The research conducted in this study was based on data derived
from the Winnipeg Health and Drinking Survey (WHADS) (Barnes
&
Murray, 1989). This survey was designed to examine the
prevalence of
alcohol consumption in a general population sample of
Winnipeg,
Manitoba by exploring the socio- demographic and personality
variables associated with its occurrence. The survey followed
a
longitudinal panel design and includes two waves of
data. The survey
attempted to capture a wide range of information on people's
drinking
and lifestyle practices across three age groups over a two
year time
period.
Partner abuse data have been
previously analyzed for Wave 1 of
the WHADS. A brief description of the overall sampling
strategy
involved the WHADS, the response rate for Wave 1 data, the
format for
the Wave 1 partner abuse data analyses as well as summaries
of its
major findings follows. This chapter concludes with:
1) a discussion of the
results and limitations of Wave 1 partner
abuse data,
2) a presentation of an
alternative theoretical model used in
this research that attempts to
overcome the methodological
weaknesses of Wave 1,
3) a list of the objectives and assumptions of Wave 2, and
4) a list of the research hypotheses tested in this research.
A discussion of the measures included in both waves of
this research
is provided in the chapter on methodology.
A random sample of adult residents of Winnipeg, Manitoba
between
the ages of 18 and 65 years who were not institutionalized
was
provided for use by the Manitoba Health Services Commission,
the
provincial medicare agency. The initial sample was
stratified by age
and sex into the following categories for males: (a) 18-34
years, (b)
35-49 years, and (c) 50-65 years; and females: (a) 18-34
years, (b)
35-49 years, and (c) 50-65 years. For each sex/age cell,
there were
667 randomly selected names of Winnipeg residents. From this
initial
sample, a total of 2761 introductory letters were mailed (See
Appendix A). Data collection for Wave 1 began during the
summer of
1989 and was completed by the fall of 1990.
In both phases of the project, respondents participated in a
90
minute face-to-face interview which involved completing a
structured
interview schedule as well as a self-administered
questionnaire. The
interviews were conducted by graduate students from the
departments
of Psychology, Sociology, and Family Studies at the
University of
Manitoba who received training during a full day workshop.
Each personal interview was
preceded by approximately one week
by a letter describing the purpose of the project.
Respondents were
invited to call the project office should they have any
questions or
concerns. A telephone appointment was made prior to the
interview.
Interviews were schedules to take place at the subject's home
(unless
otherwise arranged). The interviews were conducted at a
time most
convenient to the subject. At least five attempts were
made to
contact each individual to arrange for a suitable interview
time.
The response rates to be described are based on the entire
sample. Of the initial sampling base, 366 persons were deemed
ineligible, 722 refused to be interviewed and 336 could not
be
contacted. The number of completed interviews was 1257 (615
males and
642 females) and provided an overall response rate of 63.5%.
Data
analyzed in this research were drawn from a subsample of 447
males
and 452 females who were married or remarried. Table 1
provides a
summary of the demographic characteristics for this
subsample.
Males
Females
Category
N
%
N %
Mean Age 46.1 Years 43.5 Years
Age Groups
18-34
years
104
23.3
145 32.1
35-49
years
155
34.7
150 33.2
50 years
+
188
42.1
157 34.7
Total 447
100.0
452 100.0
Marital Status
Married
429
96.0
443 98.0
Married, but
previously 18
4.0
9 2.0
divorced
Total 447
100.0
452 100.0
Educational Status
Grade
school
27
6.0
128 26.1
Some high
school 93
20.0
110 24.1
High school
completed 88
19.7
115 25.4
Some college
or 110
24.6
79 17.5
technical diploma
University
graduate 88
19.7
17 3.8
Post university education
41
9.2
13 2.9
Total 447
100.0
452 100.0
Table 1.
cont'd.
Table 1 (continued)
Category
Males
Females
N
%
N %
Current Employment Status
Working full
time 365
81.7
172 38.1
Working part
time 12
2.7
108 23.9
Unemployed
11
2.5
10 2.2
Student
2
.4
10 2.2
Housewife
___
___
116 25.7
Retired
43
10.1
31 6.9
Other
14
2.6
10 2.2
Total* 447
100.0
457 101.5
Income
<
$10,000/yr.
5
1.1
4 .9
$10,000 -
$20,000/yr. 16
3.6
34 7.5
$20,000 -
$35,000/yr. 88
19.7
88 19.5
$35,000 -
$50,000/yr. 127
28.4
117 25.9
>
$50,000/yr
196
43.8
161 35.6
Total* 432
96.6
404 89.4
Religious Preference
Catholic
117
26.7
143 31.6
Protestant
197
44.1
205 45.4
Jewish
14
3.1
11 2.4
Other
51
11.4
47 10.4
No religious
preference 67
15.0
46 10.2
Total* 446
99.3
452 100.0
Race
White
417
93.3
417 92.3
Non
white
30
6.7
35 7.7
Total 447
100.0
452 100.0
* Note: Not all totals will equal 447 or 452 (100%) due to missing
data or multiple categories.
Data in this phase of the research were analyzed with the
Statistical Package for the Social Sciences, version X
(SPSSx) and
the Statistical Analysis Systems (SAS). The following
represents the
general format of data analyses for this phase of the
research:
1. Frequency analyses (using
percentages and means) were
conducted to describe the
frequency and severity of perpetrated
partner abuse as well as the
demographic variables that define
the sample. At this stage
of the analysis, it was also possible
to test for scale outliers,
skewness, normality, linearity and
homoscedasticity and to conduct
appropriate transformations when
necessary. Although the
CTS (Straus, 1979) was found to be
skewed, the criteria set by
Cleary and Angel (1984) suggested
the use of linear multiple
regression was an appropriate
statistical approach to analyze
these data. Finally,
examination of the
distributions of the alcohol dependence
measures established their
high/low cutpoints and facilitated
the construction of the Alcohol
Dependence Index.
2.Pearson Correlations were
computed to determine the bivariate
relationships among continuous
or dummy coded variables. As
noted previously, correlational
analyses determined the measures
needed to be included in the
Neuroticism Index. In addition,
zero order relationships
between the dependent measure, partner
abuse and the independent
measures, socio-demographic,
personality, alcohol measures
were determined.
3. A number of standard
regression models testing both main and
interaction effects were
performed on continuous and dummy coded
variables. These involved
examining both the individual and
combined effects of the
independent variables against the
dependent variable (i.e., six
items taken from the CTS).
Through this procedures, it was
possible to ascertain which
variables were most salient in
their explanation of partner
abuse.
The following represents a summary of the results of the
analyses conducted on the male and female data from Wave 1 of
this
research.
1. The prevalence of male
perpetrated abuse was 26.3 percent
with the most common abuse tactic being "pushing,
grabbing and
shoving" (Barnes, Sommer, Murray & Patton, 1994)
(See Table 2).
Number of
Type of
violence
occurrences %
Minor violence acts
Threw or smashed
something
79 15.8
(not at partner)
Threatened to throw
something
42 7.3
(not at partner)
Threw something at
partner
30 4.6
Pushed, grabbed or
shoved
85 17.2
Severe violence acts
Hit
partner
42 7.3
Hit partner with something
hard
14 .9
Violence Indexes
Minor
Violence
128 25.9
Severe
Violence
43 7.6
Overall
Violence
125 26.3
There were 10 missing cases. Overall violence scale
statistics:
Mean = 6.63, S.D. = 1.48 and
range = 6-20
2. Partner abuse by male
respondents was significantly predicted
by being nonwhite, unemployed and alcohol dependent, and by a
low
score on Eysenck's Lie Scale and a high score on the
Neuroticism
Index. In addition, an interaction effect was found for
high alcohol
consumption and high scores on the Neuroticism Index. This
interaction effect was found to be the strongest predictor of
male
perpetrated partner abuse. Table 3 illustrates the
results of a
standard regression model testing both main and interaction
effects.
Predictor r Beta R2
Income
-0.08 0.04
White
-0.09* -0.14**
Age
-0.10* -0.05
Unemployment
-0.25*** -0.23***
Years of
education
-0.08 -0.09
Catholic
-0.04 -0.06
Protestant
-0.04 0.02
Ethanol
0.13* 0.11
Alcohol
dependence
0.26*** 0.16**
MacAndrew
Scale
0.14* 0.12
EPQP
0.17** -0.12
EPQE
0.07 0.04
EPQL
-0.22** -0.22**
Neuroticism
Index
0.23*** 0.23***
Alcohol consumption &
Mac 0.16* -0.38
Alcohol consumption &
Neuroticism
Index
0.32*** 0.25**
Alcohol consumption &
EPQL -0.32*** -0.06
Alcohol consumption &
EPQE 0.11
0.07
Alcohol consumption &
EPQP 0.27*** 0.25
Equation
.26
* p<.05, ** p<.01, *** p<.001, F(19,311) = 5.92,
p<.001,
adj. R2 = .22
1. The prevalence of female
perpetrated partner abuse was 39.1
percent with the most common abuse tactic also being
"throwing or
smashing something (but not directly at partner)"
(23.6%). See Table
4.
Number of
Type of
violence
occurrences %
Minor violence acts
Threw or smashed
something
108 23.6
(not at partner)
Threatened to throw
something
69 14.9
(not at partner)
Threw something at
partner
75 16.2
Pushed, grabbed or
shoved
92 19.8
Severe violence acts
Hit
partner
73 15.8
Hit partner with something
hard
16 3.1
Violence indexes
Minor
violence
173 38.0
Severe
violence
75 16.2
Overall
violence
178 39.1
Note: There are 2 missing cases. Overall violence scale
statistics:
Mean = 7.38, S.D. = 3.37,
and range = 6-28.
2. Partner abuse by female
respondents was significantly
predicted by being young in age and having high scores on
Esyenck's
Psychoticism Scale (EPQ-R), the Neuroticism Index and the
MacAndrew
Scale. An interaction effect was found between alcohol
consumption
and the EPQ-P. The strongest predictor of partner abuse
by women was
the main effect of having high scores on the EPQ-P.
Table 5
illustrates the results of a standard multiple regression
model
testing both main and interaction effects.
Predictor r Beta R2
Age
-0.23*** -0.21***
Years of
education
-0.07 -0.02
Unemployment
0.05 0.03
Income
-0.14** -0.06
Catholic
-0.01 -0.03
Protestant
0.00 0.04
White
0.10* -0.003
Ethanol
0.001 0.14
Alcohol
dependence
0.02 -0.07
EPQP
0.39*** 0.30***
EPQE
0.10* -0.003
EPQL
-0.17** -0.03
Neuroticism
Index
0.32*** 0.35***
MacAndrew
Scale
0.16** 0.16*
Alcohol consumption & EPQP
-0.001 -0.38*
Alcohol consumption &
Neuroticism
Index
0.20*** -0.13
Alcohol consumption & EPQE
-0.001 0.19
Alcohol consumption & EPQL
-0.03 -0.15
Alcohol consumption &
Mac -0.004
0.08
Equation
.28
Note: * p < .05, ** p < .01, *** p <.001
F(19,240)=4.87, p <
.001, adj. R2 =0.22.
Wave 1 research has made an important contribution to the
family
violence literature because it is one of the first general
population
studies to investigate the prevalence of partner abuse by
including
both socio-demographic and individual variables in its
analyses.
Yet, in spite of the findings provided by this research, a
number of
theoretical and empirical questions have been raised. A
discussion of
Wave 1 results and its subsequent limitations follows.
Wave 1 data provides support
for the role of both
socio-demographic and individual risk factors in the
perpetration of
partner abuse. As might be expected, the regression
analyses yielded
abuser profiles that differed for males and females.
The one common
predictor, however, is the finding that both male and female
abusers
are most likely to score high on the Neuroticism Index.
Results demonstrated that
male and female partner abusers can be
distinguished in terms of how alcohol abuse influences the
likelihood
of perpetration. Whereas alcohol dependence and the
consumption of
alcohol (as it interacted with Neuroticism) were found to be
salient
factors in the prediction of male perpetrated violence, the
same did
not hold true in the prediction of female perpetrated
violence.
Although the interaction between Eysenck's P Scale and
alcohol
consumption was found to be a weak predictor for female
abusers, it
appears that the role of alcohol in partner abuse is
experienced
differently by males and females.
In an attempt to explain
this sex difference, our article on
female perpetrated spouse abuse (Sommer et al., 1992) relied
upon
Frieze and Schafer's (1984) cognitive model. According
to this
model, "a drinker's reactions will depend upon the
social context in
which drinking occurs and the prior expectations of the
person about
how the alcohol will affect him or her" (p.277). We
suggested that
"the effects of alcohol consumption are thought to be
dependent upon
an individual's cognitive interpretation of the physiological
arousal
experienced in a manner that is consistent with prior
sex-role
socialization" (Sommer et al., 1992, p. 1321). For
women, the
physiological effects associated with alcohol consumption may
be
interpreted as emotional warmth, whereas for men, it may be
interpreted as power.
The contextual role of
alcohol abuse in the perpetration of
partner abuse is an issue that has arisen from this research.
Specifically, the question "Is alcohol consumed during
an abuse
episode?" is in need of being answered. In doing
so, it will be
possible to determine whether it is the reaction to the
immediate
effects of alcohol consumption, or merely the alcoholic
lifestyle,
that is most influential in the perpetration of partner
abuse.
The socio-demographic risk
factors found to be significant for
male (i.e., unemployment, nonwhite) and female abusers (i.e.,
young
age) are also consistent with the findings of other general
population surveys on partner abuse.
Contrary to those derived from clinical data (Gondolf et
al., 1990),
abusers in this, and other general population based research,
were
not necessarily defined as being of low SES backgrounds even
though
the male data demonstrated that being unemployed was a risk
factor.
Further, the finding that being nonwhite was also a salient
factor in
the prediction of male perpetrated spouse abuse, needs to be
considered cautiously since our male sample was 93.3 percent
white.
Finally, the finding that female abusers are most likely to
be young
in age is consistent with other research (Kennedy &
Dutton, 1989;
Shupe & Stacey, 1987; Stacey & Shupe, 1983; Straus et
al., 1980), as
well as with the profile of the deviance prone individual
(Sommer et
al., 1992).
Although Wave 1 data
included the socio-demographic variables
most often used to define a sample, several other variables
commonly
investigated in family violence research were omitted.
For example,
the effects of violence in the family and life stress events
were not
considered in the analysis of Wave 1 data. In view of
the extensive
literature suggesting that each of these factors are
important
correlates of partner abuse, it is thought that they too, may
add to
the explanatory power of a regression model.
Finally, with respect to the
prevalence of perpetrated partner
abuse, Wave 1 data demonstrated that 26.3 percent of males
and 39.1
percent of females in this random sample of adult Winnipeg
residents
acknowledged at least one incident of partner abuse during
the course
of their relationship with their current partner. While
the
prevalence rates of partner abuse reported are consistent
with Straus
et al.'s (1980) findings, explanations regarding a
significant sex
difference (Sommer et al., 1991) in its occurrence remain a
matter of
speculation.
The recency of the partner
abuse was also not examined.
Establishing whether the abuse reported is part of a well
established
pattern or behaviour, or simply an isolated event, possibly
occurring
early on in the relationship is a matter that remains
unresolved.
Contextual issues (i.e., the perpetration of abuse occurring
in self
defence), as well as the consequences of partner abuse (i.e.,
partner's need for medical attention following a partner
abuse
episode) were not examined, and remain a challenge for future
research.
The limitations just
described indicate that the study of abuse
between intimate partners like any other area of research is
not
without its problems. It was the goal of this study to be
sensitive
to these issues and to attempt to overcome them by way of the
methods
outlined in the following chapter. In so doing, some of
the
uncertainty regarding the dynamics underlying both male and
female
perpetrated partner abuse was clarified.
In order to address both the socio-demographic and
psychological
variables represented in this research, and in order to
provide a
more definitive explanation of partner abuse, a diathesis
stress
model was selected for those purposes. This approach
"considers the
often subtle interaction between a predisposition toward
disease (or
disorders) - the diathesis - and the environmental, or life
events
disturbing people - the stress" (p. 55, Davison &
Neale, 1990).
According to Davison and Neale (1990), an important tenet of
this
perspective is that both the diathesis and stress are
necessary
conditions in the development of a particular state.
Often used to
describe the development of schizophrenia (Buskist &
Gerbing, 1990),
this conceptual framework can also be extended to a number of
phenomena whose underpinnings are thought to involve both
biologically and environmentally based vulnerabilities.
According to
Graff (1993), vulnerability factors are thought to be
relatively
stable, whereas stressors may be experienced as acute or
chronic.
The diathesis-stress model was thought to be particularly well
suited toward explaining the problem of partner abuse.
Our earlier
research on courtship violence employed this conceptual
framework
(Barnes et al., 1991) and demonstrated that the inclusion of
an
interaction term (i.e., personality and alcohol consumption)
in the
diathesis-stress regression model increased the amount of
variance
explained over the amount explained by the main effects by
seven
percent. Previously, theories that have guided research
demonstrating associations between a number of
socio-demographic
variables, personality, alcohol consumption, and the
perpetration of
abuse between intimate partners have not been successful in
providing
a complete explanation of this phenomenon. The
application of a
diathesis-stress model allowed for the integration of
principles from
a number of the theories previously discussed.
In this research, by
conceptualizing personality, alcohol
dependence, socio-demographic variables (i.e., age,
employment
status, income, education, religion and race), family
background and
past partner abuse as the diathesis, and life stress events
(i.e.,
job loss, job change, change in financial situation, change
in
residence, birth of a child, retirement, start or finish
school,
change in marital status, spouse's job loss, spouse's job
change) and
recent alcohol consumption as the stress, an examination of
their
individual and combined effects in the prediction of partner
abuse
was attempted. In the presence of both constitutional
and
environmental risk factors, the likelihood for partner abuse
was
expected to increase. It was expected that the
application of a
diathesis-stress model to the issue of partner abuse would
improve
upon the explanatory value of previously advanced
models. Figure 1
illustrates the model tested in this research.
[Will be shown here when the graphics file becomes available. --WHS]
This research project assumed that references to partner abuse
were made within the context of heterosexual relationships.
Although
it is recognized that there is a growing body of literature
on gay
and lesbian partner abuse, it was deemed to be beyond the
scope of
this research to explore the perpetration of partner abuse
within
this population.
The objectives of this project were as follows:
1) Examine the longitudinal
relationships among
socio-demographic variables, alcohol consumption and
dependence,
personality and partner abuse.
1) Identify the incidence and
prevalence of partner abuse in a
large urban
Canadian sample.
2) Compare the rates of partner abuse between males and females.
3) Compare these rates of
partner abuse with those of other
Canadian and
U.S. urban samples.
4) Examine the stability of
partner abuse during the course of
the research
project.
5) Examine whether the abuse perpetrated occurred in self defence.
6) Examine the consequences
of partner abuse episodes based on
whether
medical treatment was sought.
7) Examine whether the
consumption of alcohol was a factor at
the time of
an abuse episode.
8) Examine whether the
relationship between the amount of
alcohol
consumed and the occurrence of partner abuse is
linear (i.e.,
such that the likelihood for abuse increases
with the
amount consumed).
9) Examine the relationship
between observing violence in the
family of
origin and the perpetration of partner abuse.
10) Examine the relationship
between recent life stress events
(i.e.,
those occurring during the past two years) and the
perpetration of partner abuse.
The empirical literature and the diathesis-stress model as
applied to abuse between intimate partners suggest several
testable
hypotheses. A description of the hypotheses tested in
this research
are as follows:
Descriptive statistics and
correlational analyses were expected to
reveal that:
1. The incidence rates of
male and female perpetrated partner
abuse (i.e.,
abuse that has occurred during the past
year) as
measured in Wave 2 will be consistent with those
reported in
the literature (i.e., 10-14 percent).
2. The pattern of partner
abuse (i.e., frequency, severity, most
common abuse
tactics and sex differences) found in Wave 1
of this
research will also hold true for Wave 2 data.
3. Partner abuse scores will
be significantly higher among
respondents
who had witnessed their parents' abuse of each
other (as
measured in Wave 2).
4. Partner abuse scores will
be significantly higher among
respondents
who had reported having experienced life stress
events (as
measured in Wave 2).
5. The relationship between
the consumption of alcohol and the
perpetration
of partnerabuse will be curvilinear whereby
individuals
who consume moderate amounts of alcohol will have
higher mean
partner abuse scores than those who consume low
and high
amounts of alcohol (as measured in Wave 1 and Wave
2).
The main hypotheses, related to the diathesis-stress model
were
assessed by way of logistic regression analyses and included
measures
drawn from Wave 1 and Wave 2. The individual and
combined main
effects of the independent measures on current perpetrated
partner
abuse (as measured in Wave 2) as well as their interactive
contributions to its prediction were evaluated for the
following:
6. For males, witnessing
mother's and father's abuse of each
other,
consuming alcohol (measured in Wave 2), being
unemployed,
perpetrating past partner abuse, having low
scores on
social conformity and having high scores on alcohol
dependence,
the Neuroticism Index, (measured in Wave 1) and
stress
(measured in Wave 2) will significantly predict
current
perpetrated partner abuse in Wave 2.
7. For females, witnessing
mother's and father's abuse of each
other
(measured in Wave 2) being young in age, perpetrating
past partner
and having high scores on Esyenck's Psychoticism
Scale
(EPQ-R), the Neuroticism Index, the MacAndrew Scale
(measured in
Wave 1) and stress (measured in Wave 2) will
significantly
predict current perpetrated partner abuse in
Wave 2.
Finally, based on the findings emerging from the
literature reviewed,
the following two hypotheses were derived:
8. For males, the
interaction between the following diathesis
and stress
factors will significantly predict current
perpetrated
partner abuse in Wave 2 and add to the
explanatory
power of the main effects model:
a) deviance prone personality (i.e. high scores on the
neuroticism index) and high recent alcohol consumption,
b) deviance prone personality (i.e., high scores on the
neuroticism index) and life stress,
c) past environmental contributions (i.e. past partner
abuse and violence in the family of origin) and high
recent alcohol consumption,
d) past environmental contributions (i.e., past partner
abuse and violence in the family of origin) and life
stress, and
e) alcohol dependence and life stress.
9. For females, the
interaction between the following diathesis
and stress
factors will significantly predict current
perpetrated
partner abuse in Wave 2 and improve upon the
explanatory
power of the main effects model.
a) high scores on deviance prone personality (i.e.,
neuroticism index and psychoticism scale) and past
partner abuse,
b) high scores on deviance prone personality (i.e.,
neuroticism index and psychoticism scale) and life
stress events,
c) high scores on deviance prone personality (i.e.,
neuroticism index and psychoticism scale) and high
recent alcohol consumption, and,
d) past environmental contributions (i.e., violence in the
family of origin and past abuse) and life stress.
Next: Chapter 4
HTML by Walter H. Schneider
___________
Updates:
2001 02 10 (format changes)
|
Walter H. Schneider Children need Both Parents --
Once a Parent, a Parent for Life You are visitor # |