Yaroslava Babych's presentation slides during the conference "Economic and Social Context of Domestic Violence".

1. Perceptions of violence and their
socio-economic determinants: a
comparative analysis
Yaroslava Babych (ISET)
with Lev Lvovskiy, Aleksandr Grigoryan and Norberto Pignatti
Economic and Social Context
of Domestic Violence

2. Do Perceptions Matter?
• Socio-economic factors and their relationship with IPV (intimate partner violence) received much
attention in empirical research. (Wang, 2016; Capaldi et al, 2012 , Arthur and Clark, 2009)
• The IPV phenomenon is multi-faceted (physical as well as psychological forms are recognized) and
correlate with a variety of socio-economic determinants. The understanding of these determinants
inform policy design to counteract IPV.
A logical chain of IPV:
• Perceptions (what is or is not perceived as a form of IPV) =>
• Attitudes (level of tolerance/justification of IPV) =>
• Actions (committing or not committing IPV, seeking help, getting involved to help the victims)
• The literature focuses mostly on the latter two links in this chain, while determinants of IPV Perceptions
receive far less attention and to our knowledge have not been studied systematically.
Perceptions ofViolence andTheir Socio-economic
Determinants: a Comparative Analysis
#AcademicsStandWithUkraine

3. Socio-economic determinants of IPV - what does the literature say?
Gender – incidence: While some studies report “gender symmetry” in the incidence of IPV, women
are far more likely to sustain serious physical injuries (Kimmel, 2002) as a result of IPV
Gender- attitudes: The tendencies for women to justify IPV more have been reported by some studies
(Waltermaurer, 2012 ) but the variation is large across countries.
Age – incidence: Age-related IPV curve (Johnson et al. 2015) perpetration of IPV tend to peak in late
adolescense (17-21 y.o) for men and a bit later (21-24 y.o) for women.
Age- attitudes: Results are not consistent. Some studies show higher justification of IPV among older
adults Gracia and Tomás, 2014; Gracia et al., 2015). Others show that younger participants justified
wife beating more often (Rani and Bonu, 2009, a groups of Central and South Asian countries )
Low education, Lower social status, Unemployment, Low income are shown to correlate with both
incidence of IPV ( Capaldi et al, 2012) and higher justification of IPV (Wang, 2016)
Perceptions of Violence and Their Socio-economic
Determinants: a Comparative Analysis
#AcademicsStandWithUkraine

4. Social Capital and IPV:
• There is empirical evidence showing that an increase in social capital decreases the likelihood
of IPV (social capital conceptualized as a stock of social resources and networks such as personal
and community networks, sense of belonging, civic engagement, norms of trust) : O’Campo et
al. 1995; Browning 2002; Tomaszewski 2002; among others.
• Emery et al. (2011) find that trust as a component of collective efficacy has a protective effect
for IPV but only when the neighborhood features low non-intervention norms.
• The relationship between social capital (such as social cohesion, trust and quality of
neighborhood measures) and perceived IPV is understudied.
Perceptions of Violence and Their Socio-economic
Determinants: a Comparative Analysis
#AcademicsStandWithUkraine

5. Perceptions ofViolence andTheir Socio-economic
Determinants: a Comparative Analysis
#AcademicsStandWithUkraine
0
5
10
15
Average
Score
A
r
m
e
n
i
a
B
e
l
a
r
u
s
G
e
o
r
g
i
a
L
a
t
v
i
a
P
o
l
a
n
d
R
u
s
s
i
a
S
w
e
d
e
n
U
k
r
a
i
n
e
Average Overall Violence Perception
Male Female
• Perceptions of violence, 15 questions:
„… in your opinion these are examples of
abuse within the couple…”:
1. Beating causing severe physical harm
2. Beating without causing severe physical harm
3. Sexual act against the partner’s will
4. Threat to do harm by using the couple’s children
5. Locking the partner in a room, apartment, house
6. Verbal threats of physical violence
7. Constant humiliation, criticism
8. Quarrels, scandals, screams
9. Prohibition to communicate with friends and / or relatives
10. Forced abortion
11. Intentional refusal to get medical care
12. Prohibition to visit public places without permission
13. Restrictions on access to financial resources
14. Requirement to show SMS, correspondence in social networks
15. Prohibition to dress as one likes

6. Perceptions ofViolence andTheir Socio-economic
Determinants: a Comparative Analysis
#AcademicsStandWithUkraine
Physical violence:
1. Beating causing severe physical harm
2. Beating without causing severe physical harm
3. Sexual act against the partner’s will
4. Forced abortion
0
1
2
3
4
Average
Score
A
r
m
e
n
i
a
B
e
l
a
r
u
s
G
e
o
r
g
i
a
L
a
t
v
i
a
P
o
l
a
n
d
R
u
s
s
i
a
S
w
e
d
e
n
U
k
r
a
i
n
e
Average Physical Violence Perception
Male Female

7. Perceptions ofViolence andTheir Socio-economic
Determinants: a Comparative Analysis
#AcademicsStandWithUkraine
Emotional violence:
1. Threat to do harm by using the couple’s
children
2. Verbal threats of physical violence
3. Constant humiliation, criticism
4. Quarrels, scandals, screams
5. Intentional refusal to get medical care
6. Requirement to show SMS, correspondence in
social networks
0
2
4
6
Average
Score
A
r
m
e
n
i
a
B
e
l
a
r
u
s
G
e
o
r
g
i
a
L
a
t
v
i
a
P
o
l
a
n
d
R
u
s
s
i
a
S
w
e
d
e
n
U
k
r
a
i
n
e
Average Emotional Violence Perception
Male Female

8. Perceptions ofViolence andTheir Socio-economic
Determinants: a Comparative Analysis
#AcademicsStandWithUkraine
Restrictions as violence:
1. Locking the partner in a room, apartment,
house
2. Prohibition to communicate with friends
and / or relatives
3. Prohibition to visit public places without
permission
4. Restrictions on access to financial
resources
5. Prohibition to dress as one likes
0
1
2
3
4
5
Average
Score
A
r
m
e
n
i
a
B
e
l
a
r
u
s
G
e
o
r
g
i
a
L
a
t
v
i
a
P
o
l
a
n
d
R
u
s
s
i
a
S
w
e
d
e
n
U
k
r
a
i
n
e
Average Restrictions Perceptions
Male Female

9. Determinants of violence perceptions
(1) (2) (3) (4) (5) (6) (7) (8)
Sweden Transitiona
Physical Emotional Restriction Overall Physical Emotional Restriction Overall
<25 0.068 0.115 0.080 0.263 0.219*** 0.154** 0.337*** 0.710***
25-34 0.016 -0.155* 0.031 -0.108 0.084*** -0.042 0.101* 0.144
35-44 0.012 -0.136** -0.041 -0.165 0.059** -0.015 -0.005 0.039
65+ -0.108** 0.078 -0.120 -0.150 -0.065* -0.057 -0.089 -0.210
Female 0.091*** 0.157*** 0.133*** 0.381*** 0.129*** 0.474*** 0.397*** 1.000***
Higher education 0.061 -0.085 0.272** 0.248 0.187*** 0.300*** 0.347*** 0.833***
Upper and post
secondary
0.049 -0.124 0.231** 0.155 0.062* 0.058 0.089 0.209
Married 0.003 -0.015 0.095 0.083 0.042 0.071 0.114** 0.227**
Not married with partner 0.042 0.161** 0.113* 0.316* 0.016 0.072 0.017 0.105
Stable or improving
financial situation
-0.009 -0.019 -0.064 -0.092 -0.001 0.017 0.028 0.044
Awareness of legislation 0.043 0.167*** 0.092* 0.303** 0.030 0.067* 0.094** 0.192**
Employed 0.042 0.095 0.094* 0.232* 0.103*** 0.154*** 0.131*** 0.388***
Children 0.003 0.026 0.034 0.063 0.003 -0.011 -0.019 -0.026
Urban 0.034 -0.053 0.068 0.049 0.029 0.090* 0.134*** 0.252**
Trust others 0.005 -0.009 -0.004 -0.008 0.009** 0.022*** 0.036*** 0.066***
Pray -0.039* -0.040 -0.055** -0.134** -0.001 -0.011 -0.028*** -0.040
Bad Environment 0.006 0.015 0.011 0.032 0.056*** 0.079** 0.077* 0.213**
Bad Social -0.016 -0.022 0.016 -0.021 -0.008 0.069 0.062 0.123
Person per room -0.100* -0.157* -0.069 -0.326* -0.018 -0.039 -0.052** -0.109**
Armenia -0.251*** -0.852*** -1.214*** -2.316***
Georgia 0.007 0.050 -0.301*** -0.245*
Latvia 0.067** -0.033 -0.323*** -0.290**
Russia -0.314*** -0.997*** -1.188*** -2.500***
Ukraine -0.106*** -0.383*** -0.739*** -1.228***
Belarus -0.132*** -0.713*** -0.902*** -1.747***
Observations 866 866 866 866 6,008 6,008 6,008 6,008

10. Determinants of violence perceptions: age
(1) (2) (3) (4) (5) (6) (7) (8)
Sweden Transitiona
Physical Emotional Restriction Overall Physical Emotional Restriction Overall
<25 0.068 0.115 0.080 0.263 0.219*** 0.154** 0.337*** 0.710***
25-34 0.016 -0.155* 0.031 -0.108 0.084*** -0.042 0.101* 0.144
35-44 0.012 -0.136** -0.041 -0.165 0.059** -0.015 -0.005 0.039
65+ -0.108** 0.078 -0.120 -0.150 -0.065* -0.057 -0.089 -0.210
Reference groups are: Age 45-64
#AcademicsStandWithUkraine
a Countries in the Transition group are: Armenia, Belarus, Georgia,
Latvia, Poland, Russia, and Ukraine

11. Determinants of violence perceptions: gender and
education
Reference groups are: Male; Low secondary and lower
#AcademicsStandWithUkraine
(1) (2) (3) (4) (5) (6) (7) (8)
Sweden Transitiona
Physical Emotional Restriction Overall Physical Emotional Restriction Overall
Female 0.091*** 0.157*** 0.133*** 0.381*** 0.129*** 0.474*** 0.397*** 1.000***
Higher education 0.061 -0.085 0.272** 0.248 0.187*** 0.300*** 0.347*** 0.833***
Upper and post
secondary
0.049 -0.124 0.231** 0.155 0.062* 0.058 0.089 0.209

12. Determinants of violence perceptions: awareness,
employment, urban
Reference groups: Note aware of anti-IPV legislation; Not employed; Rural area or village
#AcademicsStandWithUkraine
(1) (2) (3) (4) (5) (6) (7) (8)
Sweden Transitiona
Physical Emotional Restriction Overall Physical Emotional Restriction Overall
Awareness of
legislation
0.043 0.167*** 0.092* 0.303** 0.030 0.067* 0.094** 0.192**
Employed 0.042 0.095 0.094* 0.232* 0.103*** 0.154*** 0.131*** 0.388***
Urban 0.034 -0.053 0.068 0.049 0.029 0.090* 0.134*** 0.252**

13. Determinants of violence perceptions: trust, religiosity,
environmental problems, persons per room
(1) (2) (3) (4) (5) (6) (7) (8)
Sweden Transitiona
Physical Emotional Restriction Overall Physical Emotional Restriction Overall
Trust others 0.005 -0.009 -0.004 -0.008 0.009** 0.022*** 0.036*** 0.066***
Pray -0.039* -0.040 -0.055** -0.134** -0.001 -0.011 -0.028*** -0.040
Bad
Environment
(noise,
pollution)
0.006 0.015 0.011 0.032 0.056*** 0.079** 0.077* 0.213**
Persons per
room
-0.100* -0.157* -0.069 -0.326* -0.018 -0.039 -0.052** -0.109**
#AcademicsStandWithUkraine

14. Determinants of violence perceptions: country effects
#AcademicsStandWithUkraine
(5) (6) (7) (8)
Transitiona
Physical Emotional Restriction Overall
Georgia 0.007 0.050 -0.301*** -0.245*
Armenia -0.251*** -0.852*** -1.214*** -2.316***
Ukraine -0.106*** -0.383*** -0.739*** -1.228***
Russia -0.314*** -0.997*** -1.188*** -2.500***
Latvia 0.067** -0.033 -0.323*** -0.290**
Belarus -0.132*** -0.713*** -0.902*** -1.747***
Reference country: Poland
Countries listed from lowest to the highest gender equality scores (UN).

15. Likelihood of “Do not know” or
“Refuse to Answer”
Perceptions of Violence and Their Socio-economic
Determinants: a Comparative Analysis
#AcademicsStandWithUkraine
(1) (9)
Sweden Transitiona
<25 -0.146 -0.213***
25-34 -0.083 -0.132***
35-44 0.006 -0.083**
65+ 0.095 0.178***
Female -0.118** -0.161***
Higher education -0.258* -0.036
Upper and post secondary -0.260* 0.029
Observations 850 5,909
R-squared 0.053 0.050

16. Perceptions ofViolence andTheir Socio-economic
Determinants: a Comparative Analysis
#AcademicsStandWithUkraine
(1) (9)
Sweden Transitiona
Number of children -0.045* 0.015
Urban -0.060 -0.086**
Pray 0.027 0.021***
Georgia 0.042
Armenia 0.182***
Ukraine 0.166***
Russia 0.403***
Latvia 0.009
Belarus 0.216***
Likelihood of “Do not know”
or “Refuse to Answer”

17. • In order to to take into consideration contemporaneous correlation of common unobserved factors, jointly
relevant for the three dimensions of violence perceptions, we estimate a two-equation seemingly
unrelated regression (SUR) model for Sweden, and transition countries jointly and separately.
• Zellner (1962) shows that with the same covariates SUR and equation-by-equation OLS estimates produce
the same coefficients, and the SUR correction refers to standard errors (hence the significance of
coefficients).
• In our case, Breusch-Pagan specification test of independent errors rejects the null hypotheses that errors
are independent implying that the model for the three dimensions should be estimated by SUR. Estimated
cross correlations of errors are in the range (0.6 – 0.7).
• When comparing the significance of SUR and OLS estimates, we observe that in most cases 5-percent or
higher significance in OLS estimates is preserved in SUR estimates. Overall, we do not observe major
significance differences in the two models.
Robustness check: controlling for correlation from common
unobserved factors
#AcademicsStandWithUkraine