define what is actually meant by the term and a more careful, and less sensational, use of it (Eades et al. 2007: 9).


In a Victorian study, Living on Edge, Bondy, Ogilvie and Astbury (2005) investigated the perceptions, motivations and experiences of young people aged between 10 and 25 years regarding the acquisition, carriage and criminal use of weapons, particularly knives. The authors examined data from several sources, including qualitative data from 82 young people, although the relatively small sample size and the limitations of the self-report methodology should be acknowledged (Merner & Delacorn 2010).

Bondy, Ogilvie and Astbury (2005) found that the majority of young people in their study did not carry weapons, with only 15 percent of respondents (n=72) reportedly carrying a knife regularly. They also found that there appeared to be two distinct populations. These were a younger cohort who carry these weapons and ultimately grow out of the habit, and an older cohort who, primarily through association with drug-related crime, are detected in crime and hospitalisation data.

Bondy, Ogilvie and Astbury (2005) noted that there had not been any longitudinal work indicating any change in the characteristics or nature of knife carriage over time. The authors pointed to the following factors as relevant to knife carriage:

  • problem proneness (eg drug use and sales, aggressive attitudes and fighting, mental health problems, risky sexual behaviour, poor academic behaviour);
  • fear and vulnerability; and
  • social influence.

The latter two factors apparently were better predictors of weapon carriage than problem proneness. Overall, they argued that the decision to acquire, carry or use a weapon 'must ultimately be viewed within a broad social framework' and that failure to do so would result in 'poorly designed and ineffectual interventions' (Bondy, Ogilvie & Astbury 2005: 113). They therefore called for 'continued efforts to engage young people in pro-social activities and reduce the perceived and actual level of risk in the environment' (Bondy, Ogilvie & Astbury 2005: 114).

Since 1999, the AIC has managed the DUMA program, which involves the quarterly collection of information on drug use and crime from police detainees in selected police stations and watch houses, and is Australia's only nationwide survey of drug use and criminal offending among police detainees.

Mouzos and Borzycki (2006) found in their analysis of DUMA data that of the 138 detainees who reported having owned a knife, the typical profile was of a man (84%); aged 30 years and under (78%); who had been arrested in the previous 12 months (73%). Just over half of the respondents (54%) had completed Year 10 at school, with a small minority (14%) working full-time. Almost one-third (32%) had been in prison in the previous year; 51 percent had been charged with a violent offence or property offence in the previous 12 months, compared with 14 percent having been charged with a drug offence. Almost all of the respondents (90%) reported engaging in some form of illicit drug use over the previous 12 months, with most (76%) having used multiple drugs. The most common drug used was cannabis (80%), followed by methamphetamine (64%). Urinanalysis for 80 respondents indicated that 80 percent had used some form of illicit drug in the previous 30 days, with 35 percent revealing multiple drug use. Cannabis (70%) and methamphetamine (32%) were again the most commonly used drugs (Mouzos & Borzycki 2006).

In another key Australian study, Brown and Sutton (2007) surveyed 150 'street youth' and 184 'school youth' in Sydney, providing a key insight into knife carriage and use by young people. Again, the research was somewhat limited, given its small sample size and a poor response rate for the school sample, which may have caused a possible bias in the data. Nevertheless, it yielded the following key findings:

  • the street sample were more likely than the school sample to know someone who frequently carried knives or implements that could be used as weapons (49% vs 27%);
  • fifty-seven and 23 percent respectively admitted that they carried such implements at least occasionally;
  • fifty-eight percent of the street sample had carried a knife/offensive implement in the last week and a further 23 percent (total of 81%) in the last few weeks. The figures for the school sample were 36 and 21 percent respectively (57% in total); and
  • the most common age at which both groups reported beginning to carry a knife/offensive implement was 13–14 years (street sample 42%; school sample 69%); a smaller but substantial number admitted carrying them between five and 12 years of age (street sample 23%; school sample 32%).

The latter finding is of particular relevance given the recent finding in Scotland that the strongest influence on carrying a knife at 16 years of age was carrying a knife at 13 years of age. This suggests that early intervention targeted at those who carry weapons at young ages could have a significant preventative impact (McVie 2010).

Brown and Sutton (2007) found that similar proportions of both groups reported carrying the implements in public places, such as sporting events (both 44%) and concerts (53% vs 51%). However, the street sample was much more likely to carry implements at railway stations (91% vs 63%) or at home (64% vs 30%; Brown and Sutton did not explain how 'home' was defined for the street sample). In the school sample, 47 percent reported carrying the implements at school. Of the respondents who carried such implements, 15 percent of the street sample and 27 percent of the school sample had used them as a weapon at least once, while 59 percent and 32 percent respectively had used them on more than one occasion. The street sample was also much more likely to have been stopped and searched by police (77% vs 17%).

Brown and Sutton (2007) indicated that the key factors that appeared to be associated with knife carriage were the same factors that relate to juvenile delinquency more widely. These were—having a history of victimisation and exposure to violence and feeling fearful and/or engaging in risky behaviours (eg drug use/sale, fighting and joining gangs). Peer and family influences also contributed to weapon carrying and were compounded by socioeconomic disadvantage, illicit drug activity, community disintegration, the availability of weapons and a lack of educational and employment opportunities. Significantly, the young people's perceptions of safety, specifically in public spaces and at night, influenced knife carriage, with Brown and Sutton (2007: 57) observing that

[a] surprising number of the young people in our study have been the object of threats and assaults with knives/offensive implements. It is sobering to realise that many of these assaults are conducted not only by strangers, but also by family members and friends. Thus, in the context of a society that condones this type of behavior, it is hardly surprising that young people would use these methods to protect themselves as well as to bolster their own sense of power and self-esteem.

National data on the use of knives for selected offences


The most recent NHMP report (Virueda & Payne 2010) indicated that in 2007–08, more homicide victims died from stab wounds than from any other single cause of death. The number of homicides involving knives had remained relatively unchanged since 1989–90, although due to the decline in firearm homicides, knife-related homicides comprised a larger proportion of homicides recorded in the more recent data. In both 2006–07 and 2007–08, knives/sharp instruments were involved in 43 percent of homicides; by way of comparison, in 2000, knives and other sharp instruments accounted for only 30 percent of homicides.

The vast majority of victims who died as a result of a stab wound in 2007–08 (92%) were stabbed with a knife. Stab wounds were the most common cause of death in Victoria, where they accounted for 59 percent of homicides; they were least common in Tasmania (20%) and the Australian Capital Territory (0%), although the small number of homicides in these jurisdictions (n=5 and n=3 respectively) should be acknowledged. Acquaintance homicides were particularly likely to be as a result of stabbing (52%), compared with 43 percent for domestic homicides and 20 percent for stranger homicides (Virueda & Payne 2010).

Analysis of more recent NHMP data indicates that victims under 25 years of age are more likely to be killed with a knife than any other weapon. Indeed, 42 percent of male victims and 31 percent of female victims aged 18–24 years were killed with a knife, compared with 20 and 15 percent respectively who were killed with a firearm. In addition, where the offender was under 25 years of age, knives were used in 34 percent of homicides (unpublished data from the NHMP).

Data from the Australian Bureau of Statistics (ABS) for the offences of murder and attempted murder are presented in Table 1. As can be seen, between 2001 and 2009, knives were involved in approximately 23 to 36 percent of murders, indicating a general rise in their proportionate use. By way of comparison, 'no weapon' was used in 33 to 45 percent of cases, while firearms accounted for nine to 16 percent of murders. The use of knives in murder has ranged from 69 to 95 per year, with 2009 the second highest on record; this demonstrates a slight increase in use, especially given that the number of murders fell from 311 to 261 during this period (ABS 2010).

Table 1: Use of knives in murder and attempted murder
Murder Attempted murder
n % n %
2001 90 28.9 151 33.0
2002 72 22.7 142 35.5
2003 86 28.6 115 32.0
2004 69 26.1 100 32.1
2005 78 30.1 81 29.9
2006 95 33.8 86 35.7
2007 81 31.8 100 40.3
2008 87 33.3 74 31.4
2009 94 36.0 87 36.9

Source: ABS 2010

Knives were involved in approximately 30 to 40 percent of attempted murders between 2001 and 2009. This represents a general increase in their use as a proportion of all attempted murders, with 2009 the second highest year on record (36.9%). Unlike murder, where the use of 'no weapon' generally exceeded the use of knives, knives were more commonly used in cases of attempted murder than 'no weapon' or firearms (the next most commonly used weapon) for each year between 2001 and 2009. In terms of offence numbers, there was a decrease in the use of knives in attempted murder from 151 in 2001 to 87 in 2009 (see Table 1). Over the same period, there was an overall decline in attempted murder from 458 to 236.

A jurisdictional breakdown indicates that between 2004 and 2009, the use of knives in murder rose as a proportion of all such crimes in Victoria (from 37% to 43%), but fell in Queensland (from 35% to 26%). The ABS (2010) data point to an increase in the proportion of murders involving knives in New South Wales between 2004 and 2009 (from 28% to 36%), although analysis from the NSW Bureau of Crime Statistics and Research (BOCSAR) suggests that the number of murders in New South Wales involving a knife or other sharp instrument (sword, scissors or screwdriver) was stable between 2005 and 2009 (BOCSAR 2010).

In 2009, knives were most likely to be involved in murder in the Northern Territory (64%) and least likely to be used in murders in Western Australia (17%), although the small number of cases in both jurisdictions must be acknowledged. There were no data for Tasmania or the Australian Capital Territory.

Armed robbery

The most recent analysis of the AIC's NARMP data (Smith & Louis 2010) sets out a wealth of information on the use of knives, which were the most commonly used weapon, accounting for 47 percent of armed robberies (down from 53% in 2006). Smith and Louis (2010) also found that knives were used against at least half of all victims, regardless of age or gender, although there were some age and gender differences in patterns of weapon use. In particular, women aged 40 to 44 years were victims of robberies where knives were used more often than any other age category (65%).

There was a small increase in the use of knives for the organisation most commonly targeted for robbery, namely retail venues (37% up from 34% in 2006). Service stations were the next most common target, at 18 percent (down from 26% in 2006). Overall, 50 percent of armed robberies of organisations involved a knife, compared with 53 percent for robbery of individuals. In 2007, a single knife was used in 45 percent of incidents, compared with 51–53 percent in 2004–06. Knives were the most common weapon used in the majority of locations, accounting for 62 percent of armed robberies in post offices and newsagents, 59 percent in open spaces and 56 percent of incidents in corner stores, supermarkets and takeaways.

Only a minority of jurisdictions were able to supply information regarding victim injury as a result of armed robbery, with injury data available for approximately one in 10 victims. Accordingly, Smith and Louis (2010) cautioned that the results should not be interpreted as representative of all armed robbery victims. Subject to this caveat, knives were the weapon least likely to result in no injury, at 10 percent, compared with 16 percent for firearms and 18 percent for syringes. However, they were equally as likely as firearms to result in emotional trauma (61% and 62% respectively), compared with 45 percent for syringes. Only three percent of the knife victims sustained trauma requiring immediate emergency medical treatment.

Other key findings from the NARMP data were:

  • knives were generally less likely to be used when more offenders were involved in the commission of the offence. Knives were used in 46 and 48 percent respectively of offences involving one or two offenders, compared with 39, 36 and 37 percent respectively for offences involving three, four or five offenders;
  • armed robberies involving knives averaged a net of $860, compared with $1,726 for robberies with a firearm and $483 for those with a syringe; and
  • there was little variation in the patterns of weapon use as a function of the various age and gender groupings of co-offenders, although mixed male/female groups aged 35–49 years used knives more frequently than any other group combination.
Table 2: Use of knives in robbery
Yearn %
2001 5,982 22.5
2002 4,051 19.3
2003 3,748 19.0
2004 2,960 17.9
2005 3,142 19.0
2006 3,734 21.5
2007 3,615 20.1
2008 3,225 19.5
2009 2,914 19.1

Source: ABS 2010

As set out in Table 2, in the ABS (2010) data, knives were involved in approximately 18 to 23 percent of robberies. This figure includes all robberies, not just armed robberies as collected in the AIC's data, and is therefore lower than the ABS' figures, where 'no weapon used' accounted for 57–63 percent of robberies. In the ABS (2010) dataset, there was a rise in the use of knives in robbery in New South Wales (as a proportion of all robberies), from 16 percent in 2004 to 20 percent in 2009, following a recorded high of 22 percent in 2006 and 2007. The number of robberies involving a knife fell from 5,982 in 2001 to 2,914 in 2009 (51% decrease), the lowest number on record. Over the same period, the total number of robberies fell from 26,591 to 15,238 (42% decrease).

By way of comparison, according to the BOCSAR (2010) data, the number of robberies involving a knife or other sharp instrument fell 12 percent between 2005 and 2009; although both the ABS and BOCSAR figures recorded a decrease in the number of robberies involving knives, the difference in magnitude of the decrease is likely due to differences in counting rules and timeframes.

In Victoria, the use of knives in robbery (as a proportion of all robberies) fell fairly steadily from 26 percent to 18 percent between 2004 and 2009, although recent data cited by Victorian Police suggest a nine percent increase in armed robberies involving knives in the last 12 months ('Crime statistics and the trouble with knives' The Law Report 12 October 2010).

In Tasmania in 2009, a knife was involved in 29 percent of robberies, the proportionately highest Australian rate, although the actual number of robberies was small (n=143). Knives were least likely to be used in robberies in Western Australia (11%; n=1,589).

Other offences

The ABS (2010) publishes national data on the use of knives in sexual assault. Approximately one percent of cases between 2001 and 2009 involved knives; knives were also used in seven to 10 percent of kidnapping/abduction cases. With regard to assault, these data are not aggregated on a national basis. In 2009, the use of knives in assaults ranged from two percent in Western Australia and the Australian Capital Territory to six percent in the Northern Territory.

DUMA data on knife carriage and use

As mentioned previously, DUMA collects information from police detainees at a number of sites around Australia. In the past, DUMA has recorded data on the use of knives by police detainees. These data provide information on the relationship between offending behaviour and contributing variables such as drug use, mental health and the possession of weapons. Mouzos and Borzycki (2006) presented data from DUMA addenda on weapons administered on three occasions in 2001, 2002 and 2004.

The present paper draws on previously unpublished data from the weapons grid that was included in the core survey between 2005 and 2009. The questions, which overlapped to some extent with the questions in the 2001–04 addenda, collected information on detainees' possession and ownership of weapons, their use of weapons in crime, how often they usually carried their weapon and the licensing/registration status of weapons owned. Although issues about the validity and reliability of self-report data of this nature are acknowledged, it is argued that this information makes a valuable contribution to the field by providing important quantitative data on detainees' experiences, thoughts and attitudes on weapons carriage and use that could not otherwise be obtained.

Table 3: Ownership or possession of knife in previous 12 months
Number of knivesn %
0 13,588 87
1 1,008 6
2 297 2
3 178 1
4–5 150 1
6–10 156 1
11–20 199 1
Total 15,576 100

Note: Percentages may not total 100 due to rounding

Source: DUMA data [computer file]

As part of the DUMA questionnaire, respondents were asked to consider only objects 'that you own or use as a weapon'. As set out in Table 3, the overwhelming majority of respondents (87%; n=13,588) answered 'none' to the question In the past 12 months how many of the following weapons [knives] have you owned or possessed?. The next most common response was one knife (6%); a small proportion of respondents (2%) reported owning or possessing six or more knives (as weapons) over the preceding year. By way of comparison, Mouzos and Borzycki (2006) found that 36 percent of respondents (n=1,365) reported having owned a knife in the previous 12 months, suggesting a significant decrease in the latter period.

Responses to the question What is your main reason for owning or possessing a knife? for each of the respondents' first three listed weapons (n=1,973) are set out in Table 4. The most common justification given was protection/self defence, which accounted for 52, 46 and 39 percent of responses for the first, second and third knife respectively.

Table 4: Main reason for owning/possessing knife
First knife Second knife Third knife
n % n % n %
Hunting/target shooting 155 8 72 7 49 7
Job requirement 111 6 37 4 22 3
Protection/self defence 1,032 52 444 46 265 39
Use in criminal activity 86 4 47 5 32 5
Part of being in a gang 14 1 12 1 1 0
Collector 380 19 291 30 247 37
Other 195 10 67 7 50 7
Total 1,973 100 970 100 675 100

Note: Percentages may not total 100 due to rounding

Source: DUMA data [computer file]

This finding supports suggestions that crime prevention campaigns that enhance perceptions of safety may have a significant role to play in reducing knife carriage (see Bartels 2011 for discussion) and is broadly consistent with Mouzos and Borzycki's (2006) findings, where 47 percent of respondents reported owning a knife for self-defence. The second most commonly cited justification for owning or possessing a knife was that the respondent had a knife collection (19%, 30% and 37% respectively). Despite media representations about the prevalence of knife crime, use in criminal activity was only cited as the main reason for possessing a knife by four to five percent of respondents, while being in a gang was cited by zero to one percent of respondents. However, the low prevalence may be due to the reliance of the study on respondents' self-report of the reasons why they carried a knife.

Of the 1,988 people who said they had owned or possessed a knife in the previous 12 months, 1,967 responded to the question Have you used or threatened to use a knife to commit a crime in the previous 12 months? Of these, 1,538 (78%) responded no, compared with 429 (22%) who indicated they had used, or threatened to use, a knife.

Table 5 sets out the responses to the question Where did you get the knife? The main source was retail sale (55%, 60% and 59% for the first, second and third knife respectively), followed by family member or friend (25%, 21% and 19% respectively). Only a very small proportion of knives were obtained from a drug dealer (1%) or over the internet (1–2%). Four percent were obtained 'on the street' and five to six percent by private sale.

Table 5: Where knife obtained
First knife Second knife Third knife
n % n % n %
On the street 78 4 34 4 26 4
Drug dealer 17 1 12 1 8 1
Family member/friend 489 25 199 21 131 19
Private sale 108 5 55 6 40 6
Retail sale 1,081 55 582 60 397 59
Internet 22 1 16 2 16 2
Other 172 9 68 7 55 8
Total 1,967 100 966 100 673 100

Note: Percentages may not total 100 due to rounding

Source: DUMA data [computer file]

The present findings have clear implications for legislative and policy decisions surrounding the sale of knives and amnesties involving their possession. However, in developing such responses, it should be noted that it has been suggested that removing offensive weapons from circulation is ineffective, as it does not address the underlying causes of the problem (Bannister et al. 2010; Eades et al. 2007; Smart Justice 2010; UK HCHAC 2009).

Respondents were also asked How often do you usually carry a knife? as set out in Table 6. Notwithstanding the fact that responses were only obtained from people who said they had owned or possessed a knife in the previous 12 months, the most common response was that the person never carried a knife (38%, 47% and 50% respectively), or did so rarely (22%, 22% and 19%), however, 11–16 percent of respondents indicated they did so every day and a further seven to 11 percent that they did so most days (18–27% in total). These findings have particular relevance for the use of weapons stop and search powers, which are in force in different forms throughout Australia (see Bartels 2011 for discussion).

Table 6: Frequency of knife carriage
First knife Second knife Third knife
n % n % n %
Every day 318 16 109 11 79 12
Most days 210 11 77 8 50 7
Some days 266 13 118 12 82 12
Rarely 435 22 209 22 129 19
Never 749 38 457 47 335 50
Total 1,978 100 970 100 675 100

Note: Percentages may not total 100 due to rounding

Source: DUMA data [computer file]


This paper has presented the most current Australian data available on the use of knives for selected offences, as well as considering recent research on the carriage of knives. Commenting on the situation in the United Kingdom, Eades et al. (2007) noted that the area suffers from a lack of useful, specific, reliable, longitudinal research on the nature, extent, cause, motivation, frequency and possible growth of knife carriage. Further, in the absence of such research, it will be difficult to design and implement programs to reduce the incidence of knife carriage and therefore reduce knife violence more broadly (Eades et al. 2007).

In other research released concurrently by the AIC, Bartels (2011) examined current and proposed legislative and policy responses on this issue in Australia and the United Kingdom. The research evidence on responses such as knife amnesties, stop and search powers, increased prison sentences and community education and awareness-raising was also considered. As Eades et al. (2007: 27) have noted however, it is 'far from clear what actually works to reduce knife carrying and knife offences'. Although it is recognised that further research in this area is required to better delineate knife carriage and use, presenting the most current Australian data is an important first step.


The author gratefully acknowledges the assistance of Antonette Gaffney in extracting and interpreting the DUMA data for this paper.


All URLs correct at March 2011

Dr Lorana Bartels is the Criminology Research Council Research Fellow and a Senior Research Analyst at the Australian Institute of Criminology.