Brief Communications

Homicide in the Canadian Prairies: Elderly and

Nonelderly Killings
AG Ahmed, MB1, Robin PD Menzies, MB, FRCPC2

Objective: To examine the psychosocial and clinical characteristics of male perpetrators of

elderly and nonelderly homicides in the Canadian Prairies.
Method: We examined data drawn from a study of 901 adult homicide offenders who were
incarcerated or on parole between 1988 and 1992 in Alberta, Saskatchewan, and Manitoba.
Results: Of those studied, 67 men were convicted of homicide involving 79 elderly vic-
tims, and 671 were convicted of homicide involving 675 nonelderly victims. Most perpe-
trators were single and engaged in irregular patterns of employment at the time of their
index offence. Fourteen (20.8%) offenders with elderly victims had a history of psychiatric
treatment, compared with 98 (14.6%) offenders with nonelderly victims; however, this dif-
ference was not statistically significant. Approximately 30% of both groups were diag-
nosed with personality disorders. A comparison of the index- offence characteristics
showed no significant differences between the 2 groups.
Conclusion: Our findings suggest that elderly individuals are more likely to be killed in
their own homes by strangers. Social isolation appears to be a significant risk factor in
cases of elderly homicide.
(Can J Psychiatry 2002;47: 875–879
See page 879 for author affiliations.

Clinical Implications
· There is a high rate of psychopathology among the perpetrators of homicide.
· Social isolation may be an important risk factor in cases of elderly homicide.
· There is a need to assess the risk of homicide in cases of domestic violence, particularly where
it involves nonelderly victims.

Limitations
· This was a retrospective study.
· There was no information on victim psychopathology.
· Diagnoses were based on clinical, rather than on structured, interview and were available for
only 40% of the subjects.

Key Words: elderly, homicide, social isolation, perpetrator, victims

anada has one of the highest life expectancies in the
C Western world, and more than 12% of the population is
aged 65 years or over (1) . This proportion is bound to increase
as the socioeconomic status and lifespan of Canadians con-
tinue to improve. Most elderly individuals lead productive
and independent lives in the community. However, a signifi-
cant proportion are isolated and frail, owing to failing physical
health. This group is often more vulnerable to various crimes,
including homicide. Elder abuse is of interest to mental health
professionals in general, and to psychiatrists in particular, be-
cause of the increased rate of mental health problems in both
perpetrators and victims (2). Although perpetrators of elder
abuse are likely to have a history of significant psychopathol-
ogy, most problems faced by the elderly are due more to ne-
glect or omission than to the commission of an act.
Perpetrators of elder abuse are also more likely to depend on

W Can J Psychiatry, Vol 47, No 9, November 2002 875

The Canadian Journal of Psychiatry—Brief Communications

their victims for financial assistance, while victims are more
likely to be socially isolated and frail (2).
Homicide generates intense emotional reaction in the popular
press, particularly when it involves elderly individuals. Most
homicides are the result of domestic discord or petty quarrels;
a relatively small proportion arise from altercations with
strangers or from other homicidal motives (3,4). Schaler has
divided homicide involving elderly individuals into 2 catego-
ries : killings independent of the victim’s age and killings in
which the perpetrator exploits the frailty and social isolation
of the victim (5). Compared with nonelderly victims, elderly
homicide victims are more likely to be killed in their own
homes, . Social isolation and weakness are significant risk fac-
tors (6). Direct physical assault in the form of blunt force inju-
ries and strangulation are relatively common causes of death
(7). Our study compares the psychosocial and clinical charac-
teristics of men convicted of elderly homicide with the charac-
teristics of men convicted of nonelderly homicide.
Method
The subjects in this study were 901 men incarcerated or on pa-
role between 1988 and 1992 in Alberta, Saskatchewan, and
Manitoba. We undertook a retrospective review of the Na-
tional Parole Board database. We reviewed the following
sources of information: police reports, Canadian Police Infor-
mation Centre (CPIC) data, autopsy reports, trial transcripts,
pre- and posttrial psychiatric reports, regular case-
management reports, psychological and psychiatric reports
documenting offender progress during incarceration, parole
board decisions, and parole officer reports. From these official
documents, we extracted offender and victim demographic
characteristics, index-offence characteristics, clinical infor-
mation, and psychosocially relevant variables. The data were
coded and tabulated for descriptive purposes and analyzed us-
ing the SPSS statistical package (8). We used chi-square with
Yates’ correction to test for significant differences between
proportions at the 5% level. We compared perpetrators of eld-
erly homicide and perpetrators of nonelderly homicide. We
corrected observed significant levels by adjusting the number
of comparisons made.
Results
We included a total of 901 men classified as homicide offend-
ers under the National Parole Board (NPB) during the study
period. In 738 (81.9%) cases, information was available on
victim age; 67 (9.1%) offenders killed victims aged 65 years
or over. The remaining 671 (90.9%) killed victims aged under
65 years. The offenders’ mean ages were 27.8 (SD 12.85)
years and 27.8 (SD 9.16) years, respectively. A higher propor-
tion of those with elderly victims were single at the time of
their index offences (61.2% vs 43.4%, c2 = 7.82, P < 0.05)
(Table 1).
Table 2 shows the psychiatric diagnoses and histories of the
perpetrators. A comparison of the psychiatric histories re-
vealed that 20.8% of perpetrators of elderly homicides and
14.6% of perpetrators of nonelderly homicides had psychiat-
ric treatment before the index offence (c2 = 1.87, P > 0.05).
Formal psychological assessment determined that 13.4% of
the former group, compared with 9.5% of the latter, had
below-average intellectual ability (c2 = 1.04, P > 0.05). Both

Table 1 Characteristics of perpetrators of elderly and non-elderly homicide

Characteristics Elderly victims Non-elderly victims c2 P value
n (%) n (%)
Marital status
Single 41 (61.2) 291 (43.4) 7.82 0.005
Married or common law 21 (31.3) 275 (41.0) 2.36 NS
Divorced or separated 05 (7.5) 105 (15.6) 3.22 NS
Irregular employment patterns 50 (75.0) 520 (77.5) 0.29 NS
Juvenile convictions 36 (53.7) 382 (57.0) 0.25 NS
Prior violent convictions 23 (34.3) 302 (45.0) 2.82 NS
Prior psychiatric treatment 14 (20.8) 98 (14.6) 1.87 NS
Family criminal record 18 (26.8) 255 (38.0) 3.24 NS
Family substance-related problem 39 (58.2) 403 (60.0) 0.09 NS
Below-average IQ 9 (13.4) 64 (9.5) 1.04 NS

Table 2 Psychiatric diagnoses in perpetrators and victims groupings

Psychiatric diagnoses Elderly victims Non-elderly victims c2 P value
n (%) n (%)
Major Axis 1 disorders 1 (1.5) 36 (5.4) 1.19 NS
Alcohol-related problem 3 (4.5) 29 (4.3) 0.06 NS
Personality disorders 20 (29.8) 198 (29.5) 0.00 NS
No diagnosis 3 (4.5) 18 (2.7) 0.21 NS
Unknown 47 (59.7) 390 (58.0) 0.07 NS

876
W Can J Psychiatry, Vol 47, No 9, November 2002
 Homicide in the Canadian Prairies: Elderly and Nonelderly Killings
groups had high rates of familial substance-related problems:
58.2% of those with elderly victims and 60% of those with
nonelderly victims had first-degree relatives with significant
substance use disorder. Both groups had similar rates of Axis I
psychiatric diagnoses. As well, similar proportions (30%) of
the 2 groups were diagnosed with personality disorders. Un-
fortunately, information on psychiatric diagnoses was not
available for almost 60% of the entire study sample. As far as
forensic history was concerned, 18 (26.9%) offenders with
elderly victims and 254 (37.9%) offenders with nonelderly
victims had a family criminal record. Twenty-three (34.3%)
offenders with elderly victims, compared with 302 (45.0%)
offenders with nonelderly victims, had been convicted for
prior violent offences. Both offender groups had similar rates
of juvenile convictions (Table 1).
We compared the assault locations in the elderly-victim group
with those in the nonelderly-victim group (Table 3). While
most elderly victims were killed in their homes (68.6% vs
32.0%, c2 = 35.73, P < 0.0005), mostnonelderly victims were
killed in public places (47.1% vs 22.4%, c2 = 15.03, P <
0.0005). Compared with elderly victims, a significantly
higher proportion of nonelderly victims were killed in homes
shared with their killers (13.6% vs 4.5%, c2 = 4.52, P < 0.05).
Sharp objects were the most common murder weapons in both
victim groups. About one-quarter (25.3%) of the elderly vic-
tims, compared with 15.2% of nonelderly victims, were killed
by beating ( c2 = 4.66, P < 0.05). Of the elderly victims, 82.0%
sustained moderate-to-severe injuries, compared with 56.3%
of nonelderly victims (c2 = 16.67, P < 0.0005).
Table 3 illustrates that the most common motive for homicide
in the elderly was robbery: 65.6% of the victims were killed
during a property crime, compared with 22.7% of nonelderly
victims killed with the same motive ( c2 = 57.80, P < 0.0005).
More than 74% of nonelderly victims, compared with 34% of
elderly victims, were killed during domestic quarrels (c2 =
47.17, P < 0.0005). Psychosis accounted for similar propor-
tions of both elderly and nonelderly homicides. Similar pro-
portions of offenders with elderly and nonelderly victims
consumed alcohol (60.7% vs 63.8%, c2 = 0.18, P > 0.05) or
street drugs (19.3% vs 23.1%, c2 = 0.47, P > 0.05) at the time
of the homicide. Perpetrators with elderly victims were more
likely to be strangers and less likely to be spouses or family
members: 7.4% of elderly victims were killed by their
spouses, compared with more than 20% of nonelderly victims
(c2 = 9.16, P < 0.005). The relationship of the offender to the
victim remains unknown in less than 10% of both groups.
Over 70% of the offenders with elderly victims, compared
with 64.8% of those with nonelderly victims, admitted re-
sponsibility at the time of arrest (c2 = 1.25, P > 0.05). About
the same proportions in the 2 groups pleaded guilty (41.5% vs
45.1%; c2 = 0.28, P > 0.05). When arrested, one-tenth (10.5%)
of the offenders with elderly victims, compared with 17.9% of
those with nonelderly victims, claimed amnesia for the homi-
cide event (c2 = 2.36, P > 0.05).
W Can J Psychiatry, Vol 47, No 9, November 2002
Discussion
Although the elderly are less likely to be victims of crime,
compared with younger adults, crime against the elderly is
real and cannot be ignored. This is of great concern for the eld-
erly, their families, and the community at large. Our study
draws attention to the demographic, offence, and clinical
characteristics of the perpetrators of an extreme crime against
the elderly. Of the 738 offenders studied, 67 (9.1%) had eld-
erly victims and caused 79 deaths. Demographically, the per-
petrators of this type of homicide are indistinguishable from
those with nonelderly victims, except that the former are more
likely to be single, while the latter are more likely to be di-
vorced or separated at the time of the index offence.
In this study, about 20% of those with elderly victims and
14.6% of those with nonelderly vicitms had psychiatric treat-
ment before the index offence. Formal psychological assess-
ment determined that 13.4% of the former group, compared
with 9.5% of the latter group, had below average intellectual
ability. The 2 groups had similar rates of personality disorders
and major Axis I disorders; this may reflect high rates of sig-
nificant psychopathology among the perpetrators of homicide
in general. The rate of substance use disorder in this study is,
however, unexpectedly low, compared with the general popu-
lation rate (9). Unfortunately, our study design did not allow
us to explore the possible explanation for these findings. They
may, however, reflect that records prior to the index offence
underrecognize substance-related problems. High alcohol
consumption at the time of the homicide was recorded in both
groups. This is significant for risk-assessment considerations
and recommendations to the National Parole Board.
Overall, strangers or acquaintances (in contrast to spouses or
relatives) killed most of the victims in this study, although this
was more common in the elderly victim group. A significantly
high proportion of nonelderly victims, compared with elderly
victims, were killed by their spouses. These findings agree
with published data, which reveal that elderly victims were
killed by either acquaintances or relatives in 23% and 43% of
cases, respectively (6). According to this published data, kill-
ings by relatives are comparatively less common (elderly vic-
tims accounted for 6.0%, and nonelderly victims accounted
for 6.4%, of those killed by relatives), although an earlier
study indicated higher rates (elderly victims 22% and noneld-
erly victims 13%) (6). We do not know what proportion of the
elderly killings in our study resulted from elder abuse, be-
cause that information was not collected in the reviewed data-
base. However, the aforementioned earlier study revealed that
elder abuse rarely accounts for homicide (6). This finding may
be the result of nondetection, underreporting, or the fact that in
many cases abuse is not always severe enough to cause death.
Our study findings are consistent with previous findings
(6,10) in that most elderly victims were killed at home.
877
The Canadian Journal of Psychiatry—Brief Communications

Table 3 Comparison of offence characteristics in elderly and non-elderly homicide

Elderly victims Nonelderly victims c2 P value
n (%) n (%)
Offence Characteristics
Location of assault
Victim’s home 46(68.6) 215(32.0) — —
Offender’s home 3(4.5) 49(7.3) 36.08 —
Shared home 3(4.5) 91(13.6) — 0.000
Public place 15(22.4) 316(47.1) — —
Method
Sharp object 20(30.0 232(34.6) — —
Blunt object 7(10.4) 57(8.5) — —
Gun shot 12(18.0) 187(27.9) 11.35 —
Beating 17(25.3) 102(15.2) — NS
Strangulation 5(7.4) 74(11.0) — —
Others 6(8.9) 19(2.8) — —
Extent of injury
Moderate to severe 55(82.0) 378(56.3) 16.67 0.000
Motive of assault
Domestic quarrel 23(34.3) 498(74.3) 47.17 0.000
Revenge 7(10.5) 126(18.8) 2.86 NS
Self defense 1(1.5) 18(2.7) 0.07 NS
During crime 44(65.6) 152(22.7) 57.80 0.000
Sexual assault 11(16.4) 112(16.7) 0.00 NS
Psychosis 4(6.0) 21(3.1) 0.85 NS
Others 1(1.5) 44(6.5) 2.04 NS
Relationship of offender to victim
Family member 4(6.0) 43(6.4) — —
Spouse 5(7.4) 158(23.5) 12.99 —
Stranger 28(41.8) 193(28.8) — 0.011
Acquaintance 27(40.3) 224(33.4) — —
Unknown 3(4.5) 53(7.9) — —
Alcohol consumption at index offense 41(60.7) 428(63.8) 0.18 NS
Street drugs use at index offense 13(19.3) 155(23.1) 0.47 NS
Admitted responsibility on arrest 48(72.2) 435(64.8) 1.25 NS
Gave self up 10(14.5) 109(16.2) 0.08 NS
Guilty plea 28(41.5) 303(45.1) 0.28 NS
Offence convicted
1st degree 11(16.9) 93(13.9) 0.33 —
2nd degree 39(58.0) 310(46.2) 3.52 0.036
Manslaughter 17(25.3) 268(39.9) 5.45 —
Motive of assault
Domestic quarrel 23 (34.3) 498 (74.3) 47.17 0.000
Revenge 7 (10.5) 126 (18.8) 2.86 NS
Self-defense 1 (1.5) 18 (2.7) 0.07 NS
During crime 44 (65.6) 152 (22.7) 57.80 NS
Sexual assault 11 (16.4) 112 (16.7) 0.00 NS
Psychosis 4 (6.0) 21 (3.1) 0.85 NS
Others 1 (1.5) 44 (6.5) 2.04 NS

878
W Can J Psychiatry, Vol 47, No 9, November 2002
 Homicide in the Canadian Prairies: Elderly and Nonelderly Killings

Nonelderly victims were comparatively less likely to be killed
in their homes and more likely to be killed in public places.
Analyses of the killers’ motives revealed that most elderly
victims were killed during a property crime, such as robbery,
whereas most nonelderly victims were killed in domestic
quarrels. These findings suggest that a significant proportion
of the elderly victims in this study fell in Schaler’s second
category, in which the perpetrators probably exploited the
physical frailty and social isolation of their victims (5). In our
study, the most common causes of death in both groups were
injuries from sharp objects, followed by beating and gunshot
injuries. However, a significantly higher proportion of eld-
erly, compared with nonelderly, victims were subjected to
more severe injuries and killed by beating. This finding agrees
with the previously reported high degree of direct physical as-
sault in elderly homicides (7).
The findings in this study indicate that elderly homicide vic-
that deserves attention in the provision of comprehensive
health care for this vulnerable group.
References
1. Statistics Canada. Population by age group. 1999 http://WWW.StatCan.CA/eng-
lish/Pgdb/People/Population/demo31a.htm
2. Wolf RS. Elder abuse: a family tragedy. Aging International 1994;21(1):60 –4.
3. Swigert VL Ronald AF. Patterns in criminal homicide: theory and research. In:
Wickman P, Whitten Initial?, editors. Readings in criminology. Lexington (MA):
Heath Lexington Books; 1978. p 191–206.
4. Goetting A. Patterns of homicide among the elderly. Violence and Victims
1992;7:203–15.
5. Schafer AT. Homicide in the elderly. A study of homicide in the aged based on
an Aachen autopsy sample of the 10-year period 1076 85. Arch Kriminol 1989;3
4:65–78.
6. Falzon AL, Davis GG. A 15-year retrospective review of homicide in the elderly,
J Forensic Sci 1998;43:371– 4.
7. Schmidt P, Dettmeyer R, Madea B. Victim-related aspects of homicide of elderly
persons in the catchment area of the Bonn Forensic Medicine Institute. Arch
Kriminol 1999;1 2:33– 41.
8. SPSS for Windows Version 7.5. Chicago (IL): SPSS Inc; 1995.
9. Kessler RC, McGonagle KA, Zhao S, Nelson CB, Hughes M, Eshleman S, and
others. Lifetime and 12-month prevalence of DSM-III-R psychiatric disorders in
the United States. Arch Gen Psychiatry 1994;51:8–19.
10. Copeland AR. Homicide among the elderly: the Metro Dade County experience,
1979 1983. Med Sci Law 1986;26:259–62.

tims are about twice as likely to be killed in their homes and

1½ times more likely to be killed by strangers than are noneld-
erly victims. Compared with nonelderly victims, elderly vic-
tims are also about 1½ times more likely to have sustained
moderate-to-severe injuries and 3 times more likely to be
killed during a crime. The foregoing underscores the signifi-
cance of social isolation as a risk factor for elderly homicide
Manuscript received January 2002 and accepted September 2002.
1
Assistant Professor, Forensic Service, Royal Ottawa Health Care Group, Ot-
tawa, Ontario.
2
Consultant Psychiatrist, Knox Manse and Regional Psychiatric Centre, Sas-
katoon, Saskatchewan.
Address for correspondence: Dr AG Ahmed, Forensic Division, Royal Ot-
tawa Hospital, 1145 Carling Avenue, Ottawa ON K1Z 7K4
e-mail: aahmed@rohcg.on.ca

Résumé : Homicides de personnes âgées et non âgées dans les Prairies

canadiennes
Objectif : Examiner les caractéristiques psychosociales et cliniques des auteurs masculins d’homi-
cides de personnes âgées et non âgées dans les Prairies canadiennes.
Méthode : Nous avons examiné des données tirées d’une étude auprès de 901 auteurs d’homicides
adultes qui ont été incarcérés ou libérés sur parole entre 1988 et 1992 en Alberta, en Saskatchewan et
au Manitoba.
Résultats : Sur les personnes étudiées, 67 hommes ont été condamnés pour homicide sur 79 victimes
âgées, et 671 ont été condamnés pour homicide sur 675 victimes non âgées. La plupart des auteurs
d’homicides étaient célibataires et présentaient des modèles d’emploi irréguliers au moment de l’in-
fraction à l’origine de la peine. Quatorze (20,8 %) auteurs d’homicides sur des victimes âgées avaient
des antécédents de traitement psychiatrique, comparativement à 98 (14,6 %) auteurs d’homicides sur
des victimes non âgées. Toutefois, cette différence n’était pas statistiquement significative. Environ
30 % des deux groupes ont reçu un diagnostic de troubles de la personnalité. Une comparaison des
caractéristiques de l’infraction à l’origine de la peine n’a indiqué aucune différence significative entre
les deux groupes.
Conclusion : Nos résultats suggèrent que les personnes âgées sont plus susceptibles d’être tuées dans
leur maison par des étrangers. L’isolement social semble être un facteur de risque significatif dans des
cas d’homicide de personnes âgées.

W Can J Psychiatry, Vol 47, No 9, November 2002 879