Daena: International Journal of Good Conscience

Psychometric Properties of the Questionnaire on Suffered Violence

Exercise of Couple

Psychometric Properties of the Suffered and Exerted Couple Violence

Questionnaire

José Moral de la Rubia, Ramos Basurto Sandra

A questionnaire that evaluates victimization and perpetration in couples has been developed in Mexico.
the so-called Partner Violence Questionnaire (PVQ). The PVQ assesses frequency, but not intensity. The
the objective of this study was to modify the CVP to assess the frequency and damage of victimization and
perpetration, and study its psychometric properties of internal consistency, factorial structure, and invariance
factorial between sexes. The modified questionnaire, called the Questionnaire on Suffered and Exercised Violence
Couple (CVSEP) was administered to a non-probabilistic sample of 240 participants (120 women and 120
men). With the 27 items of frequency of suffered violence (α = .95), two factors were defined: violence
physical/sexual (α= .88) and psychological/economic/social (α = .94). With the 27 items of damage suffered (α = .95) it
they defined four factors: damage suffered from economic/social violence (α = .93), sexual violence/blackmail (α
(.88), physical violence (α = .86) and jealousy violence (α = .83). With the 12 items of frequency of violence
exerted (α = .84) two factors were defined: psychological violence (α = .85) and another type of violence (α = .66)
Although it is recommended to reduce this scale to a factor with 10 indicators (α = .86). With the 12 items of damage
caused (α = .89) two factors were defined: damage from psychological violence (α = .88) and another type of violence
(α = .81). Victimization and perpetration indices were created. It is concluded that the CVSEP has consistency.
internal and structural validity.

Palabras clave:violencia de pareja, violencia sufrida, violencia ejercida, daño psicológico, psicometría.

A questionnaire assessing victimization and perpetration in couples has been developed in Mexico.
and it is named the Couple Violence Questionnaire (CVQ). The CVQ evaluates frequency, but does not evaluate
intensity. The aim of this research was to modify the Couple Violence Questionnaire to assess frequency and
damage of victimization and perpetration, and study its psychometric properties of internal consistency, factor
structure and sex invariance. The modified questionnaire, named Suffered and Exerted Couple Violence
Questionnaire (SECVQ) was applied to a non-probability sample of 240 participants (120 women and 120 men).
men). Two factors were defined using the 27 items on frequency of suffered violence (α = .95): physical/sexual
violence (α = .88) and psychological/economic/social violence (α = .94). Four factors were defined using the
27 items of suffered damage (α = .95): by social/economic violence (α = .93), sexual violence/blackmail (α =
.88), physical violence (α = .86) and jealousy-related violence (α = .83). Two factors were defined using the 12
items on frequency of exerted violence (α = .84): psychological violence (α = .85) and other kind of violence
(α = 66), although it is recommended that this scale be reduced to one 10-item factor (α = .86). Two factors
were defined using the 12 items on caused damage (α = .89): by psychological violence (α = .88) and other kind
of violence (α = .81). The indexes of victimization and perpetration were created. It is concluded that the
SECVQ has internal consistency and structural validity.

Key words: Couple violence, suffered violence, exerted violence, psychological damage, psychometrics.

Introduction

Partner violence can be defined as an exercise of power, in which, through

of actions or omissions, it harms or controls against the will of that person with whom
there is an intimate bond, whether it be courtship (a romantic relationship maintained between two)
people with or without the intention of getting married and not cohabitating), marriage (cohabitation relationship)
and mutual support between two legally established individuals that often implies a bond

109
 Daena: International Journal of Good Conscience

romantic cohabitation (romantic relationship between two people who live together or not)
intention to marry and that may have a legal recognition distinct from marriage
(Moral and López, 2014).

In Mexico, four types of partner violence are distinguished by modality.

harm: the physical violence directed at the body to cause pain or even death,
considering everything related to sexuality as a separate category; violence
psychological that is aimed at self-esteem, self-concept, and emotional balance;
economic violence aimed at controlling income, expenses, and assets; and violence
sexual that aims to impose or restrain sexual relationships, or to undermine self-esteem
and sexual integrity (Secretariat of the Interior, 1999).

There are scales to assess violence in couples (Health Observatory of the

Woman, 2005). In Mexico, there is the Partner Violence Severity Index (PVSI) of
Valdez et al. (2006), which evaluates the intensity of violence exercised by the partner and is applied to
women; the Domestic Violence Inventory: Frequency and Perception (VIDOFyP) of
Trujano et al. (2006) that evaluates the frequency and perception of violence in a series of
couples' behaviors or situations, initially developed for women, but
subsequently it has also been applied to men (Trujano et al., 2010); the Questionnaire of
Violence in the Couple (CVP) by Cienfuegos and Díaz-Loving (2010), which evaluates violence
suffered and exercised in the situation of living as a couple, they are answered in terms of frequency and
it is applicable to both sexes and various couple situations (marriage, cohabitation or
dating); and the Women's Abuse Questionnaire (CAM) by Linares et al. (2013), which
evaluate the presence of physical, psychological, or sexual violence against women.

Statistics emphasize the violence inflicted against women; without

It is recognized that the man also suffers violence from his female partner.
Research by authors such as Archer (2002), Straus (2007), Hattery (2009), and Fiebert
(2013) have made it clear that violence is not a unidirectional phenomenon and that violence
is exercised within the couple usually by both members. Picking this up
identification, a good instrument for partner violence should consider both the
violence exerted as suffered in the forms of physical, psychological, social violence
and sexual, evaluate not only the frequency but also the intensity, to be applicable to both
genders and various couple situations (marriage, cohabitation, or dating). Currently
there is no instrument with such characteristics (Nicholls et al., 2013; Observatory of
Women’s Health, 2005); the one that comes closest is the CVP of Cienfuegos and Díaz-Loving
(2010), but does not evaluate intensity or damage. Thus, it was decided to modify the CVP and request from the
participants who will evaluate each of their 38 items in terms of frequency and damage.
Resuming the suggestion of the study of the psychometric properties of the Moral CVP
López (2014) added a second item of exercised sexual violence, due to the problems
of internal consistency and communality of the item of sexual violence already present in the
questionnaire, with the CVP now consisting of 39 items.

Considering this background, the objectives of this research were:

1) modify the CVP to evaluate violence suffered and exerted in frequency and damage and
2) study its internal consistency and factorial structure. Values are expected to
internal consistency of the 2 scales of the questionnaire and their factors vary from acceptable to

110
 Daena: International Journal of Good Conscience

high; that the factorial structure of the 27 items of violence suffered, both in terms of frequency
like the damage, it is of 4 correlated factors: psychological/social, physical, sexual and
economic; and the factorial structure of the 12 items of violence exerted, both those of
frequency like those of damage, be it from 2 correlated factors: psychological violence and
another type of violence (Cienfuegos and Díaz-Loving, 2010; Moral and López, 2014).

Methodology

Participants

A non-probabilistic incidental sample of 240 participants was used.

volunteers from the city of Monterrey, Nuevo León, Mexico. Inclusion criteria are
required: to know how to read and write, to be over 18 years old, to have a heterosexual romantic relationship
and giving informed consent. The exclusion criteria were: questionnaire
incomplete and the pollster's impression that the participant responded in a certain way
unintentional or arbitrary.

The questionnaire was administered individually by an interviewer in households.

pedestrian streets and public parks. Participation was unpaid. The questionnaire was
preceded by a sheet in which consent was requested for participation in the
study, the objectives of the research, those responsible for it, and its
institutional affiliation; likewise, anonymity and confidentiality were guaranteed for the
information given, following the ethical standards of the Mexican Society of Psychology (2007).

Half of the participants were men and the other half women. It was alternated between
a man and a woman in the app to get the same number of participants
both sexes. The percentage of participation, giving consent and responding to the
The complete questionnaire was 82% (240 out of 293). In most cases where not
consent was given and participation was not desired, the reason given was lack of time.

The average age in the sample was 35.41 years (SD= 9.63) with a minimum of 18 years and
maximum of 57 years. Regarding education, 23 of the 240 participants (9.6%)
They indicated having primary education, 58 (24.2%) secondary education, 47 (19.6%) higher secondary education.
general, 39 (16%) technical media and 73 (30.4%) higher education. Regarding
In terms of labor activity, 45 out of the 240 participants (18.8%) reported being engaged in work related to
home, 67 (27.9%) work as office or sales employees, 54 (22.5%) work as
manual workers, 45 (18.8%) work as professionals, 14 (5.8%) have
own business, 8 (3.3%) being unemployed and 7 (2.9%) dedicating to studying. Regarding
to the self-defined socioeconomic status, 54 of the 118 participants (22.5%) defined themselves as
estatus socioeconómico bajo, 62 (25.8%) medio-bajo, 91 (37.9%) medio, 30 (12.5%) medio-
high and 3 (1.3%) high. Regarding marital status, 118 out of 240 participants (49.2%)
reported being married, 22 (9.2%) single, 44 (18.3%) divorced or separated, and 56
(23.3%) in common-law partnership. Of the 240 participants, 191 (79.6%) reported having children and 49 (20.4%)
not having. The average number of children was 2.31 (SD= 1.06), varying from 1 to 5 among those who
they did have children.

111
 Daena: International Journal of Good Conscience

Instrument

Partner Violence Suffered and Exercised Questionnaire (CVSEP). It was designed to

this study is derived from the CVP of Cienfuegos and Díaz-Loving (2010). It allows for the evaluation of
violence suffered and exercised in the couple situation in terms of frequency and harm. It is
applicable to both sexes. The questionnaire consists of four scales. Two scales
they assess the violence suffered by the partner through the same 27 direct items they have
a 5-point Likert response format. The 27 items are answered on a scale.
from 1 'never' to 5 'always' and in the other scale the same 27 items
they are answered in terms of damage suffered (from 1 'nothing' to 5 'a lot'). Regarding the
The content of these 27 items of suffered violence, 8 items are oriented to assess violence.
psychological and social (items 6, 9, 15, 17, 18, 19, 21, and 22), 7 items to assess physical violence,
intimidation and aggression (items 2, 3, 11, 12, 13, 16, and 25), 7 items to assess sexual violence
(items 1, 4, 5, 7, 8, 14, and 27) and 5 items to assess economic violence (items 10, 20, 23, 24, and
The other two scales evaluate violence exerted against the partner through the 12
same direct items that have a 5-point Likert response format. A scale
It is responded to in terms of frequency (from 1 'never' to 5 'always') and another scale is responded to.
in terms of damage inflicted (from 1 'nothing' to 5 'a lot'). Regarding the content of
These 12 items of exercised violence, 5 items are aimed at assessing violence.
psychological (items 28 to 32) and 7 items to evaluate other types of violence, such as social,
economic, physical and sexual (items 33 to 39). See the CVSEP in the Annex.
A sexual violence item was added (item 39 "I have told my partner that it is
ugly or not very attractive that was not in the CVP to complement or replace the one already
I had the questionnaire (item 34 'I do not take into account the sexual needs of my partner'),
due to the fact that this item showed low internal consistency. The wording of this item was made
in correspondence with the item that showed more consistency and factorial weight within the
frequency factor of sexual violence suffered, which was item 7 "my partner has said that I am
ugly or not very attractive” (Cienfuegos and Díaz-Loving, 2010; Moral and López, 2014).

Statistical analysis

The internal consistency was calculated using Cronbach's alpha coefficient. Some values
α ≥ .90 was considered excellent, from .80 to .89 good, from .70 to .79 acceptable, from .60 to
.69 questionable and from .50 to .59 poor and < .50 unacceptable (Cronbach and Shavelson, 2004).
The linear association was calculated using the product-moment correlation coefficient.
Pearson; values below .10 were considered trivial associations, of 10
.29 low, .30 to .49 moderate, .50 to .69 high, and .70 to .89 very high and ≥ .90 unitary
(Ellis, 2010).

The factorial structure was determined both by exploratory factor analysis, using
Principal Components and Oblimin rotation, as per confirmatory factor analysis,
using Scale-Free Least Squares (SLS). The number of factors was determined by
the Horn criterion (95th percentile) and the expectation. The SLS method was chosen due to the
non-compliance with multivariate normality and the ordinal level of measurement of the variables
manifestations (Byrne, 2010).

112
 Daena: International Journal of Good Conscience

Eight adjustment indices were considered for the confirmatory factor analysis: test
chi-square (χ)2the quotient of the chi-square statistic and its degrees of freedom (χ2/gl),
Bollen-Stine test (B-S), goodness of fit index (GFI) of Jöreskog and Sörbom and its
corrected modality (AGFI), Bentler-Bonett normalized fit index (NFI), index
relative adjustment by the Bollen rho coefficient (RFI) and the standardized squared residual
medium (SRMR). Good fit values for the indices were stipulated: pde χ2ypde
B-S > .05, χ2/gl ≤ 2, GFI ≥ .95, AGFI, NFI and RFI ≥ .90 and SRMR ≤ .05; and as values
adequate: pde χ2> .01, χ2/gl ≤ 3, SRMR ≤ .08, GFI ≥ .85 and AGFI, NFI and RFI ≥
The significance of the parameters was contrasted by the free percentile method.
bias, extracting 2,000 samples. The same number of samples was extracted for the test
the Bollen-Stine. The properties of factorial invariance between both sexes were studied by
multi-group contrast, specifying nested models in cumulative constraints. It
he considered that invariance is acceptable if the goodness of fit is good or acceptable and
statistically equivalent or next, at least, in the models without constraints and with
constraints in weight measurements (Byrne, 2010).

To define the factorial structure, significant measurement weights were required and
saturations ≥ .30, internal consistency of the factors at least acceptable (α ≥ .70),
correlations between non-unit factors and < .80, and properties of invariance between both
acceptable sexes.

Results

Internal consistency and factorial structure of the 27 items of violence frequency

suffered

The internal consistency of the 27 items of frequency of violence experienced was

excellent (α = .95). Based on Horn's parallel analysis, 2 components were extracted.
that explained 52.3% of the total variance. After the oblique rotation, the first component
was defined by 17 items related to economic violence (items 10, 20, 23, 24, and 26),
psychological/social (items 6, 9, 15, 17, 18, 19, 21, and 22), intimidation/verbal aggression (items
12, 13, and 25) and sexual (item 14) and had excellent internal consistency (α = .94). The second
The component was configured with 10 items on sexual violence (items 1, 4, 5, 7, 8, and 27)
and physics (items 2, 3, 11, and 16) and had high internal consistency (α = .88). The correlation of
both components were moderate (r= .49, p< .01).

By extracting 4 components based on expectation, 62.7% of the total variance was explained.
After the oblique rotation, the first component was configured by 6 items on violence.
physical/intimidation/aggression (items 2, 11, 13, 16, and 25) and sexual violence (item 14) and had
high internal consistency (α = .89). The second component was defined by 6 items about
sexual violence (items 1, 4, 5, 7, 8, and 27) and had high internal consistency (α = .80).
The third component consisted of 9 items on economic violence (items 10, 20, 23,
24 and 26) and psychological violence related to role performance (items 15, 18, and 19) and
Intimidation/blackmail (item 12) had excellent internal consistency (α = .91). The fourth
component was made up of 6 items on psychological/social violence (items 6, 9, 17,
21 and 22) and intimidation (item 3) and its internal consistency was high (α = .85). The correlations

113
 Daena: International Journal of Good Conscience

Among the 4 components, they varied from .28 to .51 with a mean of .36 (SD=
.08).

Confirmatory factor analysis was used to test the expected model of 4 factors.
correlated (4FC_Spa.), apart from 2 models derived from exploratory factor analysis:
of 4 and 2 correlated factors (4FC_Obs. and 2FC_Obs., respectively). Upon observing
very high correlations between the factors (from .68 to .90 in the expected 4-factor model,
from .65 to .85 in the observed of 4 factors and .82 in the observed of two factors) and when considering
the very high value of consistency among the 27 items also specified a model of a
a factor (1F) and two models of 4 hierarchically organized factors to a general one (4FJ_Sp. and 4FJ_Obs.).

According to Moral and López (2014), the expected model of 4 factors, whether they are
correlated (4FC_Spa.) or hierarchized (4FJ_Spa.), was defined by violence
psychological/social with 8 indicators (items 6, 9, 15, 17, 18, 19, 21, and 22), violence
physical/intimidation/aggression with indicators with 7 indicators (items 2, 3, 11, 12, 13, 16 and
25), sexual violence with 7 indicators (items 1, 4, 5, 7, 8, 14, and 27) and economic violence
with 5 indicators (items 10, 20, 23, 24, and 26). The internal consistency of the 4 factors varied
from acceptable (α = .71) to excellent (α = .90).

The standardized value of Mardia's multivariate kurtosis for the 27 items was
83.73, which evidenced deviation from multivariate normality. The 6 models
specified had all their significant parameters by the free percentiles method
of bias. The adjustment indices were good, although the goodness of fit was not maintained by
the probability of Bollen-Stine, nor the probability of the chi-square statistic. The
the expected model of 4 correlated factors (4FC-Eng.) was the one that showed the best
adjustment indices (see Table 1).

However, the correlations between the 4 expected factors were very high and the
The correlation between physical and sexual violence was one (.90). In the 4-factor model.
the hierarchical weight of the general factor was unitary on physical violence (.94) and
psychological/social (.90). These correlations or unit weights dismiss these models
expected. Furthermore, in the sample of men, sexual violence had internal consistency.
poor (α = .55) and questionable economic violence (α = .67). Also the model
four observed factors had a unitary correlation and two factors did not reach it
acceptable internal consistency in men.

In the model of two correlated factors, internal consistency varied from

acceptable (α = .72 in physical/sexual violence) to excellent (.88 in other types of violence) in
men and was excellent in women (.91 and .95, respectively). When contrasting the invariance
between both sexes, the problem of the lack of significance of the weight measurement persisted
factor of sexual/physical violence regarding item 8 in men, as in the models of 4
factors, with all other parameters being significant in both samples. The goodness of
the unconstrained model adjustment was rejected by the chi-square statistic and the
Bollen-Stine probability, but 5 indices reflected a good fit and RMSEA showed a value
of proper adjustment. The adjustment worsened as constraints were imposed on the weight measurements and
the fit was poor in the models with additional constraints on the variances-covariances
structural and in the residues. The correlation of the two factors was very high, at .84 in
114
 Daena: International Journal of Good Conscience

women and .70 in men. In women the residues were lower, resulting in the model
more defined (see Table 1).

Table 1
Adjustment indices for factorial models of frequency of suffered violence
Uni-group (joint sample) Multi-group among both sexes
IA* 2CO
4CE 4JE 4CO 4JO 2CO 1
SC PM CE RM
χ2 390.7 440.7 415.9 437.6 505.7 616.9 833.3 1564.1 6549.8 7067.8
p ** ** ** ** ** ** ** ** ** **
χ2/gl 1.22 1.38 1.31 1.37 1.56 1.90 1.29 2.33 9.72 10.1
B-S ** ** ** ** ** ** ** ** ** **
SRMR .07 .07 .07 .07 .08 .08 .08 .10 .34 .21
GFI .98 .98 .98 .98 .97 .97 .95 .91 .61 .59
AGFI .97 .97 .97 .97 .97 .96 .94 .90 .57 .55
NFI 0.98 .97 .97 .97 .97 .96 .94 .89 .52 .49
RFI .97 .97 0.97 .97 .97 .96 .93 .88 .50 .49
**p< .01. IA = adjustment indices. Models: Number of factors: 4, 2 or 1. C = correlated factors, J = factors
hierarchized to a general factor, E = expected model, O = observed model from the component analysis
main. SC = without constraints, PM = constraints in measurement weights, CE = in structural variance-covariances
and RM = in measurement residues.

Internal consistency and factorial structure of the 27 items of damage suffered

The internal consistency of the 27 items of suffered damage was excellent (α = .95). With
Based on Horn's parallel analysis, 3 components were extracted that explained 59.5% of
the total variance. After the oblique rotation, the first component was configured by 11
ítems sobre daño por violencia económica (ítems 10, 20, 23, 24 y 26), psicológica-social
(items 15, 17, 18, 19, and 21) and verbal aggression (item 25) and had excellent internal consistency
(α = .93). The second component was defined by 8 indicators on damage from violence.
sexual (items 1, 4, 5, 7, 8, and 27) and intimidation/blackmail (items 3 and 12) and was consistent
high internal (α = .88). The third component was made up of 8 indicators on damage caused by
physical violence (items 2, 11, 13, 16, and 22), psychological-social violence motivated by jealousy
(items 6 and 9) and sexual (item 14) and had high internal consistency (α = .88). The 3 components
they correlated with each other with values ranging from .38 to .51.

By extracting 4 components by expectation, 64.4% of the total variance was explained.

After the oblique rotation, the first component was configured with 11 items on damage.
for economic violence (items 10, 20, 23, 24, and 26), psychological-social (items 15, 17, 18, 19)
and verbal aggression (item 25) and had excellent internal consistency (α = .93). The second
it was defined by 8 items on damage from sexual violence (items 1, 4, 5, 7, 8, and 27) and
intimidation/blackmail (items 3 and 12) and had high internal consistency (α = .88). The third
it was composed of 5 items regarding damage from physical violence (items 2, 11, 13, and 16) and sexual
(item 14) and had high internal consistency (α = .86). The fourth was made up of 3
indicators on psychological-social violence motivated by jealousy (items 6, 9 and 22) and had
high internal consistency (α = .83). The 4 components correlated positively with each other
yes with values that ranged from .23 to .52.

115
 Daena: International Journal of Good Conscience

A confirmatory factor analysis tested the expected model of 4 factors.

correlated (4FC-Sp.), apart from 2 models derived from component analysis
mainly, of 4 and 3 correlated factors (4FC-Obs and 3FC-Obs, respectively). The
the standardized value of Mardia's multivariate kurtosis of the 27 items was 75.80, which
which evidenced a deviation from multivariate normality. The fit indices were good,
although the 3 models were rejected based on the chi-square test and the Bollen probability
Stine and the standardized mean squared residual reflected a good fit in all 3 cases.
The model with the best fit was the observed one with 4 correlated factors (4FC-Obs.) with
the advantage not only of having all its significant parameters, but also lesser correlations
that .80 among its factors when it was not the case in the other models (see Table 2).

When contrasting the invariance between both sexes of the observed model of 4
correlated factors (4FC_Obs.), all parameters were significant in both
samples by the unbiassed percentile method. The correlations between the factors
were less than .80 in both samples. The models without constraints and with
constriction in the weights of measurement showed good fit to the data, except for the
chi-square statistics and the Bollen-Stine repetitive sampling probability. No
however, the adjustment indices were poor as additional constraints were imposed on the
structural variances-covariances and in the measurement residuals (see Table 2). In women
the variances of the factors were greater and the residues were smaller than in men, that is,
the model proved to be more defined in women. The values of internal consistency of the 4
Factors ranged from .84 to .94 in women and from .73 to .91 in men.

Table 2
Adjustment indices for factorial models of damage suffered
Multi-group between both sexes
Uni-group
IA* (4FC_Obs)
4FC_Sp. 4FC_Obs 3FC_Obs SC PM CE RM
χ 2 469.1 406.4 453 664 1150.7 3757 4064
p ** ** ** ** ** ** **
2
χ /gl 1.48 1.28 1.41 1.04 1.75 5.62 10.03
B-S ** ** ** ** ** ** **
SRMS .07 .07 .07 .07 .09 .25 .13
GFI .98 .98 .98 .96 .94 .80 .78
AGFI .97 .98 .98 .96 .93 .77 .76
NFI .97 .98 .98 .96 .93 .76 .74
RFI .97 .97 .97 0.95 .92 .74 .73
Notes. **p< .01. IA = adjustment indices. Models: Number of factors: 4 or 3. C = correlated factors,
expected model
measure, CE = in structural variance-covariances and RM = in measurement residuals.

116
 Daena: International Journal of Good Conscience

Internal consistency and factorial structure of the 12 items of violence frequency

exercised

The internal consistency of the 12 items of frequency of violence exercised was high (α
If item 34 were to be eliminated, it would increase to .85. The correlation of item 34 with the
the rest of the scale was .23. Based on Horn's parallel analysis, 1 component was extracted
which explained 38.70% of the total variance. All saturations were ≥ .30.

By extracting 2 components based on expectation, 48.8% of the total variance was explained.
After the oblique rotation, the first one was configured by the 5 items on violence.
psychological (items 28, 29, 30, 31, and 32) and two about aggression (items 33 and 36) and had
high internal consistency (α = .86). The second was composed of 5 items about violence.
economic (items 37 and 38), social (item 35) and sexual (items 34 and 39) and had consistency
internal consistency (α = .55), which would increase if item 34 were removed (α = .57).
The correlation between components was moderate (r= .44, p< .01).

Three models were contrasted by confirmatory factor analysis. Two models were
of two correlated factors. A bifactor model was specified from the expectation,
according to Moral and López (2014): psychological violence with 5 indicators (items from 28
to 32) and another type of violence different from psychological with 7 indicators (items from 33 to
39). In the present sample, the first factor showed high internal consistency (α = .85) and
the second questionable (α = .66). The other bifactor model was specified from the result
from exploratory factor analysis; in this second model items 33 and 36 were
indicators of the first factor of psychological violence-aggression and the second factor of another
the type of violence remained with 5 indicators (2FC_Obs.). The third model was of a factor
general, supported by the parallel analysis of Horn (1F_Obs).

The standardized value of Mardia's multivariate kurtosis of the 12 items was

41.47, which reflects deviation from multivariate normality. In all three models, all the
parameters were significant by the bias-free percentiles method, including the
weight of the factor of another type of violence or the general factor on item 34. The goodness of
the adjustment of the 2 correlated factor models was greater compared to that of
unifactorial model (Δχ2= 8.86, p< .01 for expected model and Δχ2= 18.69, p< .01 for the
observed model). Due to the Bollen-Stine probability, the goodness of fit was not maintained.
with a p-value > .05 in none of the three models; only the observed model of 2 factors
correlated had a p>.01. In this last model (2FC_Obs.), the remaining indices of
The contemplated adjustments were good, including the chi-square statistic (p = .18). In the
in two other cases, the adjustment indices vary from good (χ2/gl, RMR, GFI, AGFI, NFI and RFI)
a suitable (χ2) (see Table 3).

When contrasting the invariance of the observed model of two correlated factors
between men and women (2FC_Obs.), all parameters were significant by the
bias-free percentile method, except for the weight of the factor of another type of violence on
item 8 in women. The correlation between the two factors was .60 in women and .75 in
men. The unconstrained model showed good fit values across all indices.
The goodness of fit of the model with constraints on the weights remained with a p> .01
by the chi-square test and Bollen-Stine and the remaining indices showed values of
117
 Daena: International Journal of Good Conscience

adjustment. In models with additional constraints on the variances-covariances

structural and average residual values were poorly adjusted (see Table 3). In
Men had lower waste than women, that is, the model was more defined.
in men. It should be noted that, in the expected model of 2 correlated factors
two indicators of another type of violence different from psychological violence not
were significant in the sample of women (items 34 and 35), the correlations between the two
factors were 0.84 in both sexes and the indices were worse than those of the model
2FC_Obs. In the unidimensional model, neither the measurement weights on items 7 and 8
they were significant.

However, the factors of the 2FC_Obs model had consistency issues.

internal in the sample of women. The factor of psychological violence/aggression had
excellent internal consistency in men (α = .92) and questionable in women (α = .66). The
Another type of violence factor had acceptable internal consistency in men (α = .70), but
unacceptable in women (α = .40). The expected factor of another type of violence also had
problems of internal consistency in women.

Only the unidimensional model lacked internal consistency problems, which

would be acceptable in women (α = .70) and good in men (α = .89). The unidimensional model
Without items 34 and 35 (1F-10), it would have all its significant parameters in both samples.
and in the combined sample, good internal consistency in the combined sample (α = .86), excellent in
men (α = .91) and acceptable in women (α = .72), a goodness of fit statistically
equivalent to the model 2FC_Obs in the joint sample (Δχ2[18,N= 240] = 8.3,p= .97) and
some invariance properties between both sexes better than the 2FC_Obs model (see
Table 3).

Table 3
Adjustment indices for factorial models of the frequency of violence exerted
Uni-group Multi-group between both sexes
2FC- 2FC- 2FC-Obs 1F-10
IA* 1F-12 1F-10
Esp Obs SC PM CE RM SC PM CE RM
χ2 71.9 62.2 80.9 53.9 93 146 665 750 74 119 586 669
P .04 .18 .01 .02 .82 .03 ** ** .34 ** ** **
χ2/gl 1.36 1.17 1.50 1.54 0.9 1.3 5.6 5.7 1.1 1.5 7.3 7.4
B-S ** >.01 ** .01 .08 .03 ** ** .04 0.02 ** **
SRMR .06 .06 .07 .06 .07 .10 .15 .14 .08 .11 .17 .15
GFI .98 .98 .97 .98 .97 .96 .81 .78 .98 .96 .80 .78
AGFI .97 .97 .96 .97 .96 .94 0.75 .74 .96 .94 .73 .73
NFI .96 .97 .96 .97 .95 .92 .67 .62 .96 .93 .67 .63
RFI 0.95 .96 .95 .96 .93 .91 .63 .62 .94 .92 .63 .63
Notes. **p< .01. IA = adjustment indices. Models: Number of factors: 2 or 1. C = correlated factors,
expected model
of measurement, CE = in structural variances-covariances and RM = in measurement residues.

Internal consistency and factorial structure of the 12 items of damage caused

The internal consistency of the 12 items of damage caused was high (α = .89).
The correlation of item 34 with the rest of the scale was .38. Based on the parallel analysis.

118
 Daena: International Journal of Good Conscience

From the Horn, 1 component was extracted that explained 46.5% of the total variance. The saturations
were ≥ .45. Item 34 was the lowest.

By extracting 2 components based on expectation, 57.3% of the total variance was explained.
After the oblique rotation, the first was configured with 6 items about damage caused to
the couple for another type of violence different from psychological violence (items 33, 34, 35, 37, 38, and 39) and
it had high internal consistency (α = .79). The second consisted of 6 items about damage
caused to the couple by psychological violence (items 28, 29, 30, 31, and 32) and physical violence
(item 36) and had high internal consistency (α = .87). The correlation between components was high
(r= .55, p< .01).

Three models were tested through confirmatory factor analysis. Two models were
of two correlated factors. A bifactor model was specified from the expectation
(Moral and López, 2014): psychological violence with 5 indicators (items 28 to 32) and another
type of violence different from psychological with 7 indicators (items 33 to 39) (2FC-Sp.)
In the present sample, the first factor showed high internal consistency (α = .88) and
also the second (α = .81). The other bifactor model was specified from the result of
exploratory analysis (2FC-Obs.). The third model was of one factor, supported by the analysis.
Horn parallel (1F_Obs).

The standardized value of Mardia's multivariate kurtosis of the 12 items was

24.99, which reflects deviation from multivariate normality. All parameters of the three.
models were significant by the bias-free percentile method. The model
expected two-factor model (2FC-ESP) was the one that showed the best fit; all its indices
they reflected a good fit, except for SRMR which showed an adequate fit (.06) and the probability
The Bollen-Stine was less than .01. The correlation between the two factors was high (r= .76).
but lower than .80 (see Table 4).

When contrasting the invariance of the expected model of two correlated factors between
men and women (2FC_Esp.), all parameters were significant by the method of
bias-free percentiles. The correlation between the two factors was .71 in women and .76
in men. The unconstrained model showed good fit values in all the
indices, good according to SRMR and bad according to the Bollen-Stine probability. The goodness
The model adjustment with constraints on the weights was poor for 3 indices and good for the rest.
5 remaining. In models with additional constraints on the variances-covariances
structural and measurement residues, the fit was worse (see Table 4). In men, there were
greater the variances of the factors and lower the residues than in women, that is,
the model was more defined in men. The internal consistency of the damage factor
caused by psychological violence was excellent in men (α = .92) and good in women
(α = .81). The internal consistency of the damage factor from other types of violence was good.
men (α = .85) and acceptable in women (α = .72).

119
 Daena: International Journal of Good Conscience

Table 4
Adjustment indices for factor models of damage caused by exercised violence
Multigroup between both sexes
Uni-group
IA* (2FC_Spa.)
2FC_Sp. 2FC_Obs. 1F_Obs. SC PM CE RM
χ 2 64.39 67.24 115.40 97.99 219.28 544.65 625.66
P .14 .09 < .01 .70 < .01 < .01 < .01
2
χ /gl 1.21 1.27 2.14 0.92 1.89 4.58 4.78
B-S < .01 < .01 < .01 < .01 < .01 < .01 < .01
SRMR .06 .06 0.08 .06 .11 .12 .10
GFI .99 .98 .97 .98 .95 .87 .85
AGFI .98 .98 .96 .97 .93 .83 .82
NFI .98 .98 .96 .96 .92 .81 .78
RFI .97 .97 0.95 .96 .91 .78 .77
Notas. *IA = índices de ajuste. Modelos: Número de factores: 2 ó 1. C = factores correlacionados, Esp =
expected model, Obs = observed model. SC = without constraints, PM = constraints in weights of
measure, CE = in structural variance-covariances and RM = in measurement residues.

Dimensions of violence suffered and exercised

By simple summation of items, the scores in frequency of violence were created.

sufrida (FVS-27)
(FVE-12) and damage due to violence exercised (DVE-12). The correlations between damage and frequency.
the violence suffered or exercised was high in the three samples, when the correlations
Crosses were moderate in the joint sample, from moderate to low in women and high in
men (see Table 5).

Table 5
Correlations between suffered and inflicted violence (frequency and damage)
Vio- Joint Women Men
lencia FVS DVS FVE FVS DVS FVE FVS DVS FVE
DVS .92*** .94*** .87 ***

FVE .34*** .42*** .24** **

.34 * .73*** .70***
DVE .31*** .44*** .79*** .22* .33 .70 .72 .78*** .83***
*** *** ***

Notes. N= 240, ***p< .001. FVS = sum of the 27 items of frequency of violence endured, DVS = sum
of the 27 items of harm suffered from partner violence, FVE = sum of 12 items of frequency of violence
exercised against the partner, DVE = sum of the 12 items of damage caused by violence exercised against the
couple.

Based on Horn's parallel analysis, 2 components were extracted that explained 93.1%.
of the total variance. After the oblique rotation of the components matrix, the first was
suffered violence (frequency and damage); and the second component was violence exerted
(damage and frequency). The correlation between both components was moderate (r= .40, p<
.01). In women, the number of components was also 2 according to Horn's parallel analysis. To
extracting these two components explained 91.1% of the total variance. After rotation

120
 Daena: International Journal of Good Conscience

oblique, also the first component was suffered violence and the second of violence
exerted. The correlation between both components was moderate (r= .30, p< .01). In
men the number of components was 1 by Horn's parallel analysis: The single component
it explained 82.8% of the total variance (see Table 6)

Table 6
Configuration matrices
Sub-scales Joint Women Men
of violence C1 C2 C1 C2 C1
Frequency of violence suffered .99 -.06 .99 -.06 .91
Damage suffered .95 0.08 .96 .07 .92
Damage caused -.01 .95 -0.01 .92 .91
Frequency of violence exercised .01 .94 .01 .92 .89
Note: Extraction method: Principal components. Rotation: Oblimin.

Discussion

This study aimed to examine the psychometric properties.

of internal consistency and factorial structure of the modified CVP to allow for evaluation
the frequency and the damage in violence suffered and exerted.

Regarding the new item introduced in the CVSEP, it can be stated that the result
a good indicator of frequency and harm from other types of violence exercised different from the
psychological. Item 34, which was intended to be replaced, was the weakest item in both scales.
of exercised violence. It should be noted that item 7, which was used as a model for drafting
the new item turned out to be one of the best indicators of frequency and damage suffered by
sexual violence, as in previous studies (Cienfuegos and Díaz-Loving, 2010; Moral and López,
2014).

Regarding the frequency scale of experienced violence (FVS), the consistency

The internal consistency among the 27 items was very high suggesting unidimensionality (Streiner, 2003).
4 expected factors were not reproduced exactly in the exploratory factor analysis.
although the configuration of the rotated component matrix did provide a result
next in its interpretation, being the components of economic violence and
psychological/social those with more overlapping items. In the factor analysis
confirmatory, the expected model of 4 correlated factors was the one that showed the best
adjustment, even in comparison with a specification derived from exploratory analysis; not
However, the model is forced by the very high correlations between the 4 factors. Throughout
In this case, a model of 4 hierarchical factors would need to be considered in relation to a general one, being the
Factors of physical and psychological/social violence fully explained by this general factor.
of frequency of violence and either of these two factors could represent the entire
scale.

In the tetrafactorial model, the correlations were very high especially between
physical and sexual violence and between psychological/social and economic violence. Precisely, for
the parallel analysis of Horn, the number of substantive factors not attributable to errors of
sampling and method were 2, resulting in a factor of sexual/physical violence and another type of
121
 Daena: International Journal of Good Conscience

violence. These two factors did have high internal consistency in both sexes when not
it was so in the 4-factor model. Except for the global chi-square fit indices and the
Bollen-Stine probability that reflected poor fit in all models, the remaining 6
adjustment indices were virtually equivalent between the 4-factor and 2-factor models, 5 of
they reflected a good fit and one showed an adequate fit. Also the differences in the
The invariance properties of the 4-factor and 2-factor models were minimal.

Summarizing the arguments made up to this point, the two-factor model of violence
physical-sexual and other types of violence is more substantive and appropriate for the 27 items of
frequency of violence suffered. The FVS scale could be reduced to the frequency factor of
physical or psychological violence. One might ask why sexual and physical violence
they are grouped into the same factor compared to other forms of violence. As has already been mentioned.
the correlations between these two factors were very high in the 4-factor model
in the joint sample as in the samples of both sexes; and in men it reached the value
unitary. It is mainly men who report suffering from a very high
correspondence between sexual violence (sexual rejection) and physical violence (assaults). This
combination, which is also present in women, could reflect an evolution towards
a worse dynamic or a worse state within marital conflicts. From this
interpretation, the dominance of the factor of psychological, economic, and social violence over the
The factor of physical-sexual violence would represent a state of lesser deterioration of the relationship or
the marital conflict. Against this interpretation, the average of physical/sexual violence was
statistically equivalent to the mean of another type of violence in both sexes and in the
joint sample. Furthermore, if there were two sequential stages, the correlation should be
smaller. Such high correlations reflect that they are rather simultaneous aspects of the
violence, that is, they coexist at the moment when the report is made. It can be stated
that the self-report of the frequency with which violence is received from the partner is
little nuanced in the physical, psychological/social, economic, and sexual aspects by the
data on internal consistency and factorial structure.

Regarding the damage scale suffered (DVS), despite the internal consistency
it was very high among its 27 items, possibly indicating unidimensionality (Streiner, 2003), it
achieved a model with factors with more unit variance than with the 27 frequency items
of suffered violence. The expected model was not reproduced exactly. A factor was set up
of damage from economic/social violence, another factor of damage from violence
sexual/intimidation/blackmail, a third factor of harm from physical violence and a fourth factor
of damage due to social violence motivated by jealousy. The model of 4 correlated factors.
derived from the exploratory factor analysis showed the best fit indices to the data and
the best properties of invariance between sexes; also the correlations between their
factors were less than .80, whereas in the other models they reached higher values
that .80, including the 3-factor model, suggested by Horn's parallel analysis. These 4
factors had high internal consistency in both men and women, from which it is considered that
it is the factorial model that best represents the present data. However, it should be emphasized
that the very high internal consistency and the high correlations between the factors (from .68 to .77
in the joint sample) highlight the significance of the total score and the blurred nature of the
different facets of violence or harm suffered.
Regarding the scale of frequency of exercised violence (FVE), item 34 of
sexual violence showed low internal consistency, as already observed in Moral's study.
122
 Daena: International Journal of Good Conscience

and López (2014). Initially, it was retained and in the exploratory factor analysis it did not have
factorial definition problems; it remained in the expected factor of another type of violence
different from the psychological/verbal with a load greater than .50. Item 33 of verbal aggression
("when he doesn't attend to the children the way I believe, I verbally attack him") and item 36 of
physical violence ("I have ended up physically hurting my partner") did not have problems with
internal consistency, but rather of factorial definition; remained in the unexpected factor of
psychological violence, which was renamed psychological violence/aggression. Although
the observed model (specified from the exploratory factor analysis) had good fit
to the data and better properties of invariance between the sexes than the expected model and the
from a factor with 12 indicators, the factor of another type of violence exercised had consistency
unacceptable internal and a non-significant indicator in the sample of women. The model of
a factor with 10 indicators (excluding items 34 and 35) would be the best option based on the
parallel Horn analysis (a substantive factor), a good adjustment in the joint sample and
equivalent to the observed two-factor correlated model, better properties of
invariance between both sexes that the observed model of two correlated factors,
internal consistency from acceptable to excellent and all its significant parameters in the
joint sample, of women and men. Again, the nuances about types of
violence exercised against the partner is poorly differentiated.

Regarding the scale of damage due to violence exercised (DVE), item 34 of violence
sexual was the one that showed the least internal consistency and the lowest weight in its factor, but these
Values are finally acceptable and their elimination is not recommended. Repeatedly this
the item is the weakest. It could be that ignoring the sexual needs of the partner is not
motivated not exclusively by rejection or passive aggression, but by lack of ability or desire
sexual, hence it is a weak indicator of violence exercised.

The exploratory factor analysis almost reproduces the expected model with the 12 items.
from the DVS, except for item 36. Item 36 ('I have physically harmed my partner')
in the exploratory analysis, it saturated with very similar loadings in both factors and the analysis
the internal consistency showed that it contributes to both factors. This problem appeared
with item 36 on the frequency scale of violence exercised, although with a greater difference in
loads between both factors. It could be that item 36 is interpreted in terms of possible
damage to health, that is, physical damage understood in a broad sense and independent of type
of violence (jealousy, verbal, psychological, physical), hence its ambiguity within the structure
factorial.

The expected model of two correlated factors is the best option for the 12.
items from the DVS scale. It fits the data well in the combined sample, better
invariance properties between both sexes that the other models, the correlations between
the factors are less than .80 in the joint and for both sexes, the factors have
internal consistency from acceptable to excellent and all its significant indicators in the
joint sample, of men and women. As in the model of the 12 items of FVE, the
Horn's parallel analysis indicated a unidimensional model and high correlations between the
Two factors were high (> .70). Although the one-factor model was the one with the worst fit,
it should be given importance for these two arguments, along with the high consistency of
set of 12 items in the three samples, especially in the men's one. Again the

123
 Daena: International Journal of Good Conscience

nuances about the types of violence exerted against a partner are poorly differentiated, such as
in the report of the damage inflicted on the couple.

Due to the lack of normative data, the scores can be interpreted in the 4
scales with the 5 discrete response values for the items. When each scale is divided by the
the number of items summed achieves a continuous and homogeneous range from 1 to 5. To establish
the correspondence with the 5 discrete values of response to the item can be divided
continuous range in 5 intervals of constant amplitude. This amplitude is 0.80 and is obtained
by dividing the difference between the maximum value and the minimum value by the number of intervals
(5 - 1)/5. In this way, the values from 1 to 1.80 correspond to the discrete value 1 ('never')
or "nothing") and can be interpreted as non-violence situations, the values of 1.81 to
2.60 corresponded to the discrete value 2 ("sometimes" or "very little") and can be
to interpret as situations of low violence, the values of 2.61 to 3.40 corresponded to the
discrete value 3 (“quite” or “little”) and can be interpreted as quite situations
violence, the values from 3.41 to 4.20 corresponded to the discrete value 4 ('with a lot of
"frequency" or "quite a bit") and can be interpreted as situations of high violence, and the
Values from 4.21 to 5 correspond to the discrete value of 5 ('always' or 'much') and can be
interpret as situations of extreme violence.
It is proposed to calculate two indices of violence, one of violence suffered or of
victimization and another of violence exerted or perpetration, multiplying the scores
total frequency (F) by damage (D). These total scores, beforehand,
they can be divided by the number of items summed so that they have a continuous range and
homogeneous from 1 to 5 (f = F/27 and d = D/27 for victimization and f = F/10 and d = D/12 for
perpetration). In order for the index value to have a range from 0 to 100, to the product of the
frequency by damage (f*d) is subtracted by its minimum value (1), then it is divided by the
difference between the maximum and minimum value of the product (25 - 1 = 24) and finally we
multiply by 100, that is, index = [(f*d) - 1/24] * 100). In this way, a range is achieved
0 to 100.

The high correlations between frequency and harm within the violence suffered or of the
violence exercised, even when lower cross correlations exist, indicates that
there is a clear relationship in partner violence between the frequency of violent acts and the
damage, which justifies the creation of a product index (frequency x damage) to assess
a more comprehensive form of the violence suffered and the violence exercised. It should be noted that they are
the women who determine in the joint sample that the cross relationship is moderate
when in men it is high. The justification for the 2 indices is also given by the two
components with moderate correlation that explain the interrelation between the 4 scales.

Values in the indices less than or equal to 9.3 could be interpreted as

situations of absence of violence (combinations of f*d ≤ 1.80 x 1.80), between 9.34 and 24
as low or mild levels of violence (combinations of f*d ≤ 2.60 x 2.60, maximum sometimes
with very little damage), between 24.01 and 44 as significant levels (combinations of f*d ≤
3.40 x 3.40, maximum with frequency with little damage), between 44.01 and 69.33 as levels
alarming (combinations of f*d ≤ 4.20 x 4.20, maximum very often with quite a
damage) and ≥ 69.34 as extreme levels (combinations > 4.20 x 4.20).

124
 Daena: International Journal of Good Conscience

As limitations of the study, the non-probabilistic nature of the

sample. Thus, every generalization should be treated as a hypothesis applicable to a
similar population of Mexican women from the general population. It should be noted that the
The sample size was within the range considered suitable for the application.
of structural equation modeling, which is from 200 to 500 (Byrne, 2010).

Conclusions

The 27 items of frequency of violence experienced and the 27 of harm suffered can be
retained based on their internal consistency properties and factorial configuration,
giving rise to 2 scales and 6 factors (2 of frequency and 4 of damage) with internal consistency of
acceptable to excellent in the joint sample and in both sexes. The different types of
expected violence suffered (physical, psychological/social, economic, and sexual violence)
they live together in women and men, the phenomenon being essentially of nature
unidimensional in frequency and damage. The differentiation of factors is possible, but forced.
In the frequency of violence experienced, one could distinguish physical/sexual violence and another.
psychological/economic/social. In the suffered harm, one could distinguish between physical violence,
economic/social, sexual and psychosocial motivated by jealousy. These two models are not
strictly invariant between both sexes, being more defined in women. Ten of the
12 items of frequency of violence exercised and the 12 items of damage caused can be
retained based on their properties of internal consistency and factorial configuration,
giving rise to two scales and three factors (1 of frequency and 2 of damage) with internal consistency
from acceptable to excellent. In the frequency of violence exerted, the model would be unidimensional.
With 10 indicators and the damage caused, it would be possible to distinguish damage inflicted by violence.
psychological and damage inflicted by another type of violence, although the correlation between both
the factors of damage inflicted would be very high. These two models are not strictly
invariants between both sexes, being more defined in men.

Recommendations

The importance of considering the rates of violence suffered and inflicted is emphasized.
for a more accurate measurement of the phenomenon. The analysis between damage and frequency of the
violence suffered and exercised made evident that there is a strong relationship between the
frequency of violent acts and the damage. Furthermore, the present data shows that the
the distinction between the types of violence is forced, which could be expressing that when
if you are immersed in a violent partner relationship, it is difficult to establish these nuances
women being more capable of differentiating nuances in suffered violence and men in
exercised violence. These differential nuances are consistent with the expectation of
genre that positions the woman as a victim and the man as the perpetrator, which appears
in the study of the social representation of violence in couples in Latin countries (Ariza,
2013) and in the approach to research on partner violence in very
various countries (Straus, 2007). It is suggested to investigate whether they correspond to a relational reality,
as suggested by the United Nations Population Fund (UNFPA, 2009), or are they remnants of a bias
cultural, as suggested by Dutton and Nicholls (2005).

125
 Daena: International Journal of Good Conscience

References

Archer, J. 2002. Sex differences in physically aggressive acts between heterosexual partners: A meta-analytic
review. Aggression and Violent Behavior. 7(4): 313-351.
Ariza, G.R. 2013. The social representations of violence in couple relationships in Medellín in the
21st century. CES Psychology Journal. 6(1): 134-158.
Byrne, B.M. 2010. Structural equations with AMOS: Basic concepts, applications, and programming (2nd ed.).
Routledge, New York.
Cienfuegos, M.Y. and Díaz-Loving, R. (2010). Violence in the couple relationship. Pp. 647-687 In: Anthology
psychosocial of the couple: classics and contemporaries, R. Díaz-Loving and S. Rivera Aragón (eds.). Miguel
Ángel Porrúa, Mexico City.
Cronbach, L.J. and Shavelson, R.J. 2004. My current thoughts on coefficient alpha and successor procedures.
Educational and Psychological Measurement. 64(3): 391-418.
Dutton, D.G. and Nicholls, T.L. 2005. The gender paradigm in domestic violence research and theory: Part 1 -
The conflict of theory and data. Aggression and Violent Behavior. 10, 680-714.
Ellis, P.D. 2010. The essential guide to effect sizes: An introduction to statistical power, meta-analysis and the
interpretation of research results. Cambridge University Press, Cambridge, UK.
Fiebert, M.S. 2013. References examining assaults by women on their spouses or partners: an annotated
bibliography. Sexuality and Culture. 17(2): 1-67.
Hattery, A. 2009. Intimate partner violence. Rowman and Littlefield Publishers, Lanham, MD.
Linares, A., Champion, J.D., and Salazar, B.C. 2013. Psychometric characteristics of the Abuse Questionnaire
against women, Spanish version: its use in research and clinical practice. Psychological Therapy.
31(3): 335-334.
Moral, J. and López, F. 2014. Measurement and relationship between violence received and exerted against the partner. Journal
International of Psychology. 13(2): 1-50 [En line. Available in
Unable to access the content provided in the link.Consultation date: 1 of
November 2014.
Nicholls, T.L. Pritchard, M.M., Reeves, K.A. and Hilterman, E. 2013. Risk assessment in intimate partner
violence: A systematic review of contemporary approaches. Partner Abuse. 4(1): 76-168.
Women's Health Observatory. 2005. Catalogue of instruments for screening and frequency of abuse
physical, psychological, and sexual. Andalusian School of Public Health, Granada.
Secretary of the Interior and General Coordination of the National Commission on Women. 1999. Program
National against Domestic Violence (PRONAVI), 1999-2000. Secretariat of the Interior, City
from Mexico.
Mexican Society of Psychology. 2007. Ethical Code of the Psychologist (4th ed.). Trillas, Mexico City.
Straus, M.A. 2007. Processes explaining the concealment and distortion of evidence on gender symmetry in
partner violence. European Journal of Criminal Policy and Research. 13(3-4): 227-232.
Streiner, D.L. 2003. Starting at the beginning: an introduction to coefficient alpha and internal consistency.
Journal of Personality Assessment. 80(1): 99-103.
Trujano, P., Martínez, K. and Camacho, S. 2010. Male victims of domestic violence: an exploratory study
about their perception and acceptance. Diversitas. Perspectives in Psychology. 6(2): 339-354.
Trujano, P., Nava, C., Tejeda, E. and Gutiérrez, S. 2006. Confirmatory study on the frequency and
perception of domestic violence: the VIDOFyP as an evaluation instrument and some
Psychosocial reflections. Psychosocial Intervention. 15(1): 95-110.
United Nations Population Fund. 2009. UNFPA strategy and framework for action to address gender-based violence.
violence 2008-2011. UNFPA, New York.
Valdez, M., Híjar, M., Salgado, V., Rivera, L., Ávila, L. and Rojas, R. 2006. Violence scale and index
Severity: a methodological proposal to measure partner violence in Mexican women. Health
Public of Mexico. 48(Supl. 2): 221-231.

126
 Daena: International Journal of Good Conscience

Annex

Questionnaire on Violence Suffered and Exercised in Relationships (CVSEP).

Instructions: Indicate on a scale from 1 to 5 how often the following situations occur in your
couple relationship. Also indicate on a scale of 1 to 5 how much you feel harmed or attacked by
these situations.

Frequency Damage
1 2 3 4 5 1 2 3 4 5
Some With a lot Very
Never Quite Always Nothing Little Quite Much
times frequency a little

Situations Frequent.Damage
My partner has told me that my personal grooming is unpleasant.
My partner has pushed me hard.
3. My partner gets angry with me if I don't do what he/she wants.
4. My partner criticizes me as a lover.
My partner rejects me when I want to have sexual relations with him/her.
My partner watches everything I do.
My partner has said that I am ugly or not very attractive.
My partner does not take my sexual needs into account.
9. My partner prohibits me from hanging out or gathering with friends.
10. My partner uses money to control me.
11. My partner has hit or kicked the wall, the door, or some piece of furniture to
to scare me
My partner has threatened to leave me.
I have been afraid of my partner.
My partner has forced me to have sexual intercourse when I am
sick or my health is at risk or I feel bad.
15. My partner gets upset with my successes and achievements.
My partner has hit me.
17. My partner forbids me to work or continue studying.
18. My partner verbally assaults me if I do not attend to my children like he/she does.
think that it should be.
19. My partner gets angry if I don't take care of my children the way he/she thinks I should.
it should be.
20. My partner gets upset when I tell them that the money they give me is not enough.
I win.
My partner gets angry if the food or work is not done when he/she is.
I believe I should be.
My partner gets jealous and suspects my friends.
My partner manages the money without taking me into account.
24. My partner blackmails me with their money.
My partner has insulted me.
My partner limits me financially to keep me at home.
He/She gets angry when I don't contribute or give the expense that he/she thinks it should be.
27. My partner has mocked some part of my body.
I told him he is guilty of our problems.
I have reached the point of yelling at my partner.
I get angry when he contradicts me or disagrees with me.
I have resorted to insulting my partner.
I have threatened my partner to leave him/her.

127
 Daena: International Journal of Good Conscience

33. When I don't attend to the children as I believe, I verbally attack my

couple.
34. I do not take my partner's sexual needs into account.
I have forbidden my partner from hanging out or gathering with their friends.
I have physically hurt my partner.
It bothers me that my partner spends their own money.
I have demanded that my partner gives me detailed explanations of the way in which
spend the money.
I have told my partner that they are ugly or not very attractive.

About the Authors

Faculty of Psychology, Autonomous University of Nuevo León. c/Dr. Carlos Canseco 110. Col. Mitras
Centro. Monterrey, NL, Mexico.
emailjose_moral@hotmail.com

Academic Unit of Psychology, Autonomous University of Zacatecas. Av. Preparatoria No. 301. Col
Agronomic. C.P. 98060. Zacatecas, Zac., Mexico. Phone and Fax (492) 92 41934.
espera_sa@yahoo.com

128