• Users Online: 7286
  • Home
  • Print this page
  • Email this page
Home About us Editorial board Search Ahead of print Current issue Archives Submit article Instructions Subscribe Contacts Login 

 Table of Contents  
Year : 2015  |  Volume : 2  |  Issue : 3  |  Page : 117-122

Factors which predict violence victimization in Kenya

Sociology and Medicine Research Units, Athens Institute for Education and Research, Athens, Greece

Date of Submission08-Jan-2015
Date of Acceptance06-May-2015
Date of Web Publication3-Sep-2015

Correspondence Address:
Lincoln Jacob Fry
974 SW General Patton Terrace, Port St. Lucie, Florida 354953, USA

Login to access the Email id

Source of Support: None, Conflict of Interest: None

DOI: 10.4103/2384-5147.164419

Rights and Permissions

Aims: The purpose of this paper is to identify the factors that predict interpersonal violence at the personal level in Kenya. Another aim is to interpret the implications of the study for violence prevention programs. Setting and design: Study is set in Kenya and is based on the responses of 2,399 collected in 2012. Methods and materials: The study's dependent variable is reported violence victimization. Results: The logistical regression analysis identified seven factors that predicted violence victimization. These included being the victim of a property crime, payment of bribes to the police, fear of crime in the neighborhood, poverty, whether there was a police station in the neighborhood, whether the police were visible in the area, and the respondent's trust in the neighbors. The surprising finding was that 72 percent of the violence victims were also property crime victims. Conclusion: The findings imply that target hardening should be the basis used to implement violence prevention programs in Kenya. It appears that crime prevention efforts 5 should begin with law enforcement personnel when they respond to reported crime, property or violent offenses. These findings suggest that there is an ongoing need to protect victims from re-victimization by preparing them to protect both their premises and their persons in the future.

Keywords: Fear of crime, interpersonal violence, Kenya, property crime, re-victimization, target hardening, violence prevention

How to cite this article:
Fry LJ. Factors which predict violence victimization in Kenya . Sub-Saharan Afr J Med 2015;2:117-22

How to cite this URL:
Fry LJ. Factors which predict violence victimization in Kenya . Sub-Saharan Afr J Med [serial online] 2015 [cited 2022 May 17];2:117-22. Available from: https://www.ssajm.org/text.asp?2015/2/3/117/164419

  Introduction Top

In 1996, the World Health Organization (WHO) declared violence a major public health problem. [1] In 2000, WHO created the Department for Injuries and Violence Prevention, [2] and in 2002, released the World Report on Violence and Health. [3] Violence was included in the call for improved research that highlighted public health's need to address data collection deficiencies, including hospital and police records, in order to begin to develop preventive interventions, including injury control programs. Violence is a major societal problem in Kenya, which is rated 24 th worst in the world in violence deaths, and violence as the cause of death ranks 10 th in the country. [4]

The majority of the research concerned with violence in Kenya has most recently concentrated on election, [5] domestic, [6] and youth violence, [7] the city of Nairobi [8] and rural areas of Kenya, [9] have also received some attention in the Kenyan violence literature. As Maina et al., [10] indicated, Kenya, suffers from day-to-day high levels of interpersonal violence, which rarely receives any attention. They suggest that the magnitude of interpersonal violence may be severely under-estimated by official crime statistics.

One-way to measure the extent of crime in a society is through the International Crime Victim Survey (ICVS). [11] Crime victimization surveys are a key tool for crime statistics, especially in regions where law enforcement and criminal justice information systems face capacity-related challenges. ICVS was first conducted in 1989, and continually over the years, world-wide in 2006, and in various African countries from 2007 to 2010. The ICVS's study population includes all adult males and females 18 years of age and older. Data collection consists of face-to-face personal interviews utilizing a stratified multi-stage representative sample random selection process designed to generate estimates to the national population of all adults in Kenya that is, accurate to within a margin of error of plus or minus 2% points at a confidence level of 95%.

The most recent crime victimization survey conducted in Kenya was carried out by the Kenya Institute for Public Policy Research and Analysis, in collaboration with the United Nations Office on Drugs and Crime in 2010. [12] The findings concentrated on a range of victimizations, at both the individual and household levels. The findings showed that 6.3% of households reported burglary with entry and 3.7% indicated they were victims of attempted burglary. The highest reported crime for households was theft from cars, 20.7%, while assaults or threats accounted for 5.1% of the crimes against individuals. Fear of crime measures revealed that there wide differences between urban and rural respondents. Almost half, 48.6% of urban residents reported feeling either a bit unsafe or very unsafe in the street after dark, compared to 34.4 of the rural respondents. Urban residents thought they were likely to be burglary victims in the next 12 months, 34.5%, when compared to rural residents, 26.5%. Also of interest, 15.4% of respondents were found to be victims of corruption, and 62% of respondents indicated they thought that the police were doing a fairly or very good job. Respondents reported taking various crime prevention measures, including installing grilles for windows, door locks, high fences, getting a dog, and arranging with neighbors to watch the other's houses.

There has been an increasing volume of calls to develop violence prevention programs at the country, continental and international levels, as well as the concomitant need to begin to develop violence prevention programs. One approach to crime prevention will receive some attention later in this paper, Crime prevention through environmental design, which has gained some support in Africa, and elsewhere. It is sometimes called target hardening and is derived from what is known as the built environment framework. [13] Elements in the built environment include homes, schools, workplaces, parks/recreation areas, business areas, and roads. It encompasses all buildings, spaces and products that are created or modified by people. Research in this tradition has focused mainly on housing, transportation and neighborhood characteristics, [14] emphasizing improved protection of self.

  Materials and methods Top

Study Design

This study's data source is round 5 of the Afrobarometer project, a collaborative research effort produced by social scientists from 35 African countries. The project's objectives are as follows:

  1. To produce scientifically reliable data on public opinion in sub-Saharan Africa;
  2. To strengthen institutional capacity for survey research in Africa; and
  3. To broadly disseminate and apply survey results. Begun in 1999, 5 rounds of the survey have been completed; Kenya was included in 5 waves, 2 through 5, which became available in 2013. Round 6 was conducted in 2014, and that round is in process at this time.

Like the previous surveys, round 5 consisted of face-to-face Interviews and completed by 2399 Kenyans, 18 years of age or older. These interviews were conducted in 15 different languages. The sampling frame included all 8 Kenya provinces, and the final sample supports estimates to the national population of all adults in Kenya that is, accurate to within a margin of error of plus or minus 2% points at a confidence level of 95%. The sampling procedures used in all of the Afrobarometer surveys are explained in detail in Bratton et al. [15] Briefly, Stage 1 begins by selecting primary sampling units (PSUs), which are the smallest, well-defined areas for which reliable population data are available. The sample universe is stratified by administrative area (region/province) and by locality (urban or rural). Stage 2 begins by selecting sampling starting points. Within each selected PSU, field teams consisting of one supervisor and four interviewers travel to randomly selected starting points. If a reliable list of households is available for every PSU then it is obtained from the National Census Bureau or district or local officials. If neither household lists nor maps are reliable, the field supervisor contacts a traditional leader, local government councilor or government official knowledgeable about the area. Stage 3 is the selection of households to be interviewed, and supervisors choose any starting point and the four interviewers on the field team are instructed to walk away from this point in the following directions: Interviewer 1 walks toward the sun, interviewer 2 away from the sun, interviewer 3 at a right angle to interviewer 1, interviewer 4 in the opposite direction from interviewer 3. Households are selected randomly and in Stage 4 interviewees are randomly selected individuals from households. Interviewers obtain two interviews per PSU and eight interviews per PSU is the quota. This procedure applies in well-populated areas, but in sparsely populated areas the team shall drop off only one interviewer to conduct no more than two interviews.

Measures and Statistics

Violence victimization: Survey respondents were asked about criminal victimization. One question asked "over the past year, how often, if ever, have you or anyone in your family been physically attacked?" Fixed responses were provided as follows: Never; just once or twice; several times; many times; and always. The study's dependent variable was created by treating never as one category (0) and all other affirmative responses were coded as one (1). This dichotomous variable is the study's dependent variable and provides the basis for the cross-tabulations presented in [Table 1] and [Table 2], as well the logistic regression presented in [Table 3].
Table 1: Cross-tabulation violence victimization and selected independent variables

Click here to view
Table 2: Cross-tabulation of property and violent crime

Click here to view
Table 3: Logistic regression with violence victimization
as the dependent variable

Click here to view

The Independent Variables

Another question asked was during the past year, have you or anyone in your family: Had something stolen from your house? The same fixed responses that we used the attacked question were provided for this property crime indicator and coded in the same manner. All affirmative were coded as one (1) and negative answers were coded (0). A poverty scale used in the Afrobarometer studies was adopted from Mattes et al., [16] factor scaled, scale scores were calculated and assigned to each respondent; The questions which generated the scale were "over the past year, how often, if ever, have you or anyone in your family gone without the following;" enough food to eat, enough clean water for home use, without medical care, enough fuel to cook your food and a cash income? This scale's reliability coefficient was 0.79 (Cronback's alpha). The control variables listed in [Table 4] were measured by a single item, like age and other measures were collapsed into fewer categories; for instance, education, which was reduced to five categories.
Table 4: Demographic characteristics of the Kenyan respondents by violence victimization (n = 2399)

Click here to view

Other variables that were measured by single item included the fear of crime in the home and neighborhood. Others were qualitative in nature, like the presence of a police station in the respondent's local area, whether police were visible in the local area, and residential crowding were recorded by the interviewer and supplemented/checked by the interviewer's supervisor. Note that race was not included in able 1, or in the analysis which follows. This was because 99 of the respondents were found to have been classified as Black Africans. So race was not included in [Table 4], or in the analysis which follows.

  Results Top

The sample social and demographic characteristics of the sample are displayed in [Table 4], broken down by whether respondents were or were not victims of physical violence within the last year.

[Table 4] shows that there was one statistically significant difference in violence victimization in this Kenyan sample, and that was by education.

Those with lower education levels were more likely to be violence victims. Gender just fell short at P = 0.053) in this analysis. There was no significant difference in violence victimization by age, residence (rural-urban), employment status, and by respondent ownership of certain possessions.

[Table 1] displays violence victimization in the last year for selected independent variables. These items begin with the fear of crime in the home and in the neighborhood.

By way of contrast, most of the variables in [Table 1] were found to be significant independent variables related to violence victimization. Most of these measures were single item indicators. These included property crime victimization, both fear of crime measures, in the home and the neighborhood, and most police and trust measures. Significant trust measure included the police and neighbors, and respondent perceptions of the police as corrupt and whether they had actually paid a bribe to the police. Other items included the presence of a police station in the area and whether police were visible in the area. A final item asked whether the respondent's area was on the electric grid. This item was included because night lighting is considered a significant factor in neighborhood crime prevention.

Note that, [Table 1] shows that those who reported that they were afraid of crime in the home, 191 (24.3%) had been violent crime victims, as had 194 (19.4%) of respondents who reported they were fearful about crime in the neighborhood; these findings are addressed further below.

The next step in the analysis was to insert the independent variables listed in [Table 1] and [Table 4] into a logistical regression, which is displayed in [Table 3], with violence victimization the dependent variable.

[Table 3] reveals that seven independent variables reached significance in the logistical regression analysis. All of these were highly significant, with a probability of 0.01 or better. Property crime victimization was the strongest, Z = 7.92.

Whether the respondent paid a bribe to the police was next, Z = 4.61, followed by fear of crime in the neighborhood, Z = 3.95 and the poverty measure Z = 3.14. Significant at the 0.01 level the final three were the presence of a police station in the area, whether the police were visible in the area, and the level of trust the respondent had in neighbors. The regression results produced a pseudo R2 of 0.16. The surprising finding in [Table 3] was the strength of the property crime victimization measure in the logistical regression equation. As a result, [Table 2] takes a closer look at the violence and property crime measures.

[Table 2] reveals that 191 Kenyans, 25.7%, of the property crime victims, were also violent crime victims. Of the 266 identified violence victims, 191 respondents, 71.8%, were also property crime victims. The findings reported in [Table 1] are relevant here. This was where 194 respondents reported fear of crime in the home and 212 respondents reported they feared crime in the neighborhood. Collectively these findings point to the importance of respondents being either re-victimized or victims of both types of crime in a single incident. Furthermore, these finding support that old adage, "you are not paranoid, they are after you."

  Discussion Top

Before the further implications of these findings for crime prevention in Kenya are addressed, there are several issues that need to be mentioned. One of these is what Shepard [17] defined as criminal deterrence as a public health strategy. As Shepard suggested, despite the fact that violence is now seen as a public health issue, criminal deterrence as a public health strategy has been greeted with ambivalence and even hostility. That reality needs to be addressed and clearly reassessed, with CEPTED the logical framework to be adopted to pursue that approach. The second issue is methodological. The results of the findings presented in [Table 1] and [Table 2] highlight the first of these issues and weakness in this study. This is the need to establish the time priority for the physical and property crime victimizations. We are unable to determine from this data which victimization occurred first or if they occurred at the same point in time; that is, the old problem that correlation does not necessarily mean causation. This same caution applies to the fear of crime indicators. In terms of fear of crime, the results suggest that these Kenyan respondents did have a valid reason to fear crime, because a large percentage of them had in fact been victims of crime, both property and violent crime.

In summary, the logistical regression analysis showed that there were seven significant factors that predicted violence in Kenya. In order of their magnitude, these included being a victim of property crime, whether the respondent paid a bribe to police, fear of crime in the neighborhood, poverty, whether there was a police station in the area, whether the police were visible in the area and whether the respondent trusted neighbors. These suggest that target hardening should be the basis to begin to develop and implement violence prevention programs in Kenya. Re-victimization would be the key issue here.

Crime prevention/law enforcement personnel should respond and follow-up incidents of reported property and/or violence victimization in their jurisdictions. Personal experience suggests that crime victims are much more likely to listen and comply with target suggestions and requests if they have recently been victimized. The purpose of such an approach would be to attempt to prepare and assist victims to better protect both their premises and their persons. The Kenyan Victim Survey, mentioned above, showed that although Kenyan respondents did very little to adopt measures that prevented crime in their households. What they do was consistent with the target-hardening approach, like improving locks, installing proper night lighting and clearing bushes from in front of their windows that might impede visibility of their property and neighborhoods. Personal experience with target hardening programs suggests that residents become open to target hardening approaches, and personnel, once they have been victimized. Also, once victimized, residents are more encouraged to develop local neighborhood groups that provide security for themselves and those in their own communities.

  Acknowledgment Top

Afrobarometer Data, Kenya round 5 available at http://www.afrobarometer.org.

  References Top

World Health Assembly. Prevention of Violence: Public Health Priority (WHA 49, 25). Geneva: World Health Organization; 1996.  Back to cited text no. 1
Krug E, Sharma G, Lozanno R. The global burden of injuries. Commentaries. Am J Public Health 2000;90:523-6.  Back to cited text no. 2
Krug EG, Dahlberg LL, Mercy JA, Zwi AB, Lozano R, editors. World Report on Violence and Health. Geneva: World Health Organization; 2002.  Back to cited text no. 3
World Health Rankings: Kenya, Violence; 2013. Available from: http://www.world life expectancy.com/kenya violence. [Last accessed on 2014 Sep 2].  Back to cited text no. 4
Barkan J. Kenya: Lessons from a flawed election. Journal of Democracy 1993;4:85-99.  Back to cited text no. 5
Ondicho T. Domestic Violence in Kenya: Why battered women Stay. Int J Soc Behav Sci 2013;1:105-11.  Back to cited text no. 6
Ruteere M, Mutahi P, Mitchell B, Lind J. Missing the Point: Violence reduction and policy Misadventures in Nairobi's Poor Neighborhoods. University of Sussex in Brighton, UK: Institute for Development Studies; 2013.  Back to cited text no. 7
Parks M. Urban poverty traps: Neighbourhoods and violent victimisation and offending in Nairobi, Kenya. Urban Stud 2014; 51:1812-32.  Back to cited text no. 8
Francis P, Amuyunzu-Nyamongo M. Bitter harvest: The social costs of state failure in rural Kenya. In: Moser C, Dani A, editors. Community Based Institutions and Asset Creation. Washington, DC: World Bank; 2007.  Back to cited text no. 9
Maina W, Munguti N, Mwai W, Butchart A, Cannooodt L. Estimating the Annual Incidence of Violence Related Injuries in Kenya: A Derivation from Review of Hospital Reports in 2007. Public Health Res 2013;5:136-41.  Back to cited text no. 10
Van Dijk J, Mayhew P, Killias M. Experiences of Crime Across the World: Key Findings from the 1989 International Crime Survey. Deventer: Kluwer Law and Taxation; 1990.  Back to cited text no. 11
Victimization Survey in Kenya (2010). Available online at http://www.unodc.org/unodc/en/data-and-analysis/Data-for-Africa.html.  Back to cited text no. 12
Rapoport A. The Meaning of the Built Environment: A Nonverbal Approach. Tuscon Arizona: University of Arizona Press; 1982.  Back to cited text no. 13
Sampson RJ, Raudenbush SW, Earls F. Neighborhoods and violent crime: A multilevel study of collective efficacy. Science 1997;277:918-24.  Back to cited text no. 14
Bratton M, Mattes R, Gyimah-Boadi E. Public Opinion, Democracy, and Market Reform in Africa. Cambridge: Cambridge University Press; 2005.  Back to cited text no. 15
Mattes R, Bratton M, Davids Y. Poverty, Survival, and Democracy in Southern Africa, Afrobarometer. Working Paper NWP023; 2003.  Back to cited text no. 16
Shepard J. Criminal deterrence as a public health strategy. Lancet 2001;358:1717-22.  Back to cited text no. 17


  [Table 1], [Table 2], [Table 3], [Table 4]


Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
Access Statistics
Email Alert *
Add to My List *
* Registration required (free)

  In this article
   Materials and me...
   Article Tables

 Article Access Statistics
    PDF Downloaded209    
    Comments [Add]    

Recommend this journal