The current paper is designed to consider the issue of domestic violence towards women in several different aspects. There are three contemporary studies analyzed in the scope of this research. Each study contains different information regarding domestic violence. Thus, Flake and Forste (2006) researched the issue of the reasons of domestic violence in Latin America, Crempien, Rojas, Cumsille, and Oda (2010) paid attention to the correlation between pregnancy and domestic violence occurrence, Green and Kelso (2010) considered punishment with regard of committing violence towards women.
The research aims at analysis of the reasons of committing domestic violence towards women and its consequences for petitioner and assaulter. The main purpose of this paper is to find out if the research of domestic violence can be improved using new statistical approach.
All of the studies were incorporated with the fact that statistical data collection in the sphere of domestic violence was greatly hampered by the several factors. Firstly, collected data cannot be adequately compared since there is no clear definition of domestic violence towards women in most of the countries. Secondly, data collection methods, design and sample size significantly differ in different countries. Thirdly, women often report data which is not relevant or refuse reporting. Fourthly, interpretation of comparisons across European countries is difficult because of the listed factors. Fifthly, surveys used for collecting data for the researched related domestic violence are designed for other purposes (Bloom, 2008).
In the Europe the part of the research on domestic violence is not available because of being designed for internal use only. When extracted from police statistics, it cannot be properly classified (Bloom, 2008). This fact influences relevance of the information represented. In some European countries, for example, in Sweden, accesses to statistical data related domestic violence require a written request addressed to authorities. In the United Kingdom Crown Prosecution Service started to be monitored on a regular basis (UNISEF, 2006).
The study represented by Flake and Forste (2006) reports statistical research in Latin America, namely: Haiti, Nicaragua, Colombia, Peru, and Dominican Republic. There are several research hypotheses which link domestic violence occurrence to several factors (socioeconomic and marital status, use of alcohol, decision-making power, family size, education homogeneity). For the measurement of the correlation between listed factors and the likelihood of domestic violence occurrence logistic regression techniques were used. The sample size is 6,082 in Colombia, 2,275 in Haiti, 588 in Dominican Republic, 6,728 in Nicaragua, and 15,174 in Peru. Samples were weighted in order to compensate for differences in responses.
Flake and Forste (2006) study was conducted using t-test which helped revealing correlation between independent variables. One of the main advantages of this study is that it showed applicability of Western theoretical methodology in Latin American context.
A research of Crempien, Rojas, Cumsille, and Oda (2010) represented data regarding domestic violence during pregnancy using survey of Peñalolén health centers patients. The research hypothesis was to identify if pregnant women of Peñalolén were exposed to the domestic violence in a greater extent than women who were not pregnant. The aim of the study was to explore the level of domestic violence risk for pregnant women and research their mental health with the help of correlations and frequency analysis. The research was conducted using a sample of 256 women who were screened and surveyed with regard to the domestic violence occurrences they experienced.
For the purposes of this research several statistical tests were conducted. The research identified the following control variables: weeks of pregnancy, age, number of children, and number of family members in the household. One-factor Anova was used for examination of significant associations with listed control variables. There was non-parametric correlation test used, including Cramer’s, phi, and chi-square in order to observe the correlation between variables listed above and domestic violence during women’s pregnancy (Crempien, Rojas, Cumsille, & Oda, 2010).
Another study represented by Green and Kelso (2010) examined the way an assaulter is punished and how the police officers provide petitioner’s safety in three locations of the USA. The study analyzed the sample of 285 women. In the scope of this study four hypotheses were developed, namely: assaulters are arrested when the victim is physically injured; offenders are more likely to have a warrant issued or arrested if children were injured during the incident; respondents are likely to be arrested if they committed a recurrent domestic violence; the respondents are likely to be arrested in the case if police officers care about injured person’s safety rather than those who do not support the petitioner’s safety.
For testing these hypotheses bivariate analysis was conducted. Hypothesis testing is supposed to be conducted using chi-square test and method of the cross-tabulation. Series of cross-tabulations will be used for running pairs of dependent and independent variables (Green & Kelso, 2010).
The Flake and Forste (2006) study revealed that alcohol use and marital status were the most significant predictors of domestic violence. The results showed that domestic violence occurrence was intimately tired to the national and cultural peculiarities, such as machismo and traditional perception of the roles distribution in a family. Undoubtedly, the research has certain limitations connected with the relevance of surveys used for collecting data.
Flake & Forste (2006) study represented a wide scope for practical implications since it clearly pointed at predictors of domestic violence. Thus, men in Latin America tend to abuse their cohabiters rather than wives. There is a strong tendency to cohabit in Latin America that can be estimated as a reason for alarm since women who tend to cohabit are exposed to a greater risk of domestic violence. Another significant factor of domestic violence threat is alcohol abuse expressed in encouragement and social pressure of using alcohol. An appropriate government social policy could be helpful in resolving this problem to reduce women susceptibility to this factor.
The model was designed to examine the magnitude of domestic violence issue in five Latin American countries across analysis of multiple datasets. The factors influencing domestic violence were analyzed in order to predict and eliminate such a practice in future.
The results reported by the Crempien, Rojas, Cumsille, and Oda (2010) study were as follows: 5.9% of surveyed women reported physical and 30.1% emotional violence during their pregnancy. The problem worsened with the fact that physical violence towards pregnant women was observed within the year previous to the survey. Also, the research revealed the correlation between domestic violence during pregnancy, recurrent occurrences of domestic violence and depressive states of pregnant women.
In accordance with Crempien, Rojas, Cumsille, and Oda (2010) study, psychological discomfort during pregnancy was strongly associated with all forms of violence (physical and emotional) and recurrent domestic violence experienced in preceding years. The findings showed that mental health state of pregnant women and domestic violence were linked.
The first hypothesis of Green and Kelso (2010) was supported in the course of the study and reported that offenders who caused injures to their victims were likely to be arrested rather than those who did not physically injured their victims. Police were not tend to arrest the offender when the victim exposed relatively slight trespass on the person, such as being slapping, grabbing, beating, or threatening with weapon.
The second hypothesis was supported in the first part, i.e. it was true that assaulters are more likely to be arrested in the case if children were exposed certain physical harm. The second part of the hypothesis was not supported because offenders were not arrested if children were not physically harmed. As a result of testing the third hypothesis it was not supported because there was no correlation between previous and current arrests found. The fourth hypothesis was supported because police officers were more likely to arrest the offender in case if they behave helpful towards victims. However, this hypothesis was not supported when police officer decided to issue a warrant (Green & Kelso, 2010).
There were several hypotheses stated in the Flake and Forste (2006) study, among them: women who cohabited with their male partners experienced domestic violence more frequently than married interviewees; women living in large families experienced violence more often than women living in small families; women living with partners who abuse alcohol were more likely to experience domestic violence than women whose partners did not get drunk; women from families with low socioeconomic status tended to experience marital violence more often than women who lived in families with high socioeconomic status; women engaged in no egalitarian relationships were more likely to experience domestic violence than women engaged in egalitarian relationships; women whose education level differed from education of their partners were likely to run the risk of being violated in a greater extent than women having same level of education with their partners.
The statistical tests provided the profile of a woman who wais likely to experience violence. Hereby, it is a woman who has a large family, cohabits with her partner, whose partner has alcohol issues; she does not take part in a family decision-making process, and has educational level different from her partner’s. Interestingly, the profile of a woman who is subject to domestic abuse in Latin America is very similar to the profile of abuse victims in most Western countries including Switzerland, the United Kingdom, and the USA (Flake & Forste, 2006).
The outcomes of the study conducted by Crempien, Rojas, Cumsille, and Oda (2010) are not comforting because they revealed that women are not ensured against domestic violence even being pregnant. The study had revealed the fact that emotional violence is muted by women in the most cases. This might be the reason of the emotional hurt rate growth. The fact cannot be muted anymore because emotional stress caused by an offender harms mental health of a pregnant woman. It is important to draw government attention to the domestic violence in order to reduce the number of domestic violence occurrences especially during pregnancy. It is known that injures experienced during pregnancy have the most tragic consequences for women.
In accordance with Green and Kelso (2010) study, in three locations researched, where mandatory arrests were adopted, they did not make a great impact on offenders. Thus, 56.4% of the assaulters were arrested and in 29.4% of cases a warrants were issued for them. Delaware, Denver, and Washington, D.C. reported 24.5%, 35.1%, and 40.4% of the cases with data available respectively. The percentage of arrests made up 18.9%, 52.8%, and 28.3% with respect to three listed States. The study did not reveal any significant differences between three locations mentioned above.
However, the study has several limitations, namely: the researches failed to measure the extent to which the victims were physically injured and there is no data about the offenders. These two limitations restrain thorough analysis of the results obtained. The study is based on the previous years’ research and is consistent with the previously published study. Also, there is no evidence of the fact that the officer who registered the fact of domestic violence was aware of preceding incidents with regard to a certain offender. Thus, it is not clear if the arrest or no-arrest decision was fair. However, jurisdictions reacted by establishing fast prosecution and adjudication procedures together with implementing a concept of primary aggressor. As a result, the number of arrests increased from 15% to 30% of all arrests (Green & Kelso, 2010).
Improvement of statistical data collection is a significant part of a statistical research. In most of the studies data was collected with the help of surveys designed for other researches. The research of domestic violence would benefit if specialized surveys are utilized. The surveys have to be composed using hidden questions not to hurt the victims of domestic violence.
Usually analytics choose between Fisher’s and chi-square tests when analyzing contingency tables. In this respect Fisher’s test offers better choice because it gives the exact p-value. Chi-square test is often used because of familiarity and simplicity of calculations while it yields an approximate p-value.
Correlation and linear regression methods are much alike and can be easily mixed. It sometimes makes sense to perform both linear regression and correlation at the same time. If the populations analyzed represent Gaussian set, then it makes sense to use Pearson correlation coefficient. In other cases Spearman correlation coefficient is preferable to use.
Also, it is important to choose correct variables because the regression line depends on which variable to choose. On the opposite of linear regression, linear correlations are not symmetrical with respect to variables. Like many other studies, Green and Kelso (2010) study does not take into account emotional and moral aspect of physical violence. With respect to the current research it does not consider the impact of domestic violence on children who were not physically injured but observed the occurrence.
Domestic violence is a serious problem in the world. Many countries offered studies devoted to this issue. In this paper there were three studies analyzed with respect to the issues of committing domestic violence and its consequences. There were several improvements of the statistical researches suggested in order to offer a possibility to obtain relevant data reporting domestic violence that could help drawing attention of governments of the countries all over the world to the researched problem.
Bloom, S. S. (2008). Violence against women and girls. A compendium of monitoring and evaluation indicators. National Online Resource Center on Violence Against Women. Retrieved from http://www.awid.org/Issuesand-Analysis/Library/Violence-against-Women-andGirls-a-Compendium-of-Monitoring-and-EvaluationIndicators
Crempien, R.C., Rojas, G., Cumsille, P., & Oda, M.C. (2011). Domestic violence during pregnancy and mental health: exploratory study in primary health centers in Peñalolén. ISRN Obstetrics and Gynecology, 2011 (2011), doi:10.5402/2011/265817
Flake D.F., & Forste, R. (2006). Fighting families: family characteristics associated with domestic violence in five Latin American countries. Journal of Family Violence, 21(1),
Green T., & Kelso, C. (2010). Police discretion in domestic violence incidents. The American Association of Behavioral and Social Sciences Journal, 14. Retrieved from http://aabss.org/Journal2010/AABManuscript3.pdf
UNISEF. Innocenti Research Centre. (2006). Domestic violence against women and girls. Retrieved from http://www.unicef-irc.org/publications/pdf/digest6e.pdf