Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.
Prevalence and severity of violence against women
Estimating the prevalence of violence against women—the number of women in the population who are affected by violence—is challenging due to the very private nature of these experiences. A subset of police services across Canada is able to track the gender of victims and offenders for crimes that are reported to them, but this undercounts incidents of violence against women as just over one-third of spousal assaults and less than 10% of sexual assaults are reported to police. A survey of shelters for abused women and their children has been in operation since 1992. However, women who use shelters tend to be fleeing very serious violence and may not be representative of all abused women.
Victimization surveys, as opposed to police collected data, have become the standard for estimating the nature and extent of violence against women in the general population. In 1993, Statistics Canada conducted the first dedicated survey on violence against women. This survey was important because of the breadth and depth of the questions asked and because it established a baseline for understanding and monitoring physical and sexual violence against women in Canadian society.
The Violence Against Women Survey (VAWS) was a special one-time survey funded by the (then) federal department of Health and Welfare. In order to track changes over time, Statistics Canada’s General Social Survey (GSS) on Victimization was adapted to include a module on spousal violence which was modeled on the VAWS. The GSS on Victimization is conducted every five years and reference is made in the survey to experiences of spousal violence in the preceding 12-month and 5-year periods. Sexual assault is covered as one of the eight crime types addressed routinely on the survey. Experiences of sexual assault are measured over the previous 12 months only.
Spousal assault severity and prevalence can be compared over three time periods: the five years prior to the 1993 VAWS and the five years prior to the 1999 and 2004 GSS. Although there were attempts to make these surveys as similar as possible, comparisons should be made cautiously (see the Methodology section). The VAWS contained a single focus on acts of male violence against women, while the GSS is a general crime victim survey with a special module of questions designed to measure the prevalence and consequences of spousal assault against both women and men.
Spousal assault was first identified as an important social issue in the early 1970s due to the efforts of the women’s movement. Since that time, awareness of the issue and the development of tools to estimate the prevalence of spousal violence have grown.
In this document, the terms "spouse" and "spousal" refer to both marital and common-law unions, unless otherwise specified. The data from the 2004 General Social Survey (GSS) on Victimization related to spousal violence include a small number of same-sex spousal relationships; however, the numbers were too small to make statistically reliable estimates separately for women and men.
Note on statistical significance: Telephone surveys such as the Violence Against Women Survey (VAWS) and the GSS randomly select a sample of the population to be interviewed (age 18 and over for the VAWS; 15 and over for the GSS). The responses of each person interviewed are weighted to produce estimates for the overall population. An estimate of the total population, expressed as a percentage, is expected to be within about 1% of the true percentage 19 times out of 20. Estimates of smaller subpopulations (such as smaller provinces) will fall within a wider range. As a result, estimates from two points in time, or between two subgroups in the population, may have a wide and overlapping range and therefore will not show a statistically significant difference. Estimates based on small samples are not reliable and are suppressed.
See the Methodology section for detailed description of these surveys.
Prevalence of spousal assault
According to victimization data, there has been a decline in spousal assaults since 1993. Seven percent of women who were living in a common-law or marital relationship reported to the 2004 GSS that they had been physically or sexually assaulted by a spousal partner at least once during the previous five years. This is a small but statistically significant drop from 8% in 1999. These figures represent approximately 653,000 women in 2004 and 690,000 in 1999. In 1993, 12% of women had been assaulted by a spousal partner in the preceding five years. The figures for men were 7% in 1999 and 6% in 2004 (Figure 1).
In the 12 months preceding the survey interview in 2004, 2% of women had been physically or sexually assaulted by a spousal partner, representing approximately 196,000 women. The figures for 1993 and 1999 were 201,000 and 220,000 women, respectively (Figure 2).
The most significant change in spousal violence rates between 1999 and 2004 were within relationships that had ended at the time of the interview. Women reported higher rates of violence by previous spouses compared to current spouses, even though the percentage of women who experienced violence in the previous five years by ex-partners declined from 28% in 1999 to 21% in 2004.
Although it is difficult to determine with certainty the reasons for a decline in the prevalence of spousal assault, some factors could have played a role. These include:
Figure 3 shows the five-year prevalence rates of spousal assault for each of the provinces as measured by the 1999 and 2004 GSS. Rates of spousal violence against women have remained relatively unchanged in all provinces. The largest change was recorded in Prince Edward Island where rates have dropped by half. For Canada, Newfoundland and Labrador, Prince Edward Island and Quebec, the change in prevalence rates is statistically significant. This means that the difference between 1999 and 2004 rates are likely real and not the result of sampling. Newfoundland and Labrador was the only jurisdiction to show a rise in spousal violence over this five-year period.
The estimated numbers and rates of spousal assault against women and men by province in 2004 are set out in Table 1.
Table 1 Estimated number and rate of spousal violence incidents against women and men 15 years of age and over, past five years, 2004
Severity of spousal assault
In general, women are more frequently subjected to severe forms of violence from men than men are from women. For example, in 2004, twice as many women than men were beaten by their partners, and four times as many were choked (Figure 4). Moreover, 16% of women who were victimized by a spouse were sexually assaulted, and twice as many female as male victims of spousal assault reported chronic, ongoing assaults (10 or more) (see Figure 20). This finding suggests that despite similar prevalence rates reported by women and men in the 2004 GSS, assaults on women are more serious. One shortcoming of the data is that they do not indicate the degree of force used in each of these acts. However, the impact of spousal assault, in terms of injury and other consequences, is more severe for women (see Figure 20).
Overall, women were two-and- a-half times as likely as men to report the most serious forms of violence, such as being beaten, choked, threatened with a gun or knife, and sexually assaulted (Figure 5). The estimated number of women and men who experienced these types of assaults over the five-year period was 254,000 and 89,000, respectively.
A comparison of data from these three time points presents some evidence that the severity of the assaults has diminished slightly. These surveys indicate a reduction across these three time points in the percentage of female victims of spousal violence subjected to the most severe types of assault (being beaten, choked, threatened with a gun or knife, or sexually assaulted), from 50% of all victims in 1993 to 39% in 2004 (Figure 6). There were also declines in the percentage of victims experiencing chronic, ongoing assaults (10 or more) and in the percentage who feared for their lives from a violent spouse. However, the percentage of victims who suffered physical injury increased slightly. These changes are small but statistically significant.1
These shifts may be a result of improved societal interventions that help to reduce the escalation of violence in spousal relationships and are consistent with the decline in spousal homicides. However, the apparent declines in the prevalence and severity of spousal assaults have not resulted in a decrease in the use of shelters for abused women (see Figure 43 and Table 8). The demand for shelters continues to exceed availability as reflected in the fact that some 200 women are turned away from shelters on an average day (see the section Victims’ use of services).
Police-reported data on spousal assault
Police data on spousal assault incidents come from the Incident-based Uniform Crime Reporting Survey (UCR2) and do not contain data from all provinces. Other than in Ontario and Quebec, the data are primarily from urban police services. Although all police agencies across the country provide Statistics Canada with an annual statistical accounting of all known crimes, not all can provide details of assaults involving spouses or other intimate partners. Data for the period 1998 to 2004 are based on 68 police forces that have consistently participated in the UCR2 Survey since 1998. The data presented here are not nationally representative but provide an indication of trends for these 68 police forces, which represent 37% of the national volume of crime. By 2007, over 90% of criminal incidents will be reported in this detailed format.
These data indicate that women represented 87% of victims of partner assault from 1998 to 2004 in the jurisdictions of the participating police agencies. In 2004, 14,597 cases of partner violence involving female victims and 2,413 cases involving male victims were reported to these 68 police departments. This is a consistent finding over time and suggests that incidents involving female victims are more likely to approach the level of severity requiring police intervention.
These data undercount the actual number of spousal assaults as only 36% of female victims and 17% of male victims in 2004 had reported spousal violence to the police, according to the 2004 GSS (see Figure 21).
As shown in Figure 7, current and former husbands make up the largest number of intimate partner assault offenders recorded by the police and the number in this group has declined since 2001. The number of current and former boyfriends reported to police for intimate partner violence has increased since 1998 to become the second highest category, surpassing the number of assaults perpetrated by wives. A limitation of this analysis is that rates per population cannot be calculated because population data for marital status cannot be matched to these geographic regions. These trends therefore do not take account of possible increases in the population.
Figure 7 Number of intimate partner assaults reported to police by offender relationship to victim, 1998 to 2004
Pro-charging policies have led to the unexpected result that, in some cases, both victim and perpetrator are charged. Some jurisdictions in the United States have adopted “primary aggressor” models, which require police to identify the primary aggressor based on the history of violence between the parties and evidence that one person may have been acting in self-defence (Federal/Provincial/Territorial Working Group 2003).
One in five homicides in Canada involves the killing of an intimate partner. Rates of spousal homicide against both female and male victims have fluctuated over the past 30 years but show a general overall decline (Figure 8). The rate for women decreased by 39% between 1991 and 2004, from 1.16 to 0.71 per 100,000 couples. The rate for men decreased 59%, from 0.34 to 0.14 per 100,000 couples during this same time period.
Police statistics suggest that a substantial percentage of persons accused of spousal homicide were acting in self-defence. In 41% of spousal killings of men in which police had the requisite information, the police determined that the male victim was “the first to use or threaten to use physical force or violence in the incident.” In contrast, police indicated that the victim initiated the violence in only 5% of spousal killings of women (Johnson and Hotton 2003).
Statistics Canada’s Homicide Survey reports only the original charges—there is no follow-up to determine if the charge was reduced or the accused was found guilty. With respect to initial charges laid, women accused of homicide against intimate partners were more likely than men to be charged with second degree murder and manslaughter while men were more likely to be charged with first degree murder. The percentage of men who were charged with first degree murder in spousal killings has risen over the past 30 years from 24% in the period 1975 to 1984 to 49% in the most recent decade. The percentage of women who were charged with first degree murder also rose from 16% to 25% (Figure 9).
Figure 9 Percent distribution of persons accused of spousal homicide by gender and most serious violation, 1975 to 2004
In a majority of spousal homicides, there is a history of violence between the victim and the accused person (Figure 10). Between 1991 and 2004, there was a history of domestic violence in 59% of homicides against women by their male partners and 70% of homicides against men by their female partners. A history of violence was less common for legally married women and men and higher for those in common-law, separated and divorced relationships. The Homicide Survey indicates just that there was a known history of domestic violence in the relationship but does not indicate who was the perpetrator and who was the victim of the violence that preceded the homicide.
Figure 10 Percentage of spousal homicides with a history of domestic violence between victim and offender, by offender relationship to victim, 1991 to 2004
The decrease in the spousal homicide rate in recent years may be due to, among other factors, increased community-based support, mandatory charging policies and improved training of police officers. Researchers in Canada and the United States have investigated the correlations between these various factors and changes in spousal homicide rates over time (Pottie Bunge 2002; Dawson 2001; Dugan et al. 1999; Rosenfeld 1997). They have demonstrated statistically that the decline in spousal homicide rates is linked to a combination of increased availability of resources and improvements in women’s socio-economic status, including:
Due to the small number of spousal homicides in some provinces, data have been combined for the 30-year period, from 1975 to 2004. Over this time period, rates of spousal homicide against women were lowest in Newfoundland and Labrador and Prince Edward Island, while rates of spousal homicide against men were relatively low in New Brunswick and Quebec. The highest rates for both men and women are shown in the Western provinces and the territories. In almost all provinces and territories, homicides of women outnumbered homicides of men by a ratio of at least 2 to 12 (Figure 11).
These provincial variations mirror overall rates of homicide and violence in the general population, which tend to be higher in the territories and the Western provinces.
Table 2 gives a clearer indication of the magnitude of the problem of spousal homicide in each of the provinces and territories, showing both the total and the average number of women and men killed annually by spouses over the 30-year period.
Table 2 Number of spousal homicides in Canada, the provinces and territories by sex of victim, 1975 to 2004
The most detailed information on sexual assault is available from the 1993 national VAWS. At that time, 39% of Canadian adult women reported having had at least one experience of sexual assault since the age of 16. The definition of sexual assault in this survey included violent sexual attacks and unwanted sexual touching, both of which are consistent with Criminal Code definitions of sexual assault.
The GSS does not contain the same kind of expanded data related to sexual assault outside a spousal relationship (legal marriages and common-law unions); as a result, trend data are quite limited. The survey questions used to define sexual assault in the 1993 VAWS were more detailed than the questions used in the GSS. Therefore, comparisons should be made only between the two GSS surveys. The percentage of women who reported being sexually assaulted in the previous 12-month period was 3% in both 1999 and 2004.
Trend data do exist for police-reported cases of sexual assault. However, since victimization surveys suggest that less than 10% of sexual assaults are reported to the police, police data significantly underestimate the incidence of sexual assault.
The Criminal Code definition of sexual assault encompasses conduct ranging from unwanted sexual touching to sexual violence resulting in serious physical injury to the victim. Correspondingly, an offence is assigned to one of three levels according to the seriousness of the offence or the degree of physical injury sustained by the victim:
The statistics for reported sexual assaults show a steady increase starting in 1983 and a decline beginning in 1993. Overall sexual assault rates are driven by level I sexual assaults since they account for over 90% of all incidents reported to the police (Figure 12).
Figure 12 Total sexual assaults and level I sexual assaults reported to police, Canada, 1983 to 2004
The more serious levels II and III sexual assault are charted on a separate graph due to the much lower prevalence reported to police (Figure 13). These assaults show quite a different pattern: police-reported rates of levels II and III sexual offences have declined significantly since the legal reform in 1983 that abolished the crime of rape (see Appendix 2).
It is unknown to what extent these data reflect actual trends in changing levels of sexual violence in Canadian society, or changes in the willingness of sexual assault victims to bring these to the attention of the police. According to the 2004 GSS, just 8% of sexual assault victims reported the crime to the police.
In 1993, the offence of criminal harassment, also known as “stalking,” was introduced in the Criminal Code. Although criminal harassment is not gender-specific, the legislation was mainly introduced as a response to violence against women, and in particular, spousal assaults against women (Department of Justice Canada 2004). Several highly publicized cases of women being stalked and killed by their estranged partners in the early 1990s provided the impetus for this legislation, in the hope that early intervention in response to criminally harassing behaviour might prevent the escalation of violence.
Criminal harassment is obsessive behaviour directed towards another person. Section 264 of the Criminal Code defines criminal harassment as repeatedly following a person from place to place or repeatedly attempting to contact that person over a period of time. The legislation also encompasses such behaviours as watching or keeping watch over someone’s home or workplace, and making threats against another person known to the victim. As a result of such behaviour, the victims have reasonable cause to fear for their safety or that of someone close to them.
The law related to criminal harassment has undergone amendments on three occasions:
In 2004, three-quarters of incidents of criminal harassment reported to the police were directed at female victims. In half of these incidents, women were stalked by a person with whom they had an intimate relationship. The most common situations involved male ex-spouses (including former common-law partners) and ex-boyfriends.
In 2004, 2,030 male partners and 207 female partners were reported for stalking to the 68 police departments included in the database. The number of male spouses and boyfriends known to police for stalking has risen in recent years (this includes ex-partners) (Figure 14). This may reflect a real rise in stalking behaviour or an increase in the number of incidents reported to the police. It may also reflect a change in the way police have applied the law, as similar types of behaviours can be charged under other offences, such as uttering threats. Once again, these figures do not take account of possible increases in the population.
Figure 14 illustrates only stalking events that were reported to police. The 2004 GSS included a special module on stalking in order to more fully explore stalking events that were and were not reported to police.
Figure 14 Number of criminal harassment incidents reported to police, by offender relationship to victim, 1998 to 2004
Overall, 9% of persons 15 years of age and older reported having experienced at least one stalking incident in the five-year period prior to the survey. Rates were higher for women than for men: 11% compared with 7%. (See the Methodology section for question wording used on the GSS to measure stalking.)
Rates of criminal harassment victimization varied by province, from a low of 9% to a high of 13% for women and between 4% and 9% for men (Figure 15).
The majority of stalking victims were stalked by a male (80%). Stalkers were male in 80% of cases involving female victims and 73% of those involving male victims. Only 5% of all cases involved a female stalking a male.
Similar to other forms of violence, the relationship of stalkers to victims was somewhat different for women and men. Overall, 21% of female victims were stalked by current or former intimate partners (spouse or boyfriend/girlfriend) compared with 10% of male victims. Men were more likely to be stalked by other people known to them, such as neighbours, friends, co-workers and people known by sight only (Figure 16).
Figure 16 Percentage of women and men aged 15 years and over who reported some form of stalking, by relationship of stalker to victim, past five years, 2004
Not only do women report higher rates of stalking, they also report experiencing most of the specific stalking behaviours at a higher frequency. For example, they were more likely to be harassed with repeated telephone calls, repeated requests for a date, unwanted messages and gifts or letters, and were more likely to receive other forms of unwanted communication. Women were also more likely to be followed and spied on, and to have stalkers waiting outside their homes, workplace or other locations. Men reported higher prevalence of two behaviours: intimidation or threats to a third party, and damage to pets or property. Question wording on the GSS stipulates that for all types of stalking behaviour, the act must have caused them to fear for their safety or the safety of someone known to them (Figure 17).
Figure 17 Percentage of women and men aged 15 years and over who reported some form of stalking, by type of stalking, past five years, 2004
This is a relatively new area of study, but those who have examined stalking within the context of intimate partner relationships have found strong associations between physical and sexual violence by intimate partners and stalking (Tjaden and Thoennes 1998; Logan et al. 2000). The link between partner violence and stalking is confirmed by the GSS. Three-quarters of women who were stalked by an ex-partner within the previous five years also had been physically or sexually assaulted by an ex-partner.
Some research suggests that intimate partner stalkers may be the most dangerous of all (Palarea et al. 1999). In other words, stalkers are most likely to be violent toward those individuals with whom they have had an intimate relationship. Stalking has been identified as one of the primary risk factors for attempted and actual murder of female partners (McFarlane et al. 2002).
As shown in Figure 18, results of the GSS indicate that ex-partners are more likely than other stalkers to intimidate or threaten their victims and more likely to grab or attack them. Although substantial percentages of stalking victims in all relationship categories feared their lives were in danger, this climbs to 60% for women who were stalked by former spouses.
Figure 18 Female victims stalked by an ex-spouse are more likely to experience violence or threats, 2004
Summary of prevalence and severity of violence against women
These indicators of the prevalence and severity of spousal assault, spousal homicide, sexual assault, and criminal harassment were developed using victimization data and police-reported data.
According to two consecutive cycles of the GSS, women are more likely than men to be victims of the most severe and frequent forms of spousal assault. However, the prevalence and severity of this violence are showing signs of a decline. Although the actual number of women who reported experiencing spousal assault in the previous year remains constant, there has been a decline in five-year prevalence rates.
For spousal homicide, rates have decreased in recent years. However, it is still the case that more men than women kill their intimate partners every year.
Rates of sexual assault remained constant between 1999 and 2004, but the number recorded by police has declined since 1993. The most serious types of sexual assault, levels II and III have declined more dramatically. Women experience higher rates of criminal harassment than do men, and are twice as likely to be stalked by former intimate partners. Three-quarters of incidents of criminal harassment reported to police were directed at female victims.
The decline in the prevalence and severity of spousal assault suggested by victimization surveys, together with the decrease in spousal homicide, may be a result of improved social interventions and the increased use of services by abused women. However, additional data are needed to draw definitiveconclusions on the question of how societal efforts have intervened to reduce or prevent violence.
1. Provincial variations in the changing severity of spousal assault against women cannot be assessed as sample sizes in some provinces were too small to produce reliable estimates.
2. In order to make comparisons among geographic regions with different population sizes, it is necessary to standardize rates to a common unit. Rates of spousal homicide are presented as the number per 100,000 couples because of the small number of homicides in most jurisdictions.