 

90m

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the
thesis entitled

CRIMINAL VICTIMIZATION IN THE PHILIPPINES:
A TEST OF LIFESTYLE-EXPOSURE, ROUTINE ACTIVITIES AND
SOCIAL DISORGANIZATION THEORIES

presented by

RAYMUND ESPINOSA NARAG

has been accepted towards fulﬁllment
of the requirements for the

Master of degree in
Science Criminal Justice

 

 

Major Prge’ssor’s Signature

/ 1&7. 07% 02W

Date

MSU is an afﬁnnative-action, equal-opportunity employer

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

MARUZQIZ
0831 13

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:lClRC/DateDue.indd-p.1

CRIMINAL VICTIMIZATION IN THE PHILIPPINES:
A TEST OF LIFESTYLE-EXPOSURE, ROUTINE ACTIVITIES AND
SOCIAL DISORGANIZATION THEORIES
By

Raymund Espinosa Narag

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER OF SCIENCE
School of Criminal Justice

2007

ABSTRACT
CRIMINAL VICTIMIZATION IN THE PHILIPPINES:
A TEST OF LIFESTYLE-EXPOSURE, ROUTINE ACTIVITIES AND
SOCIAL DISORGANIZATION THEORIES
By
Raymund Espinosa Narag

This research studies criminal victimization in the Philippines. Employing a 1995
multi-stage sampling data of 1200 households, this research analyzes different variables
that are associated with property and violent victimization. Traditional Western theories,
particularly lifestyle-exposure and routine activities theories emphasize individual level
variables of guardianship, target attractiveness and proximity to potential offenders, while
social disorganization theory emphasizes structural and neighborhood characteristics such
as level of urbanization, population mobility, neighborhood cohesiveness, and public
formal and informal controls, to be key determinants in criminal victimization. Using
these Western models as guide, this research evaluates whether similar measures can be
replicated in a non-western, developing society. Results Show that Western theories can
be successfully integrated in the local setting and that variables derived from these
theories have the capacity to explain risk of victimization. Speciﬁcally, property crimes
are related strongly to social disorganization variables and violent crimes to routine-
lifestyle variables. However, there are variables that are associated with victimization in
the opposite direction, indicating that some variables are context speciﬁc. Finally, the
models explain a small portion of the total variation in victimization, indicating that other

variables unique to the social, political and cultural milieu of the Philippines may need to

be incorporated in the model.

DEDICATION

For Shella and our two angels: Ella and Angel, with whom I toiled hard for the
completion of this task;

For my parents, Robert and Helming, and parents-in-law, Gabby and Ritchie, for
loving their grandchildren unconditionally;

For my three sisters: Jen, Mads and Jane and Sister-in-law, Dimple, for sharing all
their joys to our family;

For my ﬁve “brads” (Dan, JV, Warren, Chris and Mike) who are unjustly put

behind bars: may this work serve as an additional voice for their clamor for freedom.

iii

ACKNOWLEDGMENTS

I am indebted to my committee members Drs. Sheila Maxwell, Edmund
McGarrell and J esenia Pizarro for the completion of this thesis. Through their guidance,
I learned the rigors of a scientiﬁc research. I particularly thank Dr. Sheila Maxwell for
her patience in answering all the questions of an amateur-researcher and reading all my
countless revisions. I also thank Dr. Edmund McGarrell for inspiring me to look for a
topic dear to my heart and for awarding me an assistantship that kept my family aﬂoat
during the critical moments of our stay in the United States. Finally, I thank Dr. J esenia
Pizarro for sharing her insights, visions and dreams, personally and professionally, about
issues in criminal justice and beyond. I particularly appreciate her “advance” notices
which made my presentations not so dreadful. To the three of you, I am inspired to take
another step in my academic growth.

I also thank the Alociljas (Kuya Rex, Ate Vangie and JD) who became our foster
family here in the United States. My family enjoyed every moment of our stay here with
you around.

Finally, I thank all my friends in the Fellowship of Christian Internationals and
Friends and the MSU Filipino Club— who laughed with me every time I announce that I

need another “major” revision. Kasayahan na! (Let us celebrate!)

iv

TABLE OF CONTENTS

LIST OF TABLES ........................................................................... v
CHAPTER I: INTRODUCTION .......................................................... 1
CHAPTER 11: REVIEW OF LITERATURE ............................................. 3
Demographic attributes and the lifestyle-exposure theory ..................... 3
Predicates of crime and the routine activities theory ........................... 4
Individual routine-lifestyle and the criminal opportunity theory .............. 6
Contextual variables and social disorganization theory ........................ 7
Routine activities and social disorganization theories:
Convergence and divergence ................................................... 10
Integrating the demographic, routines and contextual variables ............ 12
Results of empirical studies ...................................................... 13
Violent versus property crimes ........................................... 13
Domain speciﬁcations ...................................................... 14
Detailed measurements ..................................................... 15
Depending on use of controls ............................................. 15
CHAPTER III: CONTEXT OF STUDY SITE ............................................ l7
Present study .......................................................................... 17
Socio-political and historical context ............................................... 17
The current state of Philippine neighborhoods ................................... 23
CHAPTER IV: DATA AND METHODS ................................................. 25
Sampling scheme ..................................................................... 26
Dependent variables ................................................................. 27
Independent variables and hypotheses ............................................ 31
Methods of analysis .................................................................. 47
CHAPTER V: RESULTS .................................................................... 50
Bivariate Analysis
Variables associated with property victimization ....................... 50
Variables associated with violent victimization ......................... 55
Summary discussion of the bivariate relationships ..................... 59
Multivariate Analysis
Multivariate predictors of property victimization ....................... 60
Multivariate predictors of violent victimization ......................... 67
CHAPTER VI: DISCUSSION AND CONCLUSIONS .................................. 72
Maj or Themes ....................................................................... 72
Limitations and Suggestions for Future Research ............................... 77
APPENDICES ................................................................................ 83
REFERENCES ................................................................................ 92

LIST OF TABLES

Table 1 Variable Coding and Descriptive Statistics,

sws Survey 4th Quarter, 1995 ..................................................... 29
Table 2 Percent of Property Victimization

By Demographic, Lifestyle-Routine and Contextual Characteristics ......... 51
Table 3 Percent of Violent Victimization

By Demographic, Lifestyle-Routine and Contextual Characteristics ......... 56
Table 4 Logistic Regression Model of Property Victimization

Model 1 .............................................................................. 61

Model 2 .............................................................................. 63

Model 3 .............................................................................. 66
Table 5 Logistic Regression Model of Violent Victimization

Model 1 .............................................................................. 67

Model 2 .............................................................................. 69

Model 3 .............................................................................. 71

vi

I
INTRODUCTION

Criminal victimization is a major concern of individuals in every society.
Especially with the advent of the Internet and explosion of information technology,
reports of Crime occurrences easily spread among the population. As such, individuals
may take precautions like buying sophisticated security paraphernalia to protect their
houses from burglaries. They may also opt to lessen the risk by reducing the times they
go out late at night or the frequency of visiting vulnerable places like bars and disco pubs.
It may also affect their decisions on where to buy homes and where to work. As such,
criminal victimization affects how ordinary citizens go on with their daily lives (Miethe,
1995; Sherman, Gartin and Buerger, 1989).

Governments, businesses and individuals beneﬁt when they know what speciﬁc
factors are related to criminal victimization. Crime control agencies, for example can
design ways that will curb the occurrences of crimes in speciﬁc locales, as when they
receive repeated calls of physical confrontations emanating from a tavern. Businesses
may be able to reduce the occurrences of theft by identifying which objects are most
prone to shopliﬁers. Additionally, individuals may reduce their risk by staying away
from areas that will expose them to potential victimization.

The study of criminal victimization, however, has started fairly recently. The
advent of the victimization surveys in the 1970’s, which were initially developed as an
alternative method of measuring crime, on top of the ofﬁcial crime rates reported by the
police and other criminal justice agencies, paved the way for scholars to study the

speciﬁc correlates of crime victimization (Smith and J arjoura, 1989). Through the years,

a respectable pool of knowledge was generated which warranted the development of the
area of victimology in the mainstream criminology (Smith and Jarjoura, 1989).

Generally, there have been three areas where scholars devote their study on
victimization. These are the demographic characteristics of the individuals, the lifestyle
and major daily routine activities the individuals undertake, and the contextual attributes
of where the individuals live, work and spend their leisure activities (Meier and Miethe,
1993; Sampson and Wooldredge, 1987). In the late eighties to the present, there have
been attempts to come up with a theory of victimization that accounts for these three
major areas.

Most of the studies, however, are done in the United States, the United Kingdom,
Canada and other developed countries. This is not surprising considering that
victimization surveys usually entail massive ﬁnancial expenditures, which most
developing countries do not have the luxury to afford (Lee, 2000). As such, the factors
that have been associated with criminal victimization in the western contexts have not
been rigorously tested in other contexts. Given the differences in economic, social,
political and cultural systems, it is interesting to know how Western theories of criminal

victimization measure up across cultural settings and milieus.

CHAPTER H
WHAT DO WE KNOW ABOUT CRIMINAL VICTIMIZATION?
A REVIEW OF THE LITERATURE

There are three main theories that have been offered to account for criminal
victimization. These three theories have many overlapping concepts and measures,
although in some instances, their predictions vary. The following section reviews the
main propositions of the theories and the manner in which the variables are measured.
Demographic attributes and the lifestyle-exposure theory

It is a common ﬁnding in the Western criminal victimization literature that males,
the young, the unmarried, and racial minorities have higher risk of becoming victims of
crime than females, older people, married and Caucasians (Hindelang et a1, 1978). This
common ﬁnding is explained by the lifestyle-exposure theory, which posits that
demographic attributes and ascribed status (gender, age, race, marital status, education,
occupation and income) determine the role expectations society has on particular
individuals (Hindelang et al, 1978). The demographic attributes and ascribed status also
impose constraints on the actions of the individuals. In turn, role expectations and
constraints determine how individuals adapt to their conditions and how they fashion out
a lifestyle attuned to their situations (Hindelang et al, 1978; Maxﬁeld, 1987). As
explicated by Meier and Miethe, “both ascribed and achieved status characteristics are
important correlates of predatory crime because this status attributes carry with them
shared expectations about appropriate behavioral choices” (1993:446). These variances
in lifestyles among individuals determine their differential exposure to dangerous places,

times and situations in which there are high risk of victimization (Hindelang et a1, 1978;

Meier and Miethe, 1993). For example, females are thought to be less victimized than

males due to the fact that females spend a greater proportion of their time inside the home
because, as “adolescents they are more closely supervised than males and as adults they
are more likely to assume housekeeping and child rearing responsibilities” (Hindelang et
a1, 1978; Meier and Miethe, 1993: 446). Also, the married and older people are more
focused on family activities compared to the unmarried and younger people, thus
lowering their chances of victimization. As such, lifestyle—exposure theory incorporates
individual-level variables in explaining target selection in criminal incidents and accounts
for the differences in the criminal victimization among groups (Garofalo et al, 1987).
Predicates of crime and the routine activities theory

The lifestyle-exposure theory is enhanced by the introduction of the routine
activities theory by Cohen and Felson in 1979. Routine activities are “the recurrent and
prevalent vocational and leisure activities individuals undertake in a regular day-to-day
basis, whatever their biological or cultural origins” (Cohen and Felson, 1979: 593).
Routine activities occur at home, at jobs away from home, and other activities away from
home.

Routine activities theory proposes that illegal and criminal acts feed on the daily
activities of individuals. It posits that structural changes in routine activities affect crime
rates through the convergence in time and space of three minimal elements of direct
contact predatory violations: (1) motivated offender (2) suitable targets and (3) absence
of a capable guardian against a violation (Cohen and F elson, 1979: 589). In here, Cohen
and Felson use Glaser’s (1971:4) deﬁnition of predatory crime as “illegal behavior during
which an individual takes or damages the property of another.” Targets could be a

person, an object or a place. Cohen, Kluegel and Land (1981) further reﬁned the concept

of target suitability in terms of visibility (that is, it has risk of discovery), inertia (that is,
the target could be overcome, like weight for property targets), value (it has material and
symbolic desirability) and accessibility (that is, it has site for entry as well as exit).
Capable guardian is deﬁned as a person or an object whose mere presence could deter
potential offenders from perpetrating an act. Lynch (1987) describes guardianship as to
include professional guards, laymen with an interest in preventing victimization, alarms
and the like. Cohen, Kluegel and Land (1981) extend the list to include persons such as
law enforcement ofﬁcers, housewives, neighbors, pedestrians and objects such as
window bars and locks and closed circuit televisions. These examples can be grouped in
three categories: formal social control, informal social control or target hardening
activities (Lee, 2000).

The third component is proximity to a motivated offender. This entails the
“physical distance between areas where potential targets of crime reside and areas where
relatively large populations of potential offenders are found” (Cohen, Kluegel and Land,
1981: 507). This aspect emphasizes the spatial component in explaining criminal
victimization. As Roncek and Maier (1991) note, routine activities theory stressed the
importance of the environment as a necessary component of criminal interactions
between potential offenders and victims.

Routine activities theory was proposed to account for the variations of crime rates
through time, such as when traditional criminological theories cannot account for rising
crime rates when variables believed to be related to crimes (disparities in poverty levels,
employrnents, incomes and educational attainments) were decreasing (Cohen and F elson,

1979). The massive changes in the post-war American society, like the presence of

women in the labor force and in schools, the prevalence of transportation and out of home
travel and the abundance of portable goods, were positively linked to increases in
property and violent crimes (Cohen and F elson, 1979). This conformed to the prediction
that the dispersion of activities away from households and families increases the
opportunity for crime and thus generates higher crime rates. Thus, changes in routine
activity patterns of everyday life can increase crime rates even if the social forces that
enhance criminal inclinations remain constant (Cohen and Felson, 1979; F elson and
Cohen, 1980). Rather than emphasizing the characteristics of offenders, routine
activities theory concentrates upon the circumstances in which offenders carry out
predatory criminal acts (Cohen and Felson, 1979; Maxﬁeld, 1987).
Individual routine-lifestyle and the criminal opportunity theory

Lifestyle/exposure theory and routine activities theory had been integrated to form
a “criminal opportunity” theory in explaining victimization. For example, the idea of
“exposure” in lifestyle-exposure theory is conceptually linked to the idea of “proximity”
in routine activities theory. Most scholars say that the differences between the two
theories are simply in semantics (Eck, 1995; Maxﬁeld, 1987). Also, lifestyle-exposure
and routine activities theories are widely acknowledged as complementary in that they
both begin with the assumption that crime and crime rates are products of non-random
events in which victims unwittingly participate (Garofalo, 1987; Massey, et a1, 1989).

With the integrated form of criminal opportunity theory, scholars had been able to
account for different phenomena that increase the risk of criminal victimization. For
example, the concept of “hot Spots” and “deviant places” (Sherman, et a1, 1989; Stark,

1987) like frequenting bars, disco pubs, public parks and adult stores, were proposed to

increase risk of victimization due to greater exposure to motivated offenders and lesser
guardianship in these places. “Deviant lifestyles,” drug and alcohol use, and offending
behaviors were likewise linked to criminal victimization (Sampson, 1985) on the ground
that engaging these activities makes one proximate to potential offenders. Neighborhood
daily activities were also conceptually linked to criminal victimization on the basis of
social guardianship, that is, the capacity of neighbors to supervise and protect other
neighbors and their properties (Lee, 2000). Family structures, household characteristics,
whether homes are occupied or not, and dwelling types were also related to criminal
victimization on their supposed varying degrees of guardianship (Kennedy and Forde,
1990; Massey et al, 1989; Sampson, 1987; Sampson and Lauritsen, 1990; Stahura and
Sloan, 1988). Major daytime and nighttime activities, locus of residence, work and
leisure, were theoretically connected to victimization on their differing effect on the level
of exposure and proximity to motivated offenders (Cohen and Cantor, 1980; Massey et
al, 1989; Miethe, et al, 1987; Miethe and McDowall, 1993). Likewise, ownership of
portable items, spending patterns, wearing of jewelries and bringing cash, were conceived
to impact risk of victimization on the account of their attractiveness to offenders (Cohen,
1980; Miethe and McDowall, 1993). Indeed, more speciﬁc and more reﬁned measures
are continually developed to capture the main concepts of the theory.
Contextual variables and social disorganization theory

Another theory that is proposed to account for the varying degrees of
victimization is social disorganization theory. Initially conceptualized by Shaw and
McKay (1942) to account for effects of social disorganization on youth delinquency, this

theory has been reformulated by later scholars to account for victimization risk as well

(Bursik, 1988; Bursik and Grasmick, 1993; Komhauser, 1978). This theory posits that
communities characterized by high racial and ethnic heterogeneity, population turnover
and mobility, and levels of economic disadvantage, such as unemployment, poverty and
inequality, will have members that will experience more criminal victimization. This is
because socially disorganized communities are “unable to realize the common values of
their residents and maintain effective social controls” (Bursik, 1988: 12; Komhauser,
1978: 120). For example, “high population turnover rate affects social integration by
decreasing permanence and stability in personal relationships” (Sampson, 1987: 336;
Veysey and Messner, 1999: 160). Population heterogeneity increases inter-group
tensions and cultural conﬂict among the residents, while areas characterized by low
economic opportunity have lesser capability to mobilize resources (Meier and Miethe,
1993; Sampson and Groves, 1989). On the other hand, areas with high levels of
organization are able to take note of or question strangers, watch over property, supervise
youth activities and intervene in local disturbances (Bursik, 1988: 541). Differences in
the communities’ level of social controls account for intra-and inter-city variation in
crime rates. More speciﬁcally, social disorganization theory is most applicable in
explaining total crime rates (Sampson and Groves, 1989; Veysey and Messner, 1999).
Two other variables—level of family disruptions in the community and degree of
urbanization were constructed to be facets of social disorganization. Family disruptions,
measured in terms of the number of single-parent, female—headed and divorced or
separated families in the community, were conjectured to be associated with
victimization (Sampson, 1985). Two-parent and male-headed households are thought to

provide increased supervision not only on their children and household property but also

for general activities in the community, like supervision of neighborhood youths and
surveillance of suspicious persons in the community (Sampson, 1985; Sampson and
Groves, 1989; Sampson and Wooldredge, 1987). Family disruption may also “increase
victimization risk by decreasing a community’s informal social controls” (Sampson,
1987: 333). Urban communities, on the other hand, are construed to have a decreased
capacity for social control, compared with suburban and rural areas. In particular,
“urbanization may weaken local kinship and friendship networks and impede social
participation in local affairs” (Sampson and Groves, 1989:782).

Further reﬁnements to the social disorganization theory have been made by recent
studies (Kubrin and Weitzer, 2003). The dynamics of social disorganization variables
(population heterogeneity, residential turnover and mobility, economic disadvantages,
family disruptions and level of urbanization) have been proposed to affect criminal
victimization indirectly. Sampson and Groves (1989) contend that there are intervening
dimensions, like organizational participation, social networks and supervision of youths
that affect social disorganization. These variables are thought to inﬂuence levels of
social cohesion, social capital and collective efﬁcacy, which have direct effects to levels
of criminal victimization (Kubrin and Weitzer, 2003). Social cohesion is conceived as
the denseness and sparseness of ties, measured in terms of the number of people known
in the neighborhood and memberships in clubs/organizations in the community (Kubrin
and Weitzer, 2003; Sampson, et a1, 1997). Social capital is conceived as the depth of
resources community members can maximize in their areas (Kubrin and Weitzer, 2003).
For example, poor communities are conceived to “lack money and resources and,

therefore, have fewer organization opportunities for youths and adults” (V eysey and

Messner, 1999: 166). Collective efﬁcacy is a further enrichment of social ties and social
capital, where these two attributes are mobilized toward the control of crime (as opposed
to a counter-tendency where highly-knit communities foster the development of crimes)
(Patillo, 1998; Sampson, 1997; Sampson, et al, 1999). As Kubrin and Weitzer (2003:
377) explicate, “networks and resources may be necessary, but not sufﬁcient for social
control. What is missing is the key factor of purposive action, that is, how ties are
activated and resources mobilized to enhance social control.”
Routine activities and social disorganization theories: Convergence and Divergence
Both routine activities and social disorganization theories emphasize the
contextual nature of the occurrences of crimes. Both theories also use macro-sociological
variables that explain risk of criminal victimization. However, their predictions using
speciﬁc indicators may converge or diverge. For example, communities that have
neighbors who help each other out are predicted by both theories to have a reduced risk
of criminal victimization. Using routine activities perspective, neighbors helping each
other out enhances guardianship, thus outsiders to the community are deterred ﬁom
committing crimes (Lee, 2000). Using social disorganization perspective, this same
measure manifests institutional control, thus effectively putting the neighborhood in
direct command of their environment (Sampson, 1987). This notion of social
guardianship is neatly integrated by Hunter’s (1985) concept of three levels of systemic
control in a community (see also Bursik and Grasmick, 1993). Private social control
corresponds to the guardianship by intimate members of the family and close friends,
parochial social control pertains to the protection and supervision by neighbors and

informal groups in the community and public social control refers to communities’ ability

10

to mobilize the formal guardianship of state authorities (Bursik and Grasmick, 1993;
Hunter, 1985). Scholars have successfully linked communities’ mobilization of public
social control to levels of criminal victimization (Velez, 2001).

Another example where there is convergence in prediction is the family
characteristic. Using routine activities theory, one-parent, female-headed and divorced or
separated families decrease the guardianship over members and household properties thus
foreseeing a higher risk of victimization (Kennedy and Forde, 1990; Massey et al, 1989;
Meier and Miethe, 1993). On the other hand, this measure of family disruption is a form
of social disorganization that hampers social control or lessens attachment to family
values (Sampson, 1987).

However, there are also numerous indicators where the directions of prediction
diverge. The ﬁrst example is number of members in the household. Routine activities
theory predicts that households with higher number of members enhance guardianship.
However, social disorganization theory predicts the opposite on account of the
criminogenic effect of household crowding (Meier and Miethe, 1993). Members of more
crowded households compete for household resources and thus induce conﬂict. Another
example is family income. From routine activities perspective, this represents
attractiveness to burglars and thieves, thus a positive relationship is predicted (Meier and
Miethe, 1993). However, in social disorganization perspective, this measure indicates
greater resources, thus signifying greater social capital for the community and forecasting
a negative relationship with victimization. Similarly, the measure “unemployed persons”
may be construed to represent increases in guardianship on the household on the account

that unemployed persons stay in the home, but in the social disorganization theory, this

11

variable may represent economic strains that will lead them to criminal acts and thus
higher victimization rates (Meier and Miethe, 1993).
Integrating the demographic, routine-lifestyles and contextual variables

Attempts to reconcile the sometimes-differing predictions of the theories had been
done by integrating the different variables in multivariate analysis (Meier and Miethe,
1993; Sampson and Wooldredge, 1987). Concentrating on the demographic and lifestyle
variables without including the structural features and community processes may result in
spurious relationships. Likewise, giving attention only to the spatial context may neglect
the fact that some individual-level attributes may persist despite controlling for the
community characteristics. For example, the positive relationship found between
households occupied by primary individuals and victimization could be attributed to the
fact that households occupied by single persons are more likely to be located in
neighborhoods that are poorer, have larger concentrations of single-parent households,
and are more racially heterogeneous (Sampson and Wooldredge, 1987; Smith and
Jarjoura, 1989). Also, the differing effect of family income could be explained in the
“differing levels of aggregation” (Smith and J arjoura, 1989:622). At the individual level,
household income is a measure of target attractiveness. Within neighborhoods, higher
income households may have a higher probability of victimization because they
“represent more attractive targets to potential burglars than lower income households”
(Smith and Jarjoura, 19892622). At the aggregate level, however, higher income
households are more likely to be located in higher status areas, thus representing higher

social capital and resources. Conceptual ambiguity could therefore be clariﬁed by

12

acknowledging that the same indicator is measuring both an individual level process
(target attractiveness) and an aggregate phenomenon (higher social control).
Results of empirical studies

Despite the theoretical clarity of the three theories, the empirical ﬁndings of
extant research Show that there are no variables (demographic, routine-lifestyles and
contextual) that could consistently predict the occurrence of criminal victimization. The
performance of the different exogenous variables is subject to crime type speciﬁcations,
subject-domains and level of sophistication of the measures employed and use of controls
(See also Land, McCall and Cohen, 1990). The relative strengths of the different
variables to predict risk of victimization vary across studies.
Violent versus property crimes

One recurring ﬁnding is that correlates of victimization depend on the speciﬁc
crimes (Cohen, Kluegel, and Land, 1981; Kennedy and Forde, 1990; Miethe and
McDowall, 1993; Miethe, Stafford and Long, 1987; Sampson, 1987). This is especially
true for differences between violent and property crimes. Miethe, Stafford and Long
(1987) for example ﬁnd that the variables measuring exposure (nighttime activity
measured as number of times going out during the week) are positively associated with
property crimes but have no effect on violent crimes. Miethe et a1 (1987) speculate that
the opportunity structures for property crimes, where offenders may base their decisions
to engage depending on the exposure of the target, is different from violent personal
crimes which usually happen in non-premeditated, spontaneous manner. This is similar
to the earlier ﬁnding by Cohen, Kluegel and Land (1981) which report that income

(measured as household annual income) is negatively related to assault, positively related

13

to personal larceny and parabolically related to burglary. Kennedy and Forde (1990),
using victimization survey data from Canadian cities report that the measure “percent
divorced in the community” is positively related to breaking and entering (burglary) and
vehicle theft but negatively related to assault and no signiﬁcant relation to robbery.
These studies show the need to disaggregate the different kinds of criminal victimization.
Domain speciﬁcations

Domain speciﬁcations also affect the capacity of the traditional variables to
predict criminal victimization (Lynch, 1987; Mustaine and Tewksbury, 1998;
Wooldredge, Cullen and Latessa, 1992). Domains of social life like school, work, home
and leisure (Lynch, 1987) are found to have speciﬁc social and cultural patterns of
victimization unique to their own environments. Speciﬁcally, Lynch shows that the
characteristics of one’s job are important in predicting personal victimization. For
example, the study by Wooldredge, Cullen and Latessa (1992) on the victimization risks
among university professors reports that gender, age and race have no signiﬁcant effect
on property and personal victimization. Mustaine and Tewksbury (1998) report that for
major theft victimization among college students, demographic variables like age, race
and marital status do not matter. The poor performance of the demographic variables
(which is contradictory to most ﬁndings) could be due to the level of domain
speciﬁcations where a more particular subgroup of people is studied. In speciﬁc
domains, the demographic differences may disappear and the recurrent work, school or
leisure activities are dominant predictors of victimization. These studies show the need

for further research in varied areas that looks unto different social and cultural context.

14

Detailed measurements

The performance of the concepts linked to criminal victimization depends also on
how they are speciﬁcally measured (Lynch, 1987; Meier and Miethe, 1993; Miethe and
McDowall, 1993; Massey et al, 1989; Mustaine and Tewksbury, 1998). For example,
Cohen and Felson (1979) measured “exposure” in terms of activities away from the home
and found positive correlation with victimization. However, Miethe et a1 (1987) ﬁnd that
some speciﬁc activities away from home, like going to school and work may in fact be
related to higher guardianship offered by schoolmates and ofﬁcemates. Similarly,
Mustaine and Tewksbury (1998) ﬁnd that it is more important to determine where one
goes when individuals leave the home or what activities one is participating when one is
out in the public, than using the general measure “out of home” (p. 851, italics original).
Depending on use of controls

The primary factor that yields contradicting results for the predictors of criminal
victimization, however, is whether the studies used controls in multivariate analysis.
Most of the demographic variables may lose explanatory power once speciﬁc measures
of routine activities or contextual variables are included in the model. Also, routine
activities variables may become insigniﬁcant once neighborhood characteristics are
introduced. For example, Miethe and McDowall (1993) employing multivariate logistic
analysis ﬁnd that in violent crimes, age becomes insigniﬁcant once contextual variables
are introduced, whereas in burglary, age has an independent effect. Also, Sampson
(1987), employing multivariate logistic regression on the British Crime Survey data ﬁnds
that the effects of lifestyles (frequency of going out at night) on stranger violence is

explained away by community characteristics like family disruption and heterogeneity

15

but has a strong positive and independent effect on stranger theft. On the other hand,
Smith and J arjoura (1989) ﬁnd that after controlling for a number of neighborhood
variables, single-parent households still have signiﬁcantly higher risk of victimization.

These examples Show the importance of introducing multivariate analysis.

16

CHAPTER III
THE CONTEXT OF STUDY SITE

As noted earlier, most of the research on criminal victimization is done in the
United States, United Kingdom, Canada and other western industrialized context. Given
the recurrent ﬁnding that various contexts and domains may result in different patterns of
criminal victimization, there is little knowledge on whether the variables that had been
used to account for criminal victimization in the western industrialized contexts also
predict victimization in non-westem, developing contexts. This is especially true since
individual role expectations, family structures, daily routine activities, people’s lifestyles,
interactions among neighbors, social organization of communities, formation of urban
areas and the relationship of public formal institutions to its citizens, are all products of
unique social, political, economic, cultural and historical forces in a particular society.
The present study

This research aims to expand the inquiry on criminal victimization in the
Philippine context. Given that no previous research on criminal victimization had been
conducted in the Philippines, this research tests representative variables from lifestyle-
exposure, routine activities and social disorganization theories and determine whether
they are capable of explaining victimization in a different context. This research is also
interested in establishing which among these alternative theories and speciﬁc measures
have the strongest applications in the Philippines.
Socio-political and historical context

In order to contextualize the different demographic, routine activity and social
disorganization variables that are included in this study, a brief description of the

country’s people, culture, history, politics and government is provided. Differences with

17

the western developed countries, especially the United States, are emphasized in order to
highlight how these social forces affect the exogenous variables, which in turn, determine
victimization risk. The predictions of the variables are guided by the theories but
reformulated to reﬂect Philippine local conditions.

The Philippines is a developing nation of 85 million people and almost twice the
geographic size of the state of Michigan. It is currently the 13‘h most populous country
and its gross domestic product (GDP) is 56"h in the world. It is located in Southeast Asia
and shares many of the Asian values, like communalism, ﬁlial piety and respect to
traditional authorities (Pe-pua and Protacio-Marcelino, 2000). However, western
inﬂuences are also apparent— it is the only predominantly Christian nation and the oldest
democracy in Asia and adopts English as its official and business language. It is
composed of three major islands— Luzon, Visayas and Mindanao, with many islets
totaling 7,107. There are eight major indigenous languages spoken in the different
regions of the country, but linguistic variations in the provinces are also common with a
total of 87 dialects. The original inhabitants of the Philippines are Malayo-Polynesian or
“brown” people scattered in small disparate communities called barangays (originally
balangay or “boat people”) (Agoncillo, 1960; Constantino, 1975; Jocano, 1998). These
baranganic communities are close family-knit systems that laid the familial nature of
Philippine polity.

Prior to the coming of the Spaniards in 1565, the small baranganic communities
were slowly integrated to the Sultanate of Sulu, a powerful and busy center of Islamic
trade with Malaysian and Indonesian archipelago (Constantino, 1975). This integration

to the Islamic sphere of inﬂuence has taken its roots in many areas in Mindanao. Many

18

Philippine historiographers believe that this could have been the beginning of the
formation of the Philippine state (Constantino, 1975).

This Islamic integration and state formation was cut short by Spanish colonization
which lasted for 333 years. The Spanish colonizers super-imposed a centralized
government co-managed by a civil-military and a frailocracy (or management by ﬁiars).
Religion was both used as a tool of assimilation and control. An “encomienda” system, a
feudal setup where large tract of lands were given to Spanish barons in exchange of their
support to manage “Indio” areas, was established (Agoncillo, 1960). In order to facilitate
the colonial regime, the heads of the baranganic familial structures were co-opted to
become part of the lower rung of the colonial administration. This co-opted class,
through intermarriages, become known as the “principalia”, and in the 19th century had
become a distinct social and economic force in the colonial society (Agoncillo, 1960;
Constantino, 1975). A divide and rule tactic was likewise employed where one ethnic
group would be pitted against the other. For example, revolts among Tagalog of
Southern Luzon would be quelled by reinforcements from Cebuano of the Visayas. More
importantly, however, was the demonization of the “Moro” (Muslim) in the South which,
for three hundred years, were not fully placed under Spanish control (Agoncillo, 1960;
Constantino, 1975). Such demonization ingrained a cultural wedge between the Muslim
Malays and Christian Malays, which, from time to time, erupt into open conﬂicts.

In the middle of the 19th century, the principalia class, now known as the
“illustrado” became conscious of their social, political and economic power in the
colonial society (Agoncillo, 1960; Constantino, 1975). They were exposed to the ideas of

European Enlightenment and called for reforms. Identifying with their Indio beginnings,

19

but wanting to maintain their social and economic privileges, they called for, among
other things, equal treatment with the Spaniards, educational access to all inhabitants,
separation of church and state, representation in the Spanish Cortez (legislature) and for
the recognition of the Philippines as a province of Spain (Agoncillo, 1960; Constantino,
1975). The reform movement of the principalia class eventually failed but laid the
foundation for the more aggressive revolution of the Indio masses which called for
independence. This turn of the century (1898) revolution against Spain, the ﬁrst
indigenous revolution against a colonial European power in Asia, was believed by many
Philippine historiographers to be the birth of the nation. This revolution bridged the gap
between the Indio masses and the illustrado class, the different ethnic groups, and for the
ﬁrst time, the term “Filipino” a term initially meant for Spaniards born in the Philippines,
became a unifying word for inhabitants of the Philippine islands (Constantino, 1975).

The birth of the Philippine nation, however, was short lived. At the time Filipino
revolutionaries declared their independence and promulgated their own constitution, the
United States of America declared war with Spain. In the beginning, Filipino
revolutionaries and the Americans were allies, but when Spain ceded the Philippines to
the Americans, the Philippine-American War ensued which lasted for more than a decade
(Constantino, 1975).

The American colonization lasted for nearly half a century and laid the foundation
of the modern Philippine state. Due to the strong undercurrents of the Philippine
revolution, especially among the Indio masses, American policy centered on
incorporating the illustrado to the colonial administration (Hutchcroft, 1998). As such,

the American colonial policy reinstituted the same wedge that divided the Filipino elites

20

from Filipino masses. The Americans introduced the formal elements of democracy, like
suffrage and representation, however, these were initially offered only to wealthy
Filipinos (Agoncillo, 1960; Constantino, 1975). The Americans awarded large tracts of
land formerly owned by the Spanish friars to local elites. The creation of the Philippine
Assembly, where encomienderos or mestizo landholders from the provinces were
integrated in national politics as members of the state legislature, ensured the creation of
a socio-political power base independent of the national government (Hutchcroft, 1998;
Side], 1999). This political feature is very important part of state transformation. Many
historiographers credit this policy to the institutionalization of the “power of the
strongman” or the predatory and particularistic power of national, regional, provincial or
municipal elites over the ﬁrnds, resources and spoils of a particular office of the
government. This strongman phenomenon has been described by scholars as “bossism”
(Sidel, 1999) “oligarchf’ (Hutchcroft, 1998) and similar to the “cacique” democracy that
characterized many of the Latin American countries.

The Philippines was formally granted political independence by the Americans in
1946, after the country was leveled off during World War II. Immediately after,
Philippine elites embarked to industrialize the countryside. Many of the landed elites
diversiﬁed their businesses to the manufacturing sectors (Rivera, 1994). Major cities like
Manila, Cebu in the Visayas and Davao in Mindanao begun to swell in population.
However, the rent-seeking behavior of the local elites, already well entrenched during
this period, necessitated the need for control of the political system which made
economic policy making highly politicized (Hutchcroft, 1998). Import substitution

industrialization (ISI), for example, meant the protection by the government of local

21

industries against foreign competition to some favored elites (Rivera, 1994). While many
countries in Asia developed in a fast pace during this period, the intra-elite struggles kept
Philippine economy stagnant. As such, industrialization had taken roots only in few
cities like Manila, Cebu and Davao and most of the countryside were still predominantly
agrarian. Traditional agrarian culture, like emphasis on the role of extended families,
ﬁlial piety, communal living and patriarchy (Pe-pua and Protacio-Marcelino, 2000) cc-
existed with modern lifestyles.

In the late 60’s and early 70’s, there were many disenfranchised sectors in the
Philippines. Students were restive. They were questioning the elitist political and
economic system, as manifested by wide socio-economic divide between the rich and the
poor, and called for massive societal change. Students’ ranks eventually populated two
of the most enduring insurgencies in the world— the communist New People’s Army and
the Muslim separatists in Mindanao (De Quiros, 1997). The Philippine “strongman”
Ferdinand Marcos tried to restore order and imposed Martial Law in 1972. He attempted
to eliminate the political and economic power of the local oligarchy and instituted a
strong control over the national and local affairs (De Quiros, 1997). This
experimentation to authoritarianism proved damaging— it centralized governmental
corruption, it undermined the legal foundations of the democratic polity and it destroyed
the criminal justice processes (De Quiros, 1997). The prevalence of Marcos “cronies,”
who took over the businesses of the oligarchs, coupled with an unsustainable debt-driven
policy, kept the economy further down the drain (De Quiros, 1997).

The formal elements of democracy were restored when Marcos’ 20-year

authoritarian rule ﬁnally ended through a “people power revolution.” However, the pre-

22

martial law economic and political set up was equally restored. The elites, who were
disempowered during Marcos’ reign, simply regained their old political and economic
base. As such, the intra-elite struggles, the concentration of economic, political and
social power to the hands of the few families, and the particularistic and predatory power
of the national, regional, provincial and municipal elites over the organs of government,
continue to deﬁne Philippine polity (Hutchcroft, 1998). For instance, the immediate past
president was deposed from ofﬁce through a people power revolt due to charges of
corruption and the current President continues to face problems of legitimacy on charges
of massive cheating in previous elections.

This brief historical sketch provides for a contextual background for many of the
variables that are included in this study. It provides a meaningful point of departure on
how the exogenous variables are affected by these societal and cultural forces.

The current state of Philippine neighborhoods

Local neighborhoods are characterized by a vibrant social relationship among the
members. Strong kinship and friendship networks continue to supplant the weaknesses
of the formal structures of the society. Informal sectors are prevalent in every sphere.
For example, the unavailability and inaccessibility of government and private business
ﬁnancing necessitates the prevalence of depending on ones’ friends and kin for loans,
educational expenses, and medical care and old age security.

Local neighborhoods are also characterized physical concentration of political,
economic and cultural activities. A “sentro” or center, where governmental and religious
ofﬁces, businesses and other instrumentalities are located, continues the Spanish tradition

of zoning for easy control (J ocano, 1975). Households located nearer the center usually

23

wield political, administrative and economic power in the community. Households and
individuals located farther the center are usually the members of the less afﬂuent groups.
Contrary to the location of the poor, usually racial minorities, in the inner cities in the
United States (Wilson, 1996), the poor and the disadvantaged in the Philippines are
usually located in the outskirts of the cities.

Despite the weaknesses of governmental institutions, residents of local
neighborhoods continue to depend on the services these institutions offer. Access to
these services, however, is usually availed of by using personal connections in the
particular ofﬁces. Business permits ﬁom the city government, for example, can be easily
availed if one is well connected in the business permit ofﬁces. As such, residents in local
neighborhoods are usually aligned in partisan groups designed to capture political power.
Inna-neighborhood conﬂicts may arise in the conduct of these political exercises.
Furthermore, life in the local neighborhoods is a combination of traditional agrarian
culture and modern western lifestyles. “Hilot” or quack doctors, for example, work side
by side with medical health professionals in the provision of basic services like child
delivery. These traditional practices continue to persist because they serve as an

alternative to the more expensive health services.

24

CHAPTER IV
DATA AND METHODS

The data used for this research was retrieved from the Inter-University
Consortium for Political and Social Research at the University of Michigan. The original
owner of the data is the Social Weather Stations (SWS), a respectable independent
polling agency in the Philippines. The SWS conducts quarterly surveys on Philippine
economic and social conditions and had been consistent provider of information
independent of governmental statistics. For this particular dataset, respondents were
asked about perceptions of criminality, victimization experiences and feelings of safety.
This survey was conducted in the fourth quarter of 1995 (from November 22 to
December 22) and covers four major geographical areas (Metro Manila or the National
Capital Region, Balance of Luzon, Visayas and Mindanao).

There were 1200 households randomly selected in this survey. In each household,
three persons were interviewed face to face: the household head and two other adult
members (more than 18 years of age) using the language/dialect that they understand.
The household head was asked about household socio-demographic characteristics, the
ﬁrst adult was asked about experiences of household members on victimization and
household activities and lifestyles, and the second adult was asked about the
neighborhood and other socio-political issues. The answers of the respondents were
combined. The unit of analysis therefore is the household. Though the survey provided
demographic attributes for the three respondents, in the analysis, the demographic traits
of the household head is employed since it is more theoretically relevant to the present

study. Compared to other members, the demographic traits of the heads of the household

25

have more effect (guardianship and attractiveness) on the victimization risk of the
household members.
Sampling Scheme

A multi-stage probability sampling was used in selecting the adult respondents:
for Metro Manila, there were three stages; and for the rest of the major areas, there were
ﬁve stages (see codebook for full description). For Metro Manila’s ﬁrst stage, using the
latest list of electoral precincts as sampling frame, 60 precincts were allocated to the 17
component cities/municipalities in proportion to their population size. The precincts
were selected randomly within each city/municipality. At the second stage, 5 households
were chosen from each sample precinct by an interval sampling method that was based
on a precinct map and guided by a right-coverage rule— streets, pathways, and
households on the right take precedence. The head of the selected household constituted
a sample unit of the population of household heads while all household members
constituted the sample units of the general population. At the third stage, two adults (a
male and a female) were taken in each household using separate random selection tables.
One respondent (male for odd-numbered questionnaires; female for even-numbered ones)
answers Questionnaire “A” and the other answers Questionnaire “B”.

For the rest of the country, the provinces served as the common ﬁrst stage unit.
Using updated population ﬁgures, 10 provinces in Balance of Luzon and 5 each in
Visayas and Mindanao were chosen with probability proportional to population size. For
the second stage, three cities/municipalities were drawn within each sample province,
with probability proportional to size. For the third stage, to get an urban sample, thirty

sample precincts were allocated among the selected cities/ municipalities proportional to

26

their urban population. In each sample city/municipality, the allocated sample precincts
were chosen by simple random sampling. To get a rural sample, 1 to 2 barangays (l in
Balance of Luzon and 2 each in Visayas and Mindanao) were selected in each sample
city/ municipality by simple random sampling. If a city/municipality was entirely urban,
another one with rural barangays was drawn with probability proportional to population
size, and from this, the required number of barangays was drawn (See codebook for more
complete discussion on methods).

Dependent Variables

Given past assertions that aggregating the victimization experience may mask
variations among crimes (Miethe et al, 1987), this study differentiates between property
and violent victimization. Property victimization includes incidents of house break—in,
herein referred to as burglary, pickpocketing and camapping (i.e., motor vehicle theft.)1
Violent victimization includes incidents of assault and stabbing and usually involves
physical disputes. As such, these two kinds of criminal victimization could provide us
with an understanding of their unique opportunity structures.2 Respondents were asked
whether they or any member of their household experienced any of the crimes for the
past 6 months prior to the survey. The variables were coded 1 for “YES”, 0 for “NO”,
making dichotomous dependent variables.

Table 1 shows that overall criminal victimization is a rare event. This is similar to
other studies that surveyed the respondents’ risk of criminal victimization (Miethe et al,
1987; Mustaine and Tewksbury, 1998; Sampson, 1987; Sampson and Wooldredge, 1987;
Smith and J arjoura, 1989; Velez, 2001). For this sample, property victimization occurred

in almost one out of ﬁve or 18.7 percent of the sampled households. Violent

27

victimization is even a rarer event. Table 1 shows that 5.4 percent of the households
report that one of their household members had been a victim violent physical

victimization.3

28

Table 1: Variable Coding and Descriptive Statistics, SWS Survey 4"I Quarter. 1995

 

 

I. Dependent Variables
a. Victim of property crime

b. Victim of physical violence
11. Independent Variables

Demographic Variables
a. Age of household head

b. Sex of the household head

 

c. Education level of household head

(1. Level of afﬂuence

e. Parental structure of the household

Routine and Lifestyles
f. Household composition

g. Work location

h. Spending pattern

i. Number of Portable items

j. Going out

Variable Coding Percentage
0=no 81 .3
l=yes 18.7
O=no 94.6
l=yes 5.4
Range = 18-95

Mean = 47

Standard Deviation: 14.8

l8-29= 11.7
30-54= 56.6
55-up= 31.7
1=Female 17.8
0=Male 82.2
1= Some college 27.7
0= Less than college 72.3
O= Afﬂuent 13.6
1= Poor 61.3
2=Very Poor 25.1
1=One parent 22.2
O=Two parent 77.8
O=Two or less adults 54.1
1=Three to four adults 30.9
2=Five adults and more 15.0
1= Working outside of home 29.5
0=Near or inside of home 70.5
1=Did spend 27.4
O=Did not spend 72.6
Range= 0-8

Mean= 2.25
Standard Deviation= 1.5

1=3 or more appliances 42.2
0= 0-2 appliances 57.8
Range= 0-4

Mean= .88
Standard Deviation= .88

0=Never 37.6
1=One 43.6
2=2-4 times 18.8

29

Table 1, (cont ’d)

 

k. Watch television

1. Perception of Criminality

m. Overseas worker
n. Homeownership
Contextual variables

0. Locale

p. Regional Area

q. Population mobility

r. Social cohesiveness

3. Barangay Justice

t. Satisfaction to law enforcement

 

Variable Coding Percentage
1=Light TV viewer 42.4
0=Heavy TV viewer 57.6
Range= 0-4
Mean= 1.60
Standard Deviation 1.17
1= Crime rose 16.3
0: Fell; did not change 83.7
1=With Overseas 9.4
=Without Overseas 90.6
1=Homeowner 74.9
0= Non-homeowner 25.1
1=Urban 62.5
0= Rural 37.5
O=Metro Manila 25.0
1=Luzon 25.0
2=Visayas 25.0
3=Mindanao 25.0
1=High 36.7
0=Low 63.3
Range= 0-7
Mean= 5.53
Standard Deviation: 1.73
1=High 90.7
0=Low 9.3
Range= 0-4
Mean= 2.45
Standard Deviation= .88
0= Low 18.2
1=Moderate 20.6
2=High 61.2
Range= 0-8
Mean= 4.54
Standard Deviation: 1.72
0=Low 23.4
1=Moderate 33.0
2=High 43.6

30

Independent variables and Hypotheses

Following the victimization literature, a set of demographic, lifestyles and
routines and contextual variables are included in the analysis. There are ﬁve household
demographic variables, nine lifestyle and routine variables and six contextual variables,
for a total of 20 variables examined in this research (see Table 1 for list, recoding and
summary statistics and Appendix 1 for listing of original items and deﬁnitions).
Demographic Variables

The ﬁrst demographic variable is the age of the household head. This variable is
measured on an interval-ratio scale by an item that asks the respondents their actual ages.
Table 1 shows that for this sample, the age of the household head ranges from 18 to 95
and the average (mean) age is 47 years old. For cross-tabulations purposes, the actual
ages were also categorized in three groups 4 (See table 1).

Most research ﬁnds negative relationship with age and victimization (Cohen et al,
1981; Cohen and Cantor, 1981; Hindelang, et a1 1978; Kennedy and Forde, 1990; Lee,
2000; Sampson and Lauritsen, 1990; Smith and Jarjoura, 1989). This is so as older
people may have a lifestyle that is more home-oriented and thus provide greater
guardianship to family members and to household properties. By extension, Hindelang et
a1 (1978) and Sampson and Wooldredge (1987) predict that households with older
household heads will experience lesser risk of burglary. This prediction also conforms to
the Filipino notion of the traditional respect for older people (Pe-pua and Protacio-
Marcelino, 2000, Williams and Domingo, 1993) which may further the guardianship of
households with older household heads. Thus, as the age of the household head

increases, risk of property and violent crimes should both decrease.

31

 

The second demographic variable is sex of the household head. This is a
dichotomous variable and male is coded “0” and female is “1”. Table 1 shows that
female-headed households constitute 17 .8 percent of the sample, slightly higher than the
ofﬁcially recorded 12.2 percent in the 2000 Philippine Census.

Male-headed or the traditional households are found to be less prone to
victimization on account of the guardianship they provide to the household (Sampson,
1985). Social disorganization theory also predicts that male-headed households in the
community provide stronger institutional control than female-headed households
(Sampson, 1985). This same prediction is applicable to the Filipino context given the
high regard given by Filipinos on traditional, usually male-headed, families (Alcantara,
1994). As such, male-headed households are expected to have a negative association
with both property and violent crimes.

The third demographic variable is the education of the household head. This
variable makes use of a single item that asks respondents their highest educational
attainment. This item was dichotomized to “some college” as 1 and “less than college”
as 0.5 Table 1 show that 27.7 percent of the sample of the household heads has at least
“some college” education, reﬂective of the educational levels of the general Filipino adult
population. I

On an individual level, having some college education is negatively related to
household theft, positively related to personal theft with contact and personal larceny
with contact but not related to risk of burglary, after controlling for lifestyle and
community characteristics (Sampson and Wooldredge, 1987). In a household level,

education levels of the household heads may signify greater attractiveness for property

32

crimes. However, in violent physical victimization, the higher education for household
heads may signify greater guardianship over the household members. This is especially
true in the Philippine context where “college-educated” professionals may have greater
capabilities to mobilize resources against people that may physically transgress their
individual members. Thus, it is hypothesized that individuals from households where the
head has higher levels of education will experience lesser risk of violent physical
victimization.

The fourth demographic variable is the level of afﬂuence of the household. This
variable corresponds to a single item in the questionnaire where the interviewers
designated a “socio-economic status” for the respondent (Class AB, C, D and E6) per
interviewers’ observation.7 “Classes AB and C” were collapsed and designated as
category 0 to indicate the “afﬂuent households.” Class D is designated as category 1 for
“poor households” and Class E is designated as category 2 indicating “very poor
households.” Table 1 indicates that 25.1 percent and 61.3 percent of the sample are
households considered by the interviewers as “very poor” and “poor” respectively,
whereas, only 13.6 percent of the households are considered “afﬂuent households.” This
sample generally reﬂects the highly unequal distribution of wealth in the Philippine polity
(Gerson, 1998).

The level of afﬂuence of a household had been associated with victimization
though in varying directions. Afﬂuent households were positively related to burglary and
vehicle theft (Kennedy and F orde, 1990) but negatively to assault and robbery (Kennedy

and Forde, 1990; Miethe et al, 1991; Miethe et al, 1987). Other research using family

33

income as a measure of afﬂuence ﬁnds that it has a parabolic relationship to burglary
(Cohen and Cantor, 1981; Cohen, Kluegel and Land, 1981).

In the Philippine context, the predictions are crime speciﬁc: positive association is
expected with property crimes on the account of greater attractiveness of the household
and the household members. However for violent physical victimization, a negative
relationship is predicted. Similar to the effect of “some college education,” more afﬂuent
households exhibit greater guardianship for their household members. Economically
well-off families are more politically and socially integrated in the communities which
provides them with capabilities to mobilize the particularistic powers of the state, like the
police and the judiciary (Hutchcroﬂ, 1998; Side], 1999) thus potentially giving their
household members greater protection.

The ﬁfth demographic variable is the parental structure of the household. This
item asks whether the household head is single, married, separated, widowed or lives
with a partner. Household heads, which designated themselves as “married or live in” are
coded O to mean two-parent structure and “all others” are coded as 1 for one-parent
structure. Two-parent households are predicted to provide greater guardianship
compared to one-parent households.8 Table 1 shows that 77.8 percent of the sample is
from a two—parent households, while 22.3 percent of the sample is from one-parent
households.

Research conducted in United Kingdom by Sampson and Wooldredge (1987)
reported that two-parent households are negatively related to risk of personal theft with
contact, but no effect on risk of personal larceny after controlling for lifestyle and

community variables. On the other hand, Kennedy and F orde (1991) report that in

34

Canadian cities, two-parent households are not related to burglary and vehicle theft but
negatively related with assault and robbery. Finally, a research by Smith and Jarjoura
(1989) in the United States report that after controlling for a number of neighborhood
variables, one-parent households still have signiﬁcantly higher risk of victimization. This
suggests that role expectations on the parental structure vary in each country thus
affecting levels of victimization. In the Philippines, given the traditional and religious 9
emphasis given on “complete” families (that is, with husband and wife, non-divorced), it
is expected that non two-parent structures will face greater economic and social
constraints. Thus, a one-parent familial structure is expected to correlate with higher
victimization.

Routine-lifestyle variables

For routine activities and lifestyle variables, this research employs seven items
that are traditionally used as measures. Considering, however, the imperative to include
reﬁned measures that tap on the speciﬁc socio-cultural context, two additional variables
are included. These variables are guided by the predictions of the theories.

The ﬁrst traditional routine and lifestyle measure is household composition,
which is deﬁned as the number of adults in the household. For this item, households with
“two or less adults” are designated as 0, households with “three to four adults” are
designated as l and households with “ﬁve or more adults” are designated 2.10 From
Table 1, we can see that 54.1 percent could be considered a “nuclear family” household;
30.9 percent could be considered “moderate-extended” household, while the remaining

15 percent of the sample has more than ﬁve adults in the household and could be

35

considered as “over-extended” families. The last two categories reveal the prevalence of
the extended nature of Filipino families.

As noted earlier, the number of adults in the household has been predicted in the
western context in conﬂicting directions: routine activities theory employs this as a
measure of guardianship, while household crowding is used as a measure of family
disorganization. Smith and J arjoura (1989) ﬁnd that different forms of criminal
victimization increase with the number of adults living in a household, whereas Massey
et a1 (1989) ﬁnd that the increase in the number of adults in the household is not related
to property victimization. In the Philippines, the number of adults in a household maybe
related to greater guardianship. Two or more households usually combine resources in
order to mitigate household expenses. Family members or kinfolks from the rural areas
usually stay with relatives in the urban areas while looking for work. Also, it is common
for households from the more afﬂuent classes to employ household helpers like family
drivers, gardeners and housemaids (Domingo and Asis, 1995). All these factors
contribute to greater guardianship in the household, thus predicting lesser victimization
risk for the household properties and its members. However, due to effects of household
crowding, it is argued that the guardianship capacity of adults in the household may
diminish when reaching a certain number and that “over extended households” may
experience higher risk of victimization (Meier and Miethe, 1993).

The second variable is the work location of the household head. This variable is
deﬁned whether the household head is working or staying “near or inside the home”
coded 0 or “working outside of home” coded 1. Household heads that are working in

“government” and “private companies” are assumed to be working outside of home,

36

99 6‘

whereas those who are “self-employed,” “working in the informal sector, paid family
worker” or “currently not working” are assumed to be working or staying at or near the
home. From Table 1, we can see that 29.5 percent of the sample is from households
where the head is working outside of the home. The prevalence of “non-working,” “self-
employed” and “employed in informal sectors” among the household heads is indicative
of the constraints in Philippine economy in 1995.

The locus of work is a traditional measure of exposure and guardianship. For
example, in the original formulation by Cohen and F elson (1979), they found that
increases in the prevalence of household heads working or staying out of home in post
World War 11 America were related to increases in different victimization. Also, Massey
et a1 (1989) report that having a job outside the home is signiﬁcantly related to household
theft, while Kennedy and Forde (1990) ﬁnd “fulltime work in the outside” is signiﬁcantly
related to burglary and vehicle theft. In the Philippines, working or staying near or at the
home will be related to lower victimization on account of the visibility of “istambay”
(from the word standby or unemployed) in providing guardianship to the household and
its members.

The third variable is the spending patterns of the household. This variable
includes an item that asks the spending pattern of the household for the past three months.
It asks respondents whether the household spent for new appliances, for improvements of
the physical structures of the house and for investment in business ventures. Coding for
this variable is 0 for households that “did not spend” on any of the three items and l for

household that “spent” for any of the items. Table 1 show that only little more than a

quarter (27 .4 percent) of the households was able to spend money on any of the items for

37

the past three months, again an indication of the trying times for the Filipino households
in 1995.

Spending patterns are traditional measure of attractiveness especially among
burglars (Miethe et al, 1987). It is expected that households that spent on the items from
the list have higher risk of property victimization. This same prediction applies to the
Philippine context.

The fourth variable is the number of appliances in the household. This item asks
whether the household owns appliances from the list. The list includes radios, television
sets, cellular phones, refrigerators, washing machines, microwave ovens, pagers/beepers,
and personal computers. Table 1 show that ownership of the appliances ranges from 0 to
8 with a mean ownership of 2.25 appliances. An alternative coding mechanism for cross
tabulation purposes also Show that 57.8 percent of the households have no more than two
of the above listed items and 42.3 percent of the households have more than 3 of the
items (see Table 1).

The number of appliances has been used as a measure of attractiveness of the
household to potential burglars (Cohen, 1980). Lightweight durable goods had been
associated with property crimes (Cohen, 1980) while expensive goods in the household is
related to burglary independent of contextual factors (Miethe and McDowall, 1993).
However, Sampson and Wooldredge (1987), using the British Crime Survey data, report
that number of appliances is not related to household burglary and household theft. In the
Philippines, ownership of appliances is usually associated with material wealth and social
afﬂuence. Households that own many appliances could therefore be viewed as socially

and politically integrated to the community. As such, though their properties could be

38

attractive targets for burglars and thieves, the household members could be more
protected against physical violations.

The ﬁfth variable is the frequency of going to busy places. This asks the
respondents if they and other members of the household engaged themselves in the
following activities for the past six months: watching a movie, attending concerts, and
going to parks and festivals. These activities entail going to busy commercial areas.
Going to these activities is coded 1, while not attending is coded 0 having an aggregate
range of 0 to 4. Table 1 show that average (mean) attendance to any of these activities is
less than 1. Recoding of the variables shows that 37.6 percent of the household have
“never” engaged in the in any of the activities, 43.6 percent of the households have gone
to only “one” of these activities and 18.8 percent of the households have gone to “two or
more” of the said activities.

“Going to busy places” is traditionally construed as a measure of exposure to
potential offenders, however with mixed results. Miethe and McDowall (1993) report
that going to busy places is positively related to violent criminal victimization but not
signiﬁcantly related to burglary and property victimization. For this research, it is
expected that going to busy places increase risk of violent criminal victimization.

The sixth variable is the household activity. This item asks respondents how
much time family members spend watching television. As Messner and Blau (1987)
argued, television viewing is predominantly a household activity. This variable shall be
coded as 1 “Light TV viewers” for households that spend less than one hour of TV

watching a day and those households without television sets and 0 “Heavy TV viewers”

39

for households that spend more than an hour a day. Table 1 shows that TV viewing is
almost split into half.

Heavy television viewing, as a household activity, is hypothesized to result to
lesser exposure to crime for household members. Households with heavy TV viewing
hours usually have members that stay at home thus limiting the risk of being exposed to
motivated offenders. It also increases guardianship to household properties. As such
heavy TV viewing activity among households is expected to be related to lower risk of
pickpocketing, break-in of houses and violent physical victimization. Same predictions
will be observed in the Philippine context.

The seventh variable is perception of criminality in the neighborhood. This
variable uses two items that asked the respondents about their perceptions on present
conditions of criminality and whether criminality will rise in the future or not. These two
items were aggregated (alpha reliability is .6478) to construct the perception of crime
index that range from O to 4, with higher scores signifying more perceived criminality in
the neighborhood. Table 1 shows that the mean perception of criminality is 1.60. For
cross-tabulation purposes, the crime perception index were dichotomized where 1
indicates that “crime rose/crime will rise” and 0 as either “crime fell/will fall and be the
same””. Table 1 indicates that less than one/ﬁfths or 16.3 percent of the respondents
think that criminality rose or will rise in their neighborhood.

Massey et a1 (1989), Sampson and Lauritsen (1990) and Wooldredge et a1 (1992)
used perception of safety as a measure of proximity to potential offenders, however with
differing results.12 Massey et a1 (1989) ﬁnd that “sense of proximity to potential

offenders” have no signiﬁcant effect on property victimization, while Wooldredge et a1

40

(1992) ﬁnd that “generally feel safe” is negatively related to property and physical
crimes. On the other hand, Sampson and Lauritsen (199) report that areas that
respondents feel have high crime rates are positively related to total victimization. The
same prediction will also be applied in the Philippine context.

As stated earlier, there is a need to incorporate speciﬁc measures that capture the
unique socio-cultural conditions of the research area. For this reason, the following
variables are included— overseas worker in the household and homeownership.
Overseas worker in the household as a variable is included because of the current social
phenomenon of “Filipino Diaspora” (Gonzalez, 1998). Almost 10 million Filipinos or
more than 10 percent of the total population are working abroad. This phenomenon
keeps the country’s economy aﬂoat, yet had been speculated to have damaging social
consequences. It is thought to disrupt the family, making the children left behind as
“orphans”, thus potentially increasing levels of juvenile delinquency and crime (Asis et
al, 2004). For this research, “overseas working” will be related to criminal victimization
in two ways. One, it is a measure of less guardianship. Filipinos who go abroad are
usually adults, particularly mothers (Asis et al, 2004). As such, it is predicted to result in
higher risk of physical violent victimization for household members. Secondly,
“overseas working” can be construed as a measure of more attractiveness. Families with
overseas workers are usually characterized with material affluence, that is, more
appliances, bigger houses, nicer clothes and more travel (Bryant, 2005). Thus, it is
hypothesized that households with family members working abroad are attractive targets

for property offenders. Households “with overseas” members will be coded as 1, 0 for

41

“without overseas”. Table 1 indicates that 9.4 percent of the households currently have a
member or members who are working abroad, which is reﬂective of the general trend.

One of the pressing problems in Philippine society is homelessness. It is
estimated that around 60 percent of the total population do not own their own homes,
especially in the urban areas (National Statistical Coordination Board, 2000). Homeless
households tend to squat in government and private lands and construct makeshift
dwellings in close proximity to other homeless households. Arguably, Filipinos who do
not own their own houses have diminished capacity over their household properties. '3
Home ownership could therefore be related to criminal victimization through less
guardianship. This is conceptually similar to the measure of “type of dwelling” (single
detached, duplex or multiple housing) used in the western context. More crowded
dwelling type is signiﬁcantly related to property victimization (Kennedy and Forde,
1990; Massey et al, 1989; Stahura and Sloan, 1988). Households that are “homeowners”
are coded l and “non-homeowners” are coded 0. Table 1 shows that in this sample, 74.9
percent of the households own their homes, whereas only 25.1 percent do not.14
Contextual variables

Following social disorganization theory, the succeeding contextual and
neighborhood variables are included. First is the locale, which measures whether the
household is located in “urban area” coded as 1 or “rural area”, coded as 0 (for detailed
deﬁnition of “urban” and “rural” areas as conceptualized in the Philippine context, see
Appendix 1). Table 1 shows that in this sample, 62.5 percent of the households are

located in the urban areas.

42

Social disorganization theory predicts that urban areas have reduced capacities to
control their communities due to greater anonymity, thus making them more prone to
victimization. Urban areas are also characterized by higher population density (Kennedy
and Forde, 1990), housing density (Sampson and Wooldredge, 1987) and cultural
heterogeneity (Sampson, 1987), factors that decrease the communities’ informal and
formal controls. Lee (2000) for example ﬁnds that cities with bigger populations have
more assault victimization than cities with lower populations. Central city residents are
also found to experience greater risk of burglary victimization (Cohen and Cantor, 1981).
In the Philippines, the same predictions will be applied. However, one must note that the
formation of urban centers is a product of push and pull factors (push because of rural
poverty and pull because of opportunities offered by urban areas) (Murakami et a1, 2005).
Also, the de-urbanization phenomena, where businesses are transferring from urban to
suburban or rural areas, have not occurred in the Philippines (Murakami et al, 2005). As
such, in the Philippines, the urban poor tend to be located in the peripheries of the urban
areas.

The second social disorganization variable is the regional area location of the
household. In the Philippines, there are three major islands (Luzon, Visayas and
Mindanao), which have unique histories of state formations (Constantino, 1975) and
religious assimilations (V itug and Gloria, 2000). Mindanao, for example, has some
provinces where Muslims comprise a sizeable minority of the population. Cultural and
religious heterogeneity may induce conﬂict among the residents in a community as
manifested by decades of tensions in Muslim-Christian affairs. The National Capital

Region (NCR) or Metro Manila is a small and high-density region that is considered as

43

the political, economic and cultural center of the country. It has a population of 12
million people, which comprises 13 percent of the national population. It is a melting pot
of the different ethnic groups who come to the region for employment and other
opportunities.

This regional variation is introduced to account for differing levels of social
organization. The Metro Manila area, being the country’s melting pot, and Mindanao,
due to cultural and religious heterogeneity, may experience more strains in achieving
social harmony. The Balance of Luzon and Visayas, due to their relative religious and
cultural homogeneity, may experience improved mechanisms to attain common
community goals. This mechanism propels the prediction that incidents of crime
victimization will be higher in the Metro Manila and Mindanao compared to Balance of
Luzon and Visayas. Table 1 Show that each regional grouping has 25 percent of the
sample.15

The third social disorganization variable is population mobility. This item asks
where respondents spent most of their childhood and has four response options (“in this

9 6" 9 66'

town or city’ In a different town or city’ In a different province” and “outside the

6619,

Philippines”). '6 The answers will be coded “0” for in “in this town or city” and for
all others. A higher score therefore means more mobility. Table 1 reveals that 63.3
percent of the respondents have been staying in the same town or city since their
childhood years.

It is hypothesized that lower mobility is related to less crime victimization on the

account that people who stay in their area develop social ties and supervise the person

and property of each other. Miethe et a1 (1991) using a longitudinal study, report that

44

cities with higher initial rates of mobility has larger net increases in robbery and burglary
rates, while Sampson (1985) reports that higher mobility has positive association with
violent crime rates. The Philippines tend to be characterized by generation of families
located in the same town or city. The constraints in transportation, due to deﬁciencies of
a national highway system and the island nature of the country, prohibit the geographical
mobility among the population. This population stability facilitates the familiarization
and trusts among community members. Respondents that report to have stayed in the
same area since childhood will therefore be negatively related to risk of property and
violent victimization.

The fourth social disorganization variable is social cohesiveness. This variable
includes two items on the questionnaire, which asks the respondents “how close they are

”17

to other people in their village/neighborhood and their town/city. The response

,9 6‘ 9, ‘6

options are “very close close not very close” and “not close at all.” The two items.
are aggregated to form the social cohesiveness index that ranges from 0 to 8 and a mean
score of 5.53. Higher scores will therefore mean that household members are more
socially attached to their communities. For cross-tabulation purposes, the social
cohesiveness scale was dichotomized where scores ranging ﬁ'om “O to 2” was designated
as “low cohesiveness” (0) and scores ranging from “3 to 8” was designated as “high
cohesiveness” (1). Table 1 shows that 90.7 percent of the sample household members
consider themselves to be socially “close” to their neighborhoods.

Socially cohesive neighborhoods are theorized to have lower crime victimization

(Lee, 2000; Smith and Jarjoura 1989). However, as noted earlier, an ethnographic study

had shown that socially cohesive and integrated communities may also foster the

45

development of crimes by giving support to local criminals who are part of the
communities’ social structures, like family networks (Patillo, 1998). In the Philippines,
given the extended nature of Filipino families and the premium given to particularistic
interests, social cohesiveness may work on both ways— one, as a form of social
guardianship to households and their members, and two, as a safe haven for offenders
who are pursued by state authorities and hide under the community’s social and family
networks.

The ﬁfth social disorganization variable is barangay justice efﬁcacy. This
variable includes an item that measures the satisfaction levels of the respondents on their
“barangay justice system.” The response options on the barangay justice item are “very
satisﬁed/satisﬁed” coded 2, “maybe satisﬁed/maybe not satisﬁed” coded 1, “not
satisﬁed/not at all satisﬁed” coded 0. Higher scores will mean more satisfaction on the
performance of the barangay system. Table 1 show that the mean score is 2.45 and
majority of the household respondents (61.3%) express satisfaction to the workings of the
barangay in their communities.

As introduced earlier, the barangay is the smallest political, social, religious and
cultural unit in the Philippines. The barangay has a “Lupon ng Tagapayapa” or a peace
and order council that mediates conﬂicts among neighbors, settles minor disputes and
imposes community sanctions on erring members (Silliman, 1985). The barangay
likewise maintains cleanliness on the surroundings like garbage collections and
beautiﬁcations. These informal social control capabilities of the barangays conform to

the parochial component of the systemic approach of social disorganization theory

46

(Bursik and Grasmick, 1993). Barangays that have capabilities to solve their own
problems are thought to have reduced incidences of criminal victimization.

The last social disorganization variable is satisfaction to law enforcement. This
variable aggregates two items that asks the respondents on their satisfaction rating on the
police and military18 in their respective places (alpha reliability is .6656). Table 1 show
that the average rating for law enforcement is 4.54 and that majority of the respondents
express average (33.0 percent) and high (43.6 percent) satisfaction to the police and the
military.

Satisfaction rating to the law enforcement indicates level of respondents’ approval
to the public formal control over their communities (Velez, 2001). In the systemic
formulation of the social disorganization theory, public formal control institutions like the
police and other criminal justice agencies, affect levels of criminal victimization through
mechanisms of social guardianship (Bursik and Grasmick, 1993; Hunter, 1985).
Residents who perceive that their law enforcement is effective may be deterred from
committing crimes and thus lowering levels of criminal victimization.

Method of Analysis

This study examines the variables that predict property and violent victimization
in the Philippines based upon western traditional theories (lifestyle-exposure, routine
activities and social disorganization). This research paper also examines the independent
effects of each of the variables. For example, will the effects of demographic variables
remain when routine activity variables are introduced? Also, will routine activity

variables remain signiﬁcantly related to property and violent victimization when

47

contextual variables are controlled? Finally, this research paper examines whether the
effects of predictors are crime-speciﬁc or not.

Toward this end, there are two levels of analysis conducted in this research. First,
a bivariate analysis using cross-tabulations was performed for both dependent variables
with all the independent variables. The Pearson Chi Square Test of Independence was
used to determine whether two pairs of variables were signiﬁcantly related or not. To
determine the strength of association, the Phi output (for two-level categorization) or
Cramer’s V output (for more than two-level categorization) was reported (Bachman and
Paternoster, 2004). An analysis of variance (AN OVA), using proportion (risk of
victimization), was also performed to determine whether the differences among groups
are signiﬁcant.

Second, multivariate analyses using logistic regressions were performed on each
of the dependent variables. Correlations among the independent variables are low thus
precluding a problem on multicollinearity (See appendix 3). The multivariate logistic
regression analysis is speciﬁcally designed for dichotomous dependent variables with the
possibility of skewed frequency distributions (Kennedy and Forde, 1999; Lee, 2000;
Miethe et al, 1987; Sampson and Wooldredge, 1987, Wooldredge et al, 1992). In each of
the dependent variable, three models were estimated. In the ﬁrst model, only the
demographic variables were used to predict the occurrence of the dependent variable. On
the second model, routine-lifestyle variables were included, controlling for the
demographic variables. On the third model, the social disorganization variables were
included on top of the other set of variables.19 The logistic regression coefﬁcient (b), the

Standard Errors (SE) and the Exponentiated Beta (Exp b) values were all shown to

48

compare the direction and strength of the relationships among the variables. Additional
information about Wald values was included when warranted. Also, in each of the
models, the amount and signiﬁcance level of the log likelihood output were presented to
show the improvements of the model from the previous models. Following Miethe et a1
(1987), Lee (2000) and other scholars on criminal victimization, this paper reports values
where the p <. 10. Finally, Pseudo R, using the Nagelkerke value was presented to Show
how much of the variation is explained by the models (Bachman and Paternoster,

2004).20

49

CHAPTER V
RESULTS

Tables 2 and 3 present the summary statistics of the bivariate relationships of the
dependent variables on property victimization and violent victimization, respectively.
Presented are percentages of victimization in each category. These percentages are also
called “risk of occurrence” when presented in the form of proportions (Bachman and
Paternoster, 2004).

Variables associated with Property Victimization

Table 2 shows that certain types of households are more at risk of property
victimization than others. Afﬂuent households have signiﬁcantly higher risk of property
victimization compared poor and very poor households. This is most especially true for
motor vehicle thefts, where the risk experienced by afﬂuent households is more than
eight times greater than the risk faced by the less afﬂuent groups. Also, households
where the head have some college education face a risk of property victimization that is
60 percent more compared to households where the head has less than college education.
Both variables were consistent with the predictions of traditional Western theories.

However, age and sex of the household head and the parental structure of the
household did not perform as expected. Household members whose heads are older are
more prone to property victimization than from households whose heads are younger.
Household members from female-headed families also have a lesser risk of property
victimization compared to their counterparts. Though the differences are not statistically

signiﬁcant, the direction is opposite what is predicted from traditional theories.

50

Table 2 Percent of Property Victimization by
Demographic, Routine-lifestyle and Contextual Characteristics

Variables

Demographic Variables
Age of the Household Head

Sex of the household head

Education of the Household head

Afﬂuence of the Household

Parental Structure of the Household

Routine-Lifestyle Variables
Household Composition

Work Location of the Household
Head

Spending pattern

Number of Appliances

Going out to busy places

All Cases

18—29
30-54
55-up
Chi square(Cramer’s V)

Female
Male
Chi square (Phi )

Some College
Less than College
Chi square (Phi )

Affluent

Poor

Very Poor

Chi square(Cramer’s V)

One parent structure
Two parent Structure
Chi square (Phi )

Two or less adults
Three to four adults
Five adults or more

Chi square(Cramer’s V)

Outside of home

Near or inside the home
Chi square (Phi )

Household Spend
Not Spend
Chi square (Phi )

0-2 appliances
3- more appliances
Chi square (Phi )

Never

Once

2-more

Chi square(Cramer’s V)

51

Number
of Cases
1200

140
678
382

214
986

332
868

163
736
301

267
933

649
371
I80

354

846

871
329

693
507

451
523
226

Percent
Victimized
18.7

15.7
17.8
21.2
2.724 (.048)

18.2
18.8
.034 (005)

25.6

16.0
l4.543***
(.110)

35.0

16.2

15.9
33.026***
(.166)

19.5
18.4
.148 (.011)

16.6
18.6
26.1
8.326“ (.083)

20.6

17.8
1.264 (.032)

18.0
20.4
.861 (.027)
14.3

24.7
20.735***
(.131)

15.3
18.9
24.8
8.953"
(086)

Table 2 (cont ’d)
Variables

Television Watching

Perception of Criminality

Overseas Worker

Homeownership

Contextual Variables

Locale

Regional Area

Population Mobility

Social Cohesiveness

Barangay Justice

Satisfaction with Law Enforcement

Phi or Cramer’s V Values are in parenthesis for the strength of association

“*p< .01 "p < .05

All Cases

Light TV viewer
Heavy TV viewer
Chi square (Phi )

Crime fell; no change
Crime rose
Chi square (Phi )

With Overseas
Without Overseas
Chi square (Phi )

Homeowner
Non-homeowner
Chi square (Phi )

Urban
Rural
Chi square (Phi )

Metro Manila

Balance of Luzon
Visayas

Mindanao

Chi square(Cramer’s V)

Low Mobility
High Mobility
Chi square (Phi )

Low Cohesiveness
High Cohesiveness
Chi square (Phi )

Low Efﬁcacy

Average Efﬁcacy

High Efﬁcacy

Chi square(Cramer’s V)

Low Satisfaction
Average Satisfaction

High Satisfaction
Chi square(Cramer’s V)

*p<.10

52

Number

of Cases
1200
509
691

1004
196

113
1087

899
301

750
450

300
300
300
300

760
440

112
1088

218
247
735

281
396
523

Percent
Victimized
18.7

15.3

21.1
6.505"
(.074)

16.4
30.1
20.178***
(.130)

25.7
17.9
4.023"
(.058)
16.8
24.3
8257*"
(.083)

25.2

7.8
56.230***
(.216)

30.3
13.0

12.7

18.7
40.354!m
(.183)

15.7
23.9
12.359***
(.101)

22.3
18.3
1.087 (.030)

22.9
23.9

15.6
12.318***
(.101)
26.7

22.0

11.9
30.746***
(.160)

Among the routine-lifestyle variables, ownership of household appliances and
perception of crime in the neighborhood have the strongest associations with property
victimization. Households which own three or more appliances, experience one and half
times greater risk than households which own two or fewer appliances. Household
members who perceive that criminality rose or will rise in their neighborhood also report
greater risk of property victimization. Going to busy places, overseas working and
homeownership are also signiﬁcantly related to property victimization, though the .
strengths of associations are weaker. All these variables are associated with property
victimization in the expected directions.

Two routine-lifestyle variables are associated with property victimization but in
the opposite direction. Contrary to the guardianship claims of TV watching (Messner and
Blau, 1987), households that engaged in heavy TV watching hours face a one and a half
times the risk of property victimization compared to households that are light TV
viewers. However, scrutinizing the speciﬁc signiﬁcance of television ownership explains
this discrepancy. In the Philippines, TV ownership is associated with ownership of other
household appliances. As such, it is TV ownership, not TV watching per se, that makes
the household attractive targets for burglars and thieves.

Household composition performed in the opposite direction. The risk of property
victimization for households with two-or-less-adults (the nuclear families) is almost
identical to the risk faced by three-to-four-adult household (the moderately extended
families). This generally conforms to guardianship effects of having “moderate” number
of adults in the household. However, the risk of property victimization signiﬁcantly

increases for households that have more-than-ﬁve adults. This suggests that the

53

guardianship accorded by the number of adults in a household pertains only to
households that are not “overly extended”. Property victimization risk increases
dramatically for households with ﬁve or more adults, perhaps on account of the
criminogenic effect of household crowding (Miethe et al, 1987). More than ﬁve adults in
the household may result in the invitation of more people to come inside the house, thus
providing potential burglars and thieves cues on the attractiveness of the household and
its properties. Finally, there is very slight variation between risk of property
victimization and work location of the household head and spending patterns of the
household. Though the difference is toward the prediction of the routine activities
theory, the differences are not statistically signiﬁcant.

More signiﬁcant variations in property victimizations were recorded for social
disorganization attributes of the households. Households located in the Metro Manila
regional area appear most at risk. Slightly more than 30 percent of these households had
been a victim either of a burglary, a pickpocketing or a motor vehicle theft during the
past six months. Metro Manila households experience more than twice the risk of
property victimization compared to the risk faced by households located in the Balance of
Luzon and Visayas regional areas. The households in the Mindanao regional area also
experience a higher risk of property victimization compared to Visayas and Balance of
Luzon.21

Of all the variables, locale has the strongest association with the risk of property
victimization (See Table 2). Households in the urban areas experience more than three

times the risk of property victimization compared to households in the rural areas.

Households which are less mobile, situated in efﬁcient barangays and which have higher

54

satisfaction to law enforcement in their areas, also face lesser risk of property
victimization compared to their counterparts. All these conform to the traditional
expectations of the social disorganization variables.

However, social cohesion is not signiﬁcantly associated with property
victimization. Households that report to be closely attached to people in their
neighborhoods share equal risk levels to households that report otherwise. This is
contrary to traditional predictions that households in socially cohesive neighborhoods
should have lower risk of property victimization. In fact, an opposite direction, though
non-signiﬁcant, is observed in the occurrence of household burglary. This pattern
conforms to an ethnographic study in a Chicago neighborhood where the denseness of
social ties among community members may promote, and not stiﬂe, the violation of
ofﬁcial laws as long as it will advance the parochial needs of the members (Patillo,
1998)

Variables associated with Violent Victimization

Table 3 shows that, when compared to property victimization, fewer variables
predict violent victimization of household members. None of the demographic variables
are signiﬁcantly related and only seven out of the remaining 15 variables are associated
with violent victimization of household members. For routine-lifestyle variables, only
three are signiﬁcantly related to risk of violent victimization. Households whose
members perceived that criminality rose or will rise in the future in their neighborhoods
reported a 250 percent increase in the risk of violent victimization than households whose
members reported that criminality declined or stayed the same in their neighborhoods.

Households that have a member who works abroad also experience greater risk than

55

Table 3 Percent of Violent Victimization by

Demographic, Routine-lifestyle and Contextual Characteristics

Variables

Demographic Variables

Age of the Household Head

Sex of the household head

Education of the Household head

Afﬂuence of the Household

Parental Structure of the Household

Routine-Lifeswle Variables

Household Composition

Work Location of the Household

Head

Spending pattern

Number of Appliances

Going out to busy places

All Cases

18-29
30-54
55-up
Chi square(Cramer’s V)

Female
Male
Chi square (Phi )

Some College
Less than College
Chi square (Phi )

Afﬂuent

Poor

Very Poor

Chi square(Cramer’s V)

One parent structure
Two parent Structure
Chi square (Phi )

Two or less adults
Three to four adults
Five adults or more

Chi square(Cramer’s V)

Outside of home

Near or inside the home
Chi square (Phi )

Household Spend
Not Spend
Chi square (Phi )

0-2 appliances
3- more appliances
Chi square (Phi )

Never

Once

2-more

Chi square(Cramer’s V)

56

Number
of Cases
1200

140
678
382

214
'986

332
868

163
736
301

267
933

649
371
180

354

846

871
329

693
507

451
523
226

Percent
Victimized
5.4

5.0
5.5
5.5
.054 (007)

'15
51)
2.157 (.042)

5.7
5.3
.084 (.008)

6.7
4.5
7.0
3.245 (.052)

7.1
4.9
1.936 (.164)

51
539
:56
.337(rn7)

6.8

4.8
1.821 (.039)

53
5t;
.lb4 (010)

4L8
63
1.373 (241)

(12
(L8
53
.971 (028)

Table 3 (cont 'd)

Variables

Television Watching

Perception of Criminality

Overseas Worker

Homeownership

Contextual Variables

Locale

Regional Area

Population Mobility

Social Cohesiveness

Barangay Justice

Satisfaction with Law Enforcement

Phi or Cramer’s V Values are in parenthesis for the strength of association
it
p < .05

aeap< _01

All Cases

Light TV viewer
Heavy TV viewer
Chi square (Phi )

Crime fell; no change
Crime rose
Chi square (Phi )

With Overseas
Without Overseas
Chi square (Phi )

Homeowner
Non-homeowner
Chi square (Phi )

Urban
Rural
Chi square (Phi )

Metro Manila

Balance of Luzon
Visayas

Mindanao

Chi square(Cramer’s V)

Low Mobility
High Mobility
Chi square (Phi )

Low Cohesiveness
High Cohesiveness
Chi square (Phi )

Low Efﬁcacy

Average Efﬁcacy

High Efﬁcacy

Chi square(Cramer’s V)

Low Satisfaction
Average Satisfaction

High Satisfaction
Chi square(Cramer’s V)

*p<.10

57

Number

of Cases
1200
509
69 1

1004
196

113
1087

899
301

750
450

300
300
300
300

760
440

112
1088

218
247
735

281
396
523

Percent
Victimized

5.4

3.3

6.9

7441*" (.079)

4.4
10.7
12.833***(.103)

8.8
5.1
2.869" (.049)

5.5
5.3
.008 (.003)

7.3
2.2
14.341*** (.109)

9.7
3.3
3.7
5.0
15.013*** (.112)

4.6
6.8
2.664* (.047)

8.0
5.1
1.654 (.037)

5.0
7.3
4.9
2.594 (.046)

8.9
6.6
2.7
15.326*** (.113)

households that do not have an overseas worker. Both variables are associated with
violent victimization in the expected direction of routine activities theory.

Similar to property victimization, TV watching is associated with violent
victimization but in the opposite direction. Households, which spend longer TV viewing
hours, experience twice the risk of victimization than households that spend less TV
viewing hours. This is in contrast to the less exposure claims of TV watching. This
ﬁnding seems to suggest that given the nature of television viewing to be primarily a
household activity (Messner and Blau, 1987), longer television viewing is associated with
the risk of having a household member being violently victimized but in the home
domain. Finally, there is little variation between risk of violent victimization and
household composition, work location of the household head, spending patterns of the
household, number of appliances, going out to busy places, and homeownership.

Similar to property victimization, households located in the Metro Manila
regional area appear most at risk of violent victimization. Nearly 10 percent of these
households reported that a household member had been a victim of assault, stabbing or
other forms of physical violence in the past six months. Metro Manila households
experience almost thrice the risk of violent victimization compared to households located
in the Balance of Luzon and Visayas regional areas and one a half times the risk
compared to households located in Mindanao regional area.

Locale is also strongly associated with the risk of violent victimization.
Households in the urban areas experience almost four times the risk of violent
victimization compared to households in the rural areas. Households which are less

mobile and have higher satisfaction with law enforcement in their areas also face lesser

58

risk of violent victimization compared to their counterparts. All these conform to the
traditional expectations of the social disorganization theory. However, social cohesion
and barangay justice efﬁcacy of the neighborhood are not signiﬁcantly associated with
violent victimization.

Summary of the Bivariate Relationships

The results of bivariate relationships answer some of the key issues that are of
interest to this research. First, it is clear that some of the western traditional variables
have the potential to explain victimization risk even in a non-western developing country
setting. This is especially true for the social disorganization variables. On a bivariate
level of analysis, 14 and 7 out of the 20 variables tested were signiﬁcantly related with
property victimization and violent victimization, respectively. Second, the difference
between the opportunity structures of property and violent victimization, as reported in
the western literature, is also manifested in this research.

However, the bivariate analysis also indicates that many of the western traditional
variables were not associated with criminal victimization in the Philippine context. This
is especially true for violent victimization. Finally, as can be seen by the Phi and
Cramer’s V values, the associations of all the independent to the dependent variables are
weak. Locale stood as the strongest among the predictors yet its Phi value of .216 is
conventionally regarded a “weak” association (Bachman and Paternoster, 2004).
Multivariate Analysis

The data presented from Tables 2 and 3 show that victimization risk for
households and its members vary with demographic characteristics of the household head

and the lifestyles and routines of the household members. Victimization risk also varies

59

with the neighborhood context in which the households are embedded. While many of
the bivariate patterns of association are consistent with the proposed expectations of
western theories, a multivariate method is necessary to identify which variables are
related to victimization once other characteristics of the household or neighborhoods are
controlled (Smith and Jarjoura, 1989).

Property Victimization

Tables 4- A, 4- B and 4- C present three models predicting property victimization.
Model 1 (Table 4-A) shows only the effects of demographic characteristics. Model 2
(Table 4-B) shows the effect of routine-lifestyle variables while controlling for
demographic variables and Model 3 (Table 4-C) shows the effect of contextual variables
while controlling for demographic and routine-lifestyle variables.

Overall, Model 1 predicts about 4.5% of the variation in property victimization, as
indicated by the Nagelkerke Pseudo R value. It is also an improvement of ﬁt compared
to the model with the intercept only, as shown by the Model Chi Square value. This
indicates that demographic variables, as proposed by Hindelang et al (1978) explain a
small proportion of the variation in property victimization.

From the logistic regression coefﬁcients (b) of each variable in Model 1, we can
see that age and sex of the household head and the parental structure of the household are
not statistically signiﬁcant predictors of property victimization when other demographic
variables are controlled. This is similar to results reported in the bivariate levels.
Educational level of the household head and the level of afﬂuence of the household
remained signiﬁcantly related to property victimization even when other demographic

variables are controlled. The odds of victimization faced by households where the head

60

have some college education increased by 42% compared to households whose head have
less than college education. The odds of property victimization for poor and very poor
households are signiﬁcantly lower (about half) than the odds of property victimization
faced by afﬂuent households (reference category).

4-A Regression Model

Independent

Variables

(Ref category: Male)

(Ref category: Less Than College) 352“

-.815**"'

(Ref category: Two parent)

-l.346""

 

Model 2 of Table 4-B shows the effect of routine-lifestyle variables, while
controlling for the demographic variables. Model 2 is a signiﬁcant improvement from
Model 1, as shown by the reduction in the log likelihood ratio by 36.909. Furthermore,
Model 2 was able to explain 9.3% of the variation in the property victimization as shown
by the Pseudo R value. This indicates that household members’ routines and lifestyles

affect property victimization independent of the demographic variables.

61

Moreover, the independent effects of the demographic variables are reduced when
routine-lifestyle variables are controlled. It is noticeable that the effect of education of
the household head is not Signiﬁcant in this model. This indicates that the routine-
lifestyles of households where the head had some college education, not college
education per se, that make the household prone to property victimization.

The contribution of the afﬂuence of the household is also reduced to more than
half, as indicated by the Wald values. After controlling for routine-lifestyle variables,
there is no longer a signiﬁcant difference between the odds of property victimization
between the more afﬂuent and poor households, though the very poor households are still
signiﬁcantly less victimized compared to the afﬂuent households.

The logistic regression coefﬁcients indicate that four of the routine-lifestyle
variables predict property victimization, controlling for other measures. A unit increase
in the perception of the residents that crime rose or will rise in the future in their
neighborhood translates to 27% increase in the odds of property victimization occurring.
Homeowners, on the other hand, have reduced odds of property victimization by more
than half compared to their non-homeowner counterparts. An increase in the number of
adults in the household and frequency of going to busy places likewise translated to
increased odds of property victimization.

Three of the routine-lifestyle variables lost signiﬁcance in the multivariate
analysis. Household appliances, television watching and overseas working, are all
associated with property victimization in the bivariate levels, but are no longer predictive

in a multivariate level. Additionally, work location of the household head and spending

62

patterns are not predictive of the odds of property victimization. These are similar to the

bivariate results shown earlier.

 

Table 4-B Logistic Regression Model of Property Victimization
Model 2
Independent Variables B SE Exp (1))

 

 

 

Demographic Variables
Age of Household Head
Female

(Ref category: Male)

 

 

 

Some College
(Ref category: Less Than College)

Afﬂuence of the Household **
( Poor)
( Very Poor) -.637*** .242 .529
(Ref category: Afﬂuent )
One Parent
(Ref category: Two parent)
Routine-lifestyle Variables
Composition .037* .021 1.03 8
Work Outside of Home
(Ref category: Near or Inside )

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Family spent
(Ref category: Family did not )
Appliances
Going out to Busy places .162* .084 1.176
Light TV watching
(Ref category: Heavy TV Watching)
Perception of Criminality 238*“ .067 1.269
With Overseas
(Ref category: W/O Overseas )
Homeowner
(Ref category: Non-Homeowner) -.467*** .177 .627
“”5““ -2.569*** .572 .077
Model 2
-2 Log Likelihood 1084.189
Model Chi Square 36.909***
Pseudo R Square (Nagelkerke) .093

 

** p<.05 ***p<.01

63

 

Model 3 in Table 4- C shows the effects of contextual variables in the predictors
of property victimization while simultaneously controlling for demographic and routine-
lifestyle variables. As can be seen by the reduction in the log likelihood ratio by 44.545,
Model 3 is a further improvement of ﬁt than the previous two models. This generally
conforms to the notion that neighborhood characteristics where households are embedded
independently affect property victimization. The Model Chi Square value also indicates,
that taken together, social disorganization variables have the greatest potency in
explaining property victimization in this data set. As such, inclusion of these variables
was able to explain 14.9% of the variations in property victimization.

The logistic regression coefﬁcients indicate that locale, population mobility,
barangay justice efﬁcacy and satisfaction with law enforcement signiﬁcantly affect
property victimization, over and above some of the effects of demographic, routine-
lifestyles and other contextual variables. Households located in urban areas have more
than twice the odds (Exp B= 2.653) of being a property victim compared to households
located in rural areas. Households whose members are relatively new to the
neighborhoods have one and half times odds of being a victim of burglary, pickpocketing
or car theft than household members who report to be staying in the same neighborhood
since childhood. Also, one unit increase in barangay justice efﬁcacy index translates in
the reduction of the odds of property victimization by 20%. An increase in the
satisfaction of the respondents on the local enforcement in their area translates to a
reduction of the odds of property victimization by almost 10%. These suggest that
structural factors like urbanization and population mobility and community processes like

barangay justice efﬁcacy and public formal controls both have direct effects on property

64

victimization. Moreover, afﬂuence of the household and perception of criminality still
remain signiﬁcant even when contextual variables are introduced, indicating that these
variables have strong independent effect. However, regional area and social
cohesiveness are not related to odds of property victimization when other variables are
controlled.

In summary, households which have the following characteristics are signiﬁcantly
more prone to property victimization— those that are afﬂuent, those which perceived that
crime rose or will rise in the future, those located in an urban area, those that have high
mobility, those that have low satisfaction with their barangay justice and those that have
low satisfaction with their law enforcement. Among all these, location in an urban area

stands as the strongest single predictor of property victimization.

65

(Ref category: Less Than College)

976*"

 

66

Violent Victimization

Tables 5-A, 5-B and 5-C present the three models predicting violent
victimization, in similar fashion as Table 4. Model 1 (Table 5-A) shows only the effects
of demographic characteristics. Overall, Model 1 predicts only 1.4% of the variation in
violent victimization, as indicated by the Nagelkerke Pseudo R. It is not an improvement
of ﬁt when compared to the model with the intercept only, as shown by the very low
Model Chi Square value. This indicates that, in this data set, demographic variables
like— age, sex and educational levels of the household head, and the afﬂuence and
parental structure of the household— are not signiﬁcantly related to violent victimization

of household members.

 

Table 5-A Logistic Regiession Model of Violent Victimization
Model 1
Independent Variables B SE Exp (b)

 

 

 

 

 

 

1. Demographic Variables
Age of Household Head
Female

(Ref category: Male)

Some College
(Ref category: Less Than College)

Afﬂuence of the Household

( Poor)

( Very Poor )

(Ref category: Afﬂuent )
One Parent

(Ref category: Two parent)
Constant -2.569*** 11.347 .077

 

 

 

 

 

 

 

 

 

 

 

 

Model 1

-2 Log Likelihood 499.727

Model Chi Square 5.727

Pseudo R Square (N agelkerke) .014
** p<.05 ***p<.01

 

 

 

 

 

 

 

67

Model 2 of Table 5-B shows the effect of routine activities variables, while
controlling for the demographic variables. Model 2 is a signiﬁcant improvement from
Model 1, as shown by the reduction in the log likelihood ratio by 25.088. Furthermore,
Model 2 explains 7.4% of the variation in violent victimization as shown by the Pseudo R
value. This indicates that household members’ routines and lifestyles affect the odds of
violent victimization, independently and more strongly than the demographic variables.

The logistic regression coefﬁcients indicate however that only two of the routine-
lifestyle variables predict violent victimization, all other things being equal. A unit
increase in the perception of the residents that crime rose or will rise in the future in their
neighborhood translates to 38% increase in the odds of violent victimization occurring.
Meanwhile, the odds that a violent victimization occurs are more than twice for heavy
TV viewers than light TV viewers. As proposed earlier, this indicates that violent crimes
may be happening in the home domain.

Overseas working, which is signiﬁcant in a bivariate level, is reduced to non-
signiﬁcance when other routine-lifestyle variables are introduced. This indicates that it is
the routine-lifestyles of households with overseas workers that make them prone to
violent victimization, and not overseas working per se. Finally, the afﬂuence of the
household becomes signiﬁcant when routine-lifestyle variables are introduced, indicating
that routine-lifestyle variables specify the impact of household afﬂuence on violent
victimization. Members of more afﬂuent households are more prone to violent
victimization than members of less afﬂuent households who engage in similar routines

and lifestyles.

68

 

Table S-B Logistic Regression Model of Violent Victimization
Model 2
Independent Variables B SE Exp (b)

 

 

 

Demographic Variables
Age of Household Head
Female
(Ref category: Male)
Some College
(Ref category: Less Than College)

Affluence of the Household **"‘
(Poor) .856 .541 2.354
( Very Poor ) .022 .444 1.022
(Ref category: Afﬂuent )
One Parent
(Ref category: Two parent)
Routine-Lifestyle Variables
Composition
Work Outside of Home
(Ref category: Near or Inside )
Family spent
(Ref category: Family did not )
Appliances
Going out to Busy places
Light TV watching
(Ref category: Heavy TV Watching) 771" 3'07 2-161

Perception of Criminality 324*" .110 1.383
With Overseas

(Ref category: W/O Overseas )
Homeowner

(Ref category: Non-Homeowner)
Constant -4.681*** 1.006 .009

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Model 2
-2 Log Likelihood 474.639
Model Chi Square 25.088***
Pseudo R Square (Nagelkerke) .074

** p<.05 ***p<.01

 

 

 

 

 

 

 

Model 3 in Table 5- C shows the effects of contextual variables in the predictors
of violent victimization while simultaneously controlling for demographic and routine-
lifestyle variables. As can be seen by the reduction in the log likelihood ratio by 19.048,
Model 3 is a further improvement of ﬁt than the previous two models. This generally
conforms to the notion that neighborhood characteristics where households are embedded

independently affect violent victimization. However, the Model Chi value indicates that

69

taken together, social disorganization variables have lesser potency in explaining violent
victimization, compared to routine-lifestyle variables. Still, the inclusion of these social
disorganization variables resulted to an increase in the explanatory power of the whole
model (Pseudo R= .118 or 11.8%).

The logistic regression coefﬁcients indicate that only locale and satisfaction to
law enforcement signiﬁcantly affect violent victimization, after controlling for different
demographic, routine-lifestyles and other contextual variables. Household members
located in urban areas have twice the odds (Exp B= 2.159) of being assaulted or stabbed
compared to household members located in rural areas. An increase in the perceived
satisfaction of the respondents with the local enforcement in their area translates to a
reduction of the odds of property victimization by almost 16%.

Population mobility, social cohesiveness and barangay justice efﬁcacy are not
related-to violent victimization, though the directions of their relationships are toward
what is predicted. This suggests that community processes may have little capability to
explain violent victimization in the Philippine context. Similar to property victimization,
regional area is also not a predictor of violent victimization in the full model, despite
being a very strong predictor in the bivariate level. Finally, households that are heavy
TV viewers and in close proximity to potential offenders still face signiﬁcantly greater
odds of property victimization, even when contextual variables are controlled.

In summary, households which have the following characteristics have members
who are more prone to violent victimization— those that perceived crime increased or
will increase in the future in their neighborhood, those that are heavy TV viewers, those

that are located in an urban area and those that have low satisfaction with their law

70

enforcement. Among all these, perception of criminality stands to be strongest single
predictor of violent victimization.22

Table Victimization

Less Than College)

 

7l

CHAPTER VI
DISCUSSIONS AND CONCLUSIONS

This study examines whether western traditional theories of criminal
victimization have the potency to explain the same phenomenon in a different social,
political, cultural and historical context. Seventeen representative variables that have
been found to be robust predictors of criminal victimization in the western societies were
selected for this study. In addition, three variables (overseas working, homeownership
and regional area) were included to represent the unique context of the study site. All
these 20 exogenous variables were guided by the predictions of lifestyle—exposure,
routine activities and social disorganization theories.

From the analyses that were conducted, the following themes emerged. First, the
different theories can be successfully operationalized in the Philippine context. For
example, overseas working is a phenomenon unique in the Philippines. It affects the role
expectations for parents in the households (Asis et a1, 2004). As such, its potential
contribution to victimization experience can be properly integrated through the
guardianship and attractiveness thesis of the routine activity theory. Similarly, regional
area, a major survey categorization in the Philippines, can be used to account for
victimization based on the cultural heterogeneity thesis of the social disorganization
theory.

Second, from this successful operationalization, western traditional theories have
some potency in explaining victimization in the Philippine context. In fact, the recurring
ﬁnding in the Western literature regarding the difference of the opportunity structures of
property and violent victimization are replicated in this study. Speciﬁcally, this study

ﬁnds that in property victimization, social disorganization variables are the most potent

72

predictors, though demographic (household afﬂuence) and routine-lifestyles (exposure to
potential offenders) also affect property victimization. On the other hand, in violent
victimization, routine-lifestyle variables are the strongest predictors, with some
contributions from social disorganization variables (locale and public formal control) and
little effect from demographic variables. In simple terms, the results show that property
victimization in the Philippines can be determined by where Filipino households live,
while violent victimization are determined by what Filipino households do.

Third, similar to the concerns raised by many scholars (Miethe et al, 1987;
Sampson and Wooldredge, 1987; Smith and Jarjoura, 1989), there is a need to investigate
the variables in a multivariate analysis. In this study, many of the demographic and
routine-lifestyle variables are signiﬁcant in the bivariate levels; yet lose their explanatory
power once contextual variables are introduced in the model. For example, the concern
raised by many Filipino scholars about the devastating effects of overseas working,
where children left behind are supposed to suffer on account of becoming “emotional
orphans” (Asis, 2004) are not supported, when all other variables are included. These
ﬁndings suggest that it is the daily routines and lifestyles of members in the household
where there is an overseas worker; not overseas working per se, that makes the household
and its members prone to victimization. Without the multivariate analysis, therefore, the
relationships among the variables would be spurious. Similarly, the strong bivariate
effects of regional area, where Metro Manila households are found to be more at risk than
households in the three other regions in both property and violent victimization, are not
statistically signiﬁcant once other demographic and routine-lifestyle variables are

controlled. This suggests that the effect of regional area is rather spurious. Metro Manila

73

households tend to be more mobile, more afﬂuent, more educated and have more
appliances than their regional counterparts, which account for their various
victimizations.23

Fourth, the concern raised by some scholars on domain speciﬁcations (Maxﬁeld,
1987) is very relevant, especially when testing a theory outside the area of its conception.
This is emphasized by the opposite direction of the performance of TV watching. In the
United States, due to the prevalence of television sets in every household, it is almost
natural to conceive of TV watching purely on its effect on household guardianship (that
is, members stay longer at home) (Messner and Blau, 1987). In the Philippines, however,
many homes still do not have TV sets. TV ownership also signiﬁes material prosperity,
thus making the household attractive to burglars and thieves. Therefore, it is the TV
ownership, not longer TV viewing per se, that is associated with higher property
victimization.24

Finally, despite the numerous variables used to explain variations in victimization
risk, the Pseudo R values of the full models indicate that there are still much to be
accounted for. This is particularly true for violent victimization where only 11.8% of the
variation was explained. This could be due to four factors. One, there are still many
other variables that were not included in the models. For example, Sampson and
Lauritsen (1990) ﬁnd that deviant and offending behaviors are intimately linked to
victimization risk. However, due to the limitations of the dataset, these variables were
not included.

Two, many variables used in this research are indirect measures of the concepts

that were originally proffered. Household afﬂuence, usually measured in terms of annual

74

family income, was substituted by using the interviewers’ assessment of the respondent’s
dwelling. This indirect measure may explain the weak performance of household
afﬂuence in violent victimization of household members.

Three, in this research, the dependent variables are not fully speciﬁed. For
example, studies done in the western context have subcategorized stranger/non-stranger
violence (Sampson, 1987) and home, ofﬁce, school, leisure or other domains (Maxﬁeld,
1987). These speciﬁcations are necessary because each kind of victimization has unique
opportunity structures that could be masked when they are aggregated. In this research,
however, the dependent variables are not sub-categorized due to data limitations, thus
possibly hiding the effects of the predictor variables.

Four, the unit of analysis used in this research is the household. Most researches
done in the western hemisphere is individual victimization. The demographic traits of the
household head may have little capability in explaining the victimization risks of
household members. This is because victimization is usually an individual affront
(except in cases of burglary and other household crimes). This could possibly explain
why the demographic variables used in this research, such as age and sex, did not
perform as predicted.

However, the weak performance of the predictor variables, especially on violent
victimization, may also be attributed to the unique social, economic, political and cultural
dynamics of the Philippines. For example, in the United States, United Kingdom and
Canada, households headed by younger people, females and single-parents are vulnerable
to different forms of victimization (Sampson and Wooldredge, 1987; Smith and J arjoura,

1989). In the Philippines, as this data set had shown, these demographic variables (age of

75

household heads, sex of household heads and parental structure of the household) are not
signiﬁcant predictors of property and violent victimization of the household and its
members.

This could be possibly explained by the differences in how households are
organized. In the Philippines, there is a prevalence of extended family membership
where household help is readily available. The prevalence of household help may
attenuate the deﬁciencies faced by households headed by younger people, females and
single-parents. It is culturally encouraged, for example, that grandparents usually stay
with single-parent, female-headed and newly starting households. The lifestyles of these
supposedly vulnerable groups are attenuated thus making these kinds of households not
particularly susceptible to victimization.

Also, the contrary ﬁnding could be explained by the differing socio-economic
characteristics of female-headed households in the Philippines. Female-headed
households in the Philippines are not as vulnerable as their counterparts in the United
States. In the United States, female-headed households are strongly associated with
economic and social disadvantages like lower family incomes, dependency on social
welfare and having more dependent children (Wilson, 1996), which possibly explains
their higher risks of victimization. In the Philippines, female-headed households tend to
be smaller, (with an average of four members) and have higher family incomes compared
to male-headed households (National Council on the Role of the Filipino Women, 2005).

Similarly, the weak performance of social cohesion emphasizes the importance of
culture in the study of criminal victimization. In this data set, household members who

report to be closely attached to people in their neighborhoods and villages share equal

76

risk levels to household members who report otherwise (See Table 2). This is contrary to
predictions that households in socially cohesive neighborhoods should have lower risk of
criminal victimization. In fact, an opposite direction, though non-signiﬁcant, is observed
in the occurrence of household burglary.” This pattern conforms to an ethnographic
study in a Chicago neighborhood where the denseness of social ties among community
members may promote, and not stiﬂe, the violation of ofﬁcial laws as long as it will
advance the parochial needs of the members (Patillo, 1998). In the Philippines, while
many household members claim to be close to other residents in their neighborhood on
account of extended kinship, this strong social cohesion fails to inﬂuence levels of
victimization. While social cohesion may increase guardianship, it may likewise provide
a venue for networking of potential criminals, thus negating its positive beneﬁts.

These suggest that other forms of conceptualization may be formulated to account
for victimization in the Philippine context. While western theories are rich in their
predictions, they do not fully explain the phenomenon in a different setting. The cultural
uniqueness in family organization, the differing level of maturity of formal public
institutions, the prevalence of elitism and particularistic interests in political life, the
differing economic responses to urban situations (like absence of de-urbanization) and
the diverging meaning attached to particular symbols (like TV ownership) must be
incorporated in any attempt of conceptualizing criminal victimization in the Philippines.
Limitations and Suggestions for Future Research

This research is the ﬁrst empirical study on criminal victimization in the

Philippines. It provides preliminary ﬁndings on what predict victimization in a non-

77

western developing society. It tailor ﬁts the predictions of western theories in the local
Philippine context.

However, these ﬁndings are tentative and should be taken with caution. One, this
study makes use of a secondary data set, which was gathered not for the primary purpose
of studying victimization risk. As such, many of the items in the survey were indirect
measures and were not originally conceptualized to gauge the ideas propounded in the
western traditional theories. Future research may therefore improve immensely if the
data gathered are guided by more direct measures of the concepts. For example, in this
research, social cohesion is measured by how close the respondents are to other members
of the neighborhood. It is anchored on the respondent’s feelings. While this may be a
plausible measure of attachment to a certain neighborhood, this is rather arbitrary. Other
researchers captured social cohesion in practical and easy to verify acts, like the number
of times neighbors exchange food items, look out at each other’s house and supervise
each other’s children (Villarreal and Silva, 2006).

Second, future research may also beneﬁt by specifying clearly the different
dependent variables. For example, determining the speciﬁc correlates of stranger-
violence or school-domain violence may ﬁrrther the understanding of criminal
victimization in the Philippines. In this research, the domain of the victimization is not
fully speciﬁed, thus possibly confounding the speciﬁc effect of predictor variables.
Furthermore, focus on individual victimization, in contrast to household victimization,
would clarify whether demographic variables, such as age and sex, are indeed weak

predictors of violent crimes in the Philippine context.

78

Third, better measures of the contextual variables where the geographical/spatial
characteristics of the neighborhood should be included in succeeding researches. Instead
of simply aggregating the similar-characteristic neighborhoods (as what has been done
here), scholars should be able to incorporate neighborhood traits that may be independent
of the individual traits by members of the neighborhood.

Fourth, an improvement of the methodology used in this study should be in order.
Due to the highly skewed nature of the data, simple logistic regressions were used.
Though appropriate in this kind of analysis, this methodology fails to incorporate the
speciﬁc context where the individual respondents are nested. Identifying the attributes of
individuals in a particular neighborhood provides a clearer delineation of the
contributions of individual-level and neighborhood-level effects. More advanced
methodological tools, like Hierarchical Linear Modeling, which takes into consideration
the individual-neighborhood interactions, had been found to yield more powerful results
(Roundtree et al, 1994).

Finally, the western theories that predict criminal victimization needs to be re-
conceptualized when used to explain the same phenomenon in a different setting. Though
some variables are universal in application, the unique organizational, political and
familial culture of a particular study site may render some variables impotent. Concepts
from other academic disciplines may be integrated to the criminological theories. In the
ﬁeld of political science, for example, the unequal development of the state organs vis-a-
vis other competing interests resulted in the popularization of the “soft state” concept
(Lipset, 1960). This concept characterizes most developing countries; the Philippines

included (Hutchcroft, 1998). This alternative form of state formation, contrary to the

79

mature and established democracies of the western industrialized countries, may
necessitate the inclusion of variables not included in the study. In particular, variables
like personal connections to key people in government, number of friends and relatives
working in the criminal justice agencies, quality and breadth of contacts, etc— variables
that capture mechanisms that increase ones’ access to the dispersers of political power——

are variables that may explain the variations in the victimization experiences.

 

NOTES

‘ I initially intended to analyze these three property crimes separately. However due to the skewed
distribution of the data, some analyses are not very meaningful. An example is motor vehicle theft that
occurred in only 1.3% of the sample. Thus, I aggregated the three incidents together. However, to answer
the concerns raised by previous scholars that aggregation may mask the unique opportunity structure of
each crime type, 1 performed concurrent analyses where burglary and pickpocketing were disaggregated.
Generally, the results are similar; otherwise, their differences will be noted in the text. Finally, given the
assertions that social disorganization theory is best in describing total crimes (Veysey and Messner, 1999),
I further aggregated property and violent victimization. The results generally reﬂect the results in the
property victimization, considering that most crimes recorded in the dataset are property crimes. Thus,
reference to the total crimes will be made only when warranted.

2 As a further check, I also tried the alternative notion that camapping involves forcible taking of one’s car
and it may involve physical violence. This implies that it is better to consider it as a violent crime. When
aggregated, however, the results are not consistent when taking violent physical victimization alone,
suggesting differences in opportunity structures.

3 The occurrence of property and violent crimes is somewhat higher compared to what is commonly
reported in the international and cross-country literature. As such, I tried to cross-validate this data by
ofﬁcial measures from the Philippine National Police. The Philippine National Police reports that in 1995,
there is a total of “1 12.8 crimes per 100,000 population” (National Statistical Coordination Board, 2002)
which (surprisingly) is one of the lowest not only in Asia but the world as well. This wide discrepancy
could be attributed to the differences in reporting mechanisms. As such, I culled additional victimization
data from different data sets of the Social Weather Stations for the ﬁrst and second quarter of 1995 (and is

80

 

 

also retrievable from the ICPSR.) Risks of victimization based on surveys are consistent with the ones
reported here.

4 Some of the independent variables are categorized to illustrate their bivariate relationship with the
dependent variables. In the multivariate analysis, however, their original metric was used.

5 I also recoded education by designating “ﬁnish college” and “post-baccalaureate” as category 1 and
“some college” and lower levels of educational attainment as category 0. The result is basically the same. I
chose to use “some college, ﬁnish college and post baccalaureate” as a distinct group considering that these
levels all indicate that they ﬁnish High School. In the Philippines, most companies require at least some
college (72 credits) to be admitted for a job, thus theoretically meeting the requirement that the category
should express similar traits. Statistically, this grouping is better considering that it increases the observed
frequency in one of the categories.

6 As used in this survey, class categorization is based on the type of dwelling of the respondents.
Respondents residing fully furnished homes and located in more affluent neighborhoods are considered
either class A, B or C. Respondents residing in more decrepit areas and with houses not fully furnished are
considered either class D or E.

7 I initially intended to measure afﬂuence of the household using two items in the questionnaire: an item
determining the economic class of the household and an item that asks the “self rated poverty” of the
respondents. However, the reliability analysis shows an alpha of .4299 indicating that the two items are
measuring two different things. As such, I dropped the “self rated poverty” and used the economic class of
the household instead. I believe that this is a more objective measure of level of afﬂuence considering that
it is the trained interviewers that classiﬁed the class of the respondents. Compared to “self rated poverty”
this measure is prone to problems associated with social desirability, where rich respondents may “humbly”
declare that they are poor. This is consistent with the Filipino cultural mentality that disdains “yabang” or
bragging and promotes “pagpapakumbaba” or humility especially when asked about economic conditions
of the household (Enriquez, 1975).

8 Sampson and Wooldredge (1987) raise some concerns on where to categorize “widow.” For though
“widows” may share the one-parent structure of “separated” “divorced” and “single” households, they may
not share its vulnerabilities. Following the theoretical emphasis of two-parent guardianship however, I
categorized “widow” in the one-parent structure.

9 The Philippines is one of the few countries in the world that do not allow a divorce. This is due to the
strong opposition of the Catholic Church against any bill in Congress proposing to legalize it.

1° Three items and different coding variations were tested before I ﬁnally settled to use this coding
mechanism. First, I used the item “resident family size” (range 1 to 14) of the household were helpers,
family drivers and gardeners were included and categorized it into three groups (1-3 “small household
size”; 4-6 “average household size” and more than 6 as “above average household size”). Second, I made
use of the item “actual number of household members” where all household members like transients (that
is, relatives who temporarily stay) are included (range from 1 to 32) and categorized in two (1-6 “not so
crowded”; 7 and up “crowded”). I settled in using the third item where the “number of adults in the
household” was identiﬁed for two reasons: one, it is theoretically more consistent to the idea of
guardianship, and two, it more speciﬁes the relationship among household members, that is, an adult
transient is different from a young transient. In any case, the results reported herein are similar to the results
of other items.

11 I also used three-group categorization (0-1 as fell/will fall=0; 2 as did not change =1; 3-4 a rose/will

rise=2). Initial results show similarity when two-group categorization was used. To increase observed cell
frequencies, I used the two-group categorization.

81

 

‘2 Scholars generally raise concern on the validity of using this measure (Wooldredge et al, 1992). It is
possible that crime perception increased as a result of victimization experience. Given the cross-sectional
nature of the data, one cannot ascertain which of the variable cause the other. As such, I performed a
separate multivariate analysis without the perception of crime index. It did not dramatically change the
statistical signiﬁcance levels of the remaining variables in the model. This same approach was taken by
Wooldredge et al (1992).

13 It is noted that in the Philippines, there are households which squat in private or public lands and
perceive that they rightfully own their houses. While this is very plausible, what I am currently interested
in evaluating is the home ownership structure of the household. This is with the assumption that similar
home ownership structure will be in close proximity of each other (that is, owners will be in close
proximity with other owners, squatters with other squatters, etc).

1’ The over-representation of household owners versus renters and other arrangements could be explained
by two factors. First, the question distinguishes between house owners and lot owners and one therefore
could be a house owner but a lot renter. Lots are usually more expensive, as manifested by the fact that
only 47 percent of the households in this same sample own their lots. Second, there is a difﬁculty in
surveying squatters and other illegal settlers, which could possibly explain why non-home owners are
under-represented.

'5 Due to the nature of the sampling, Metro Manila is over-sampled and Balance of Luzon is under-
sampled. The current population distribution is: Metro Manila is 13%, Balance of Luzon is 44%, Visayas is
20.3% and Mindanao is 23% (Census on Population and Housing, 2000).

'6 Another item used in the questionnaire is the respondent’s length of residence in the present town and is
answerable by “less than a year” or “one year or more.” This item is more commonly used in the studies in
the western context. However, considering the nature of Filipino households to stay in the same locality for
long period of time, for example, in this survey only 2.2 percent of the total respondents answered to have
stayed in the present neighborhood for less than a year, the independent variable becomes highly skewed
and precludes making any meaningful analysis. As such, I opted to use a more strict deﬁnition of
population stability, that is, the respondent should be staying in the same locality since childhood. Thus, in
this research, older respondents who may have stayed in the same locality for the past 20 years or so, as
long as they spent their childhood in a different town, are considered more mobile than those who stayed in
the same neighborhood since childhood.

'7 Other items in the questionnaire include questions on how close the respondents are to people in the
province and to the country. I declined to include these items on the analysis for the reason that closeness to
the province and country are theoretically different to the closeness to village and towns, which may dilute
the community processes emphasized by the social disorganization theory. The alpha reliability of the two
items is .5720.

'8 In the Philippines, the military has a supportive police functions. The military is involved in the general
peace and order, anti-insurgency, and counter-criminal measures like kidnappings and drugs. (Filler, 2002)
As such, the Philippine Military is directly involved in the communities’ formal public control.

‘9 Separate analyses (not reported) were also conducted where routine-lifestyle variables and contextual
variables were introduced ﬁrst.

20 A more powerful tool, the hierarchical linear modeling (HLM), is more desirable in multivariate
researches. However, a major limitation of the dataset, where very few individuals are nested in the same
neighborhood, precludes the use of HLM.

2' To verify mean differences among the four regions, I performed an analysis of variance speciﬁcally for
burglary. The results are similar. See table below.

82

 

Cross-tabulation of Burglary victimization by Regional Area

Regional Area Non-victim Victim
Metro Manila 271(90.3%) 29(9.7%)
Balance of Luzon 289(96.3%) ll(3.7%)
Visayas 286(95.3%) 14(4.7%)
Mindanao 264(88.0%) 36(12.0%)

Total 1 l 10(92.5%) 90(7.5%)

Chi Square is 20.613“** *** p<.01
22 This ﬁnding should be taken with caution. As explained earlier, higher perception of criminality may be
a direct result of a previous criminal victimization experience. A longitudinal study would clarify more
speciﬁcally the effects of negative crime perceptions.
2’ I performed a cross-tabulation of regional area with the variables that are signiﬁcantly related to
victimization. Results show that Metro Manila residents are more mobile, more afﬂuent, more educated and

have more household appliances than their regional counterparts.

2’ To verify this, I made a cross-tabulation of TV viewing and household appliance. Results shows that the
two are variables are associated (Chi Square is 96.452***; p<.01).

25 This data was disaggregated from the total property crimes.

83

APPENDIX 1 LIST OF VARIABLES AND ORIGINAL CODINGS

DEPENDENT VARIABLES:

Now, we would also like to know your experiences and that of your immediate family
members regarding crime. In the past six months, have you or any member of your

family been a victim of

 

‘CYeSO9 or “N0,,)

? (Answerable by

l. PANDURUKOT O PAGNANAKAW NG PANSARILING KAGAMITAN
(Pick-pocketing/theft of personal property).
2. PAGPASOK O BREAK-IN SA TAHANAN (Break-in at respondent’s residence)
3. PAGNANAKAW NG KOTSE O SASAKYAN (Automobile theft)
4. PAMBUGBOG, PANANAKSAK O IBA PANG KARAHASAN (Assault,
Stabbing and other Physical violence)
5. Total number respondent’s family was victimized
INDEPENDENT VARIABLES:
1. Age of the head of the households (interval level measure).
2. Sex (“Male” or “Female)
3. Educational Attainment of the Household Head
Frequency distribution of Educational Attainment
Frequency Percent
1 No formal education 28 2,3
2 Some elementary 195 163
3 Completed elementary 200 16,7
4 Some high school 166 13.8
5 Completed high school 216 18,0
6 Some vocational 26 2.2
7 Completed vocational 37 3_1
8 Some college 127 10.6
9 Completed college 188 15.7
10 Post college 17 1,4
Total 1200 100.0
4. Afﬂuence of the household (Class of dwelling).
Frequency distribution of Class of dwelling
Frequency Percent
1 AB 33 2.8
2 C 130 10.8
3 D 736 61.3
4 E 301 25.1
Total 1200 100.0

84

5. Parental structure of the household (Civil Status of Household head)
Frequency distribution of the Civil Status of the Household head

Frequency Percent
1 Single 70 5.8
2 Married 916 76.3
3 Widow/widower 183 153
4 Separated 14 1.2
5 Living-in as married 17 1.4
Total 1200 100.0

6. Household composition (This variable is constructed by counting the total number
of adults in the household).
Frequency distribution of the number of adults in the household

Frequency Percent
1 Single 22 1.8
2 One adult/one child 14 1.2
3 One adult/two children 14 1.2
4 One adult/three or more
children 13 1'1
5 Two adults 90 7.5
6 Two adults/one child 117 9.8
7 Two adults/two child 130 10.8
8 Two adults/three or more 249 20.8
9 Three adults 62 5.2
10 Three adults with children 148 12.3
11 Four adults 41 3.4
12 Four adults with children 120 10.0
13 Five adults 27 2.3
14 Five adults with children 61 5.1
15 Six adults 13 1.1
16 Six adults with children 32 2.7
17 Seven adults 9 .8
18 Seven adults with children 16 1.3
19 Eight adults 2 .2
20 Eight adults with children 8 .7
21 Nine adults 2 .2
22 Nine adults with children 5 ,4
24 Ten adults with children 3 .3
25 11 adults 2 .2
Total 1200 100.0

7. Work location of household head. This variable is constructed from an item that
asks respondents of their working status: working in government; working in the
private sector; self-employed/informal sector; unpaid family worker; not working;
never worked before. Working in “Government” and in “Private” form one
category, the rest, another category.

85

10.

11.

Frequency distribution of the working location of the household head

Frequency Percent
Government 98 8.1
Private 256 21.3
Self-employed/informal sector 592 49.3
Unpaid family worker 2 .01
Not working 237 19.7
Never worked before 15 1.2
Total 1200 100.00

Spending patterns. This variable is constructed by adding three items that asks:
Which of the following did your family do in the past three months and
answerable by l=yes; 0=no.
a. Bought a major home appliance worth no less than Php 1,000.00.
b. Repaired/ Remodeled the house for worth no less than Php 1,000,00.
c. Invested or increased investment in any income-generating venture for not
less than Php 1,000.00.

Number of portable items in the household. This variable is constructed by adding
the items that the household has (cellular phones, black and white TV, colored
TV, personal computers, pager/beepers/microwave ovens)

Going out. This variable is constructed by adding the responses on the question:
Which of the following activities have you done in the past 6 months which is
answerable by yes or no.

Watched a movie

Watched a concert or recital
Watched a festival Show
Went to a park

P-P 9‘!”

Television Watching. This variable uses a single item that asks respondents “In an
ordinary day, how much time do respondents spend watching television?”

Frequency distribution of the television watching

Frequency Percent
1 Have no TV 251 20.9
2 Have TV but no time 63 5.3
3 Less than 1 hour 195 16.3
4 1-2 hours 366 30.5
5 3 or more hours 325 27.1
Total 1200 100.0

86

12. Perception of criminality. This variable is constructed by adding the scores of two
items that ask respondents about perceptions of criminality in the neighborhood:
a. In the past year, did criminality in your neighborhood ?
Answerable by: “rise” “did not change” “fall” “not know
b. In the coming year, will criminality in your neighbor ?
Answerable by: “will rise” “will not change” “will fall” “not know”

“none”.
“Not know” and “none” were considered in the neutral category and assigned with the

’91

same value as “will not change .

’9 6‘

none”.

Frequency distribution of the perception about criminality in the past year

Frequency Percent
1 Rose 178 14.8
2 Fell 427 35.6
3 Did not change 581 48.4
4 Not know 13 1.1
5 None 1 .1
Total 1200 100.0

Frequency distribution of the perception about criminality in the future

Frequency Percent
1 Will rise 165 13.8
2 Will fall 385 32.1
3 Will not change 599 49.9
4 Not know 50 4.2
5 None 1 .1
Total 1200 100.0

13. Overseas worker in the household. This item asks respondents whether there are
overseas contract workers in the family and answerable by: “currently”
“previously” and “never”. “Previously and “never” were collapsed into one

category.
Frequency distribution of overseas working
Frequency Percent
1 Currently 1 13 94
2 Previously 104 8.7
3 New 983 81.9
Total 1200 100.0

87

14. Homeownership. This variable includes two items that ask whether respondents

own their house and their lots with answers: “own” “renting” “neither own nor
rent” “owned by relatives” “owned by employer” “other”).
Frequency distribution of homeownership
Frequency Percent

1 Own house 899 74,9

2 Renting 163 13.6

3 Neither own nor rent 138 11.5

Total 1200 100.0

15. Locale. This item determines whether respondents live in a “rural” or “urban”
area. Interviewers were guided by the deﬁnition used by the National Statistical
Coordination Board, the Philippines’ policy making and coordinating body on
statistical matters. “Urban” areas fall under the following categories:

a. In their entirety, all municipal jurisdictions which, whether designated
chartered cities, provincial capital or not, have a population density of at
least 1,000 persons per square kilometer;

b. all barangays, poblaciones or central districts of municipalities and cities
which have a population density of at least 500 persons square kilometer;

c. Poblaciones or central districts not included in (1) and (2) regardless of the
population size which have the following:

1. street pattern or network of streets in either parallel or right angel
orientation;
ii. at least six establishments (commercial, manufacturing,
recreational and/or personal services);
iii. at least three of the following:
1. a town hall, church or chapel with religious service at least
once a month;
2. a public plaza, park or cemetery;
3. a market place, or building, where trading activities are
carried on at least once a week;
4. a public building, like a school, hospital, puericulture and
health center or library.

(1. Barangays having at least 1,000 inhabitants which meet the conditions set
forth in (3) above and where the occupation of the inhabitants is
predominantly non-farming or ﬁshing.

“Rural” areas are all poblaciones or central districts and all barrios that do not
meet the requirements for classiﬁcation of urban.

16. Regional Area (Metro Manila, Balance of Luzon, Visayas and Mindanao)

17. Population Mobility. This item asks respondents “where did you spend most of

your childhood” and answerable by: in “this town or city’ , in a different town”
“in a different province” and “outside the Philippines”.

88

 

Frequency distribution of where respondents spent most of their childhood

Frequency Percent
l Inthis town/city 760 63.3
2 In a different town 203 16.9
3 In a different province 235 19.6
4 Outside the Philippines 2 .2
Total 1200 100.0

18. Social cohesiveness. This variable makes use of two items measuring how closely
the respondents feel to the neighborhood (village and town) that they come from.
The answers are: “very close” “close” “not very close” “not close at all” and
“Can’t choose.” “Can’t choose” was assigned a middle value. 2

Frequency distribution of the closeness to neighborhood/village

Frequency Percent
1 Very close 333 27.8
2 Close 695 ‘ 57.9
3 Not very close 151 12.6
4 Not close at all 19 1.6
5 Cant choose 2 .2
Total 1200 100.0

Frequency distribution of the closeness to own town or city

Frequency Percent
1 Very close 179 14.9
2 Close 681 56.8
3 Not very close 278 23.2
4 Not close at all 47 3.9
5 Cant choose 15 1.3
Total 1200 100.0

19. Barangay Justice. This variable includes an item that measures the perceptions of
respondents on the efﬁcacy of the barangay in the delivery of justice services
answerable by “very satisﬁed” “satisﬁed” “maybe satisﬁed/maybe not satisﬁed”
“not satisﬁed” “not satisﬁed at all” and “not know”. “Not know” was assigned a

middle value.3
Frequency distribution of the satisfaction rating on barangay justice efficacy

Frequency Percent
1 Very satisﬁed 50 4.2
2 Satisﬁed 685 57.1
3 Maybe satisﬁed/maybe not 244 20.3
4 Not satisﬁed 191 15.9
5 Not at all satisﬁed 27 2.3
6 Not know 3 .3
Total 1200 100.0

89

20. Satisfaction to Law enforcement. This variable combines two items about
satisfaction about police and military performance in the locality. “Not Know”
and “Not Aware” are given middle values.4

Frequency distribution of the satisfaction rating with the military

1 Very satisﬁed

2 Satisﬁed

3 Maybe satisﬁed/maybe not
4 Not satisﬁed

5 Not at all satisﬁed

6 Not know

8 Not aware

Total

Frequency Percent
33 2.8
595 49.6
255 21.3
210 17.5
52 4.3
18 1.5
37 3.1
1200 100.0

Frequency distribution of the satisfaction rating with the police

1 Very satisﬁed

2 Satisﬁed

3 Maybe satisﬁed/maybe not
4 Not satisﬁed

5 Not at all satisﬁed

6 Not know

8 Not aware

Total

 

NOTES TO APPENDIX 1

Frequency Percent

36 3.0
590 49.2 _ .-_
256 21.3
230 19.2

64 5.3

16 1.3

8 .7
1200 100.0

' I also assigned missing values to “Not know” and “none.” The bivariate and multivariate results were not

substantially altered.

2 “Can’t choose” was also assigned a missing value. The results are also not substantially altered.

3 “Not know” was also assigned a missing value. The results are also basically the same.

4 “Not know” and “not aware” were also designated as missing. The results are not substantially altered.

90

EoEooS-«em 33 ">33
353. xmwgm Ham
mmoaozmoaoonmoo
€368 cesaaoa "mo:
83 Reomwouuwmm
282 "004
hogooﬁonuzom
coo—83 05325 36326 "3&0

cooﬁonammo: 2: E 25.8 mo cosmoocoa "EU
maﬁa; >912.

So wﬁow 2:: no 528:: "PDQ

22682 65 a 86: anacofmg
238:0: 06 mo Quota wamocoamumm

one: 29.095: 05 me 5:82 #83 M;
536388 20:88: "0mm

Becca-so: 05 mo 830:5 guacamnmm
20:83: 2: mo oozes—be oo Bare—um“?

one: 228:0: mo cosmosvouDQm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

wmo. wvo. No_.- 2:. Nvm. owo. onor No_.- ~N_.- mo_.- _mm.- who: go. moor w_o.- VS. o2. mono. to; >33
oNN. mmo. wmo. _N_. wmo. for Nmo. mNor ovor owor ohor mNo. omor omor 08. N2. mno. mmo. mm
mmor vmo. mi. 3o. mNor omor 2o: moor moor owor Nmor omor Nvor mmo. who. vmo. _oo.- $00
mvor mtr ooor So. So. mmo. wmo. owo. _No. Eb.- mmor moor mmor ~mo.- Go. nmor mOE
own. omﬁ #5.. N3: mom: mNo... wmvr ohor moo. moor mNo. emm. of. voo. ooor 0mm
3N. gov wwor ommr vomr Nov: ooor wmﬁ wit moor hmm. com. oo—. Nvor DOA
Nvo. ooor mnor 02.- moor 2o. 3N. Nmo. omor woo. :_. omor mmm. EOE
moo. moo. No. of. mg. hmo. :Vo. to. var ovor 0N7 mwo. 3&0
Noo. mvo. o2. owor w_o.- ﬁmo. oNor :7 mnor oNo. ovo. EU
2 _. Sm. who. Nmor ﬁoﬁ N8. mmNr _o_ .- who: :o. >H
we. on“. ooor ono. coo. om _ .- of .- vmor who: PDQ
mo o. ovor m3. vmo. amt w :4.- owor 3 ﬂ. MOm
hmor woor moo. mgr ooNr 2o. Swot mom
mwo. w_o.- mvor woﬁ om_.- 5mm. A3
mmo. 52.- o~o.- moor mmm. Um:
mmo. oo~.- 3 ~.- 37 mm
com. So. _oN.- HE<
_oo.- ooo. UDQm
ovmr Xmm
mm 300 mOE Om; 004 203 3&0 EU >H HBO MOA mmm .55 033 mm a: Dam Xmm m0<

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mmqm§<> PZmDmemQ/m m0 mZOﬁ-«AEOU ”N wag-63%

91

REFERENCES

Agoncillo, T. C. (1960) History of the Filipino People. Quezon City: Garotech
Publishing.

Alcantara, A. N. (1994). Gender Roles, Fertility, and the Status of Married Filipino Men
and Women. Philippine Sociological Review 42, 94-109.

Andresen, M. A. (2006). A Spatial Analysis of Crime in Vancouver, British Columbia:
A Synthesis of Social Disorganization Theory and Routine Activity Theory.
Canadian Geographer, 50, 487-502.

Asis, M.M., Huang, 8., & Yeoh, B. (2004). When the Light of the Home is Abroad:
Unskilled Female Migration and the Filipino Family. Singapore Journal of
Tropical Geography, 25, 198-215.

Bachman, R. & Paternoster, R. (2004). Statistics for Criminology and Criminal Justice.
New York: McGraw Hill Companies.

Basserer, S. (1996). Criminal Victimization: An International Perspective. Juristat.
Canadian Center for Justice Studies, 18. 6

Bennett, RR. (1991). Routine Activities: A Cross National Assessment of a
Criminological Perspective. Social Forces, 70, 147-163.

Bryant, J. (1995). Children of International Migrants in Indonesia, Thailand and the
Philippines. A Review of Evidence and Policies. UNICEF: Innocenti Research
Centre. http://www.unicef-icdc.org/publications/pdf/iwp2005_05.pdf

Bursik, R.J. (1988). Social Disorganization and Theories of Crime and Delinquency:
Problems and Prospects. Criminology. 26, 519-551.

Bursik, R. J. & Grasmick, H.G. (1993). Neighborhoods and Crime. New York:
Lexington Books.

Clarke, R.V. (1983). Situational Crime Prevention: Its Theoretical Basis and Practical
Scope. Crime and Justice. 225-256.

Cohen, LE. (1981). Modeling Crime Trends: A Criminal Opportunity Perspective.
Journal of Research in Crime and Delinquency. 138-164.

Cohen, L.E., & Cantor, D. (1980). The Determinants of Larceny: An Empirical and
Theoretical Study. Journal of Research in Crime and Delinquency, 140-159.

Cohen, L.E., & Cantor, D. (1981). Residential Burglary in the United States: Lifestyle

92

 

and Demographic Factors Associated with the Probability of Victimization.
Journal of Research in Crime and Delinquency, 1 13-127.

Cohen, LE, and Felson, M. (1979). Social Change and Crime Rate Trends: A Routine
Activity Approach. American Sociological Review. 44. 588-608.

Cohen, L.E., Felson, M., Land, KC. (1980). Property Crime Rates in the United States:
A Macrodynamic Analysis, 1947-1977; With Ex Ante Forecasts for the Mid-
l980’s. American Journal of Sociology, 86, 90-1 18.

Cohen, L.E., Kluegel, J .R. & Land, KC. (1981). Social Inequality and Predatory
Criminal Victimization: An Exposition and Test of a Formal Theory. American
Sociological Review, 46, 505-524.

Coleman, J .S. (1988). Social Capital in the Creation of Human Capital. American
Journal of Sociology, 94, 95-120.

 

Cook, RI. (1986). The Demand and Supply of Criminal Opportunities. Crime and
Justice, 7, 1-27.

Constantino, R. (1975). The Philippines: A Past Revisited. Vol. I. Quezon City:
Foundation for Nationalist Studies.

De Quiros, C. (1997). Dead Aim: How Marcos Ambushed Philippine Democracy.
Manila: Foundation for Worldwide People Power.

Domingo, LI. and Asis, M.M. (1995). Living Arrangements and the Flow of Support

Between Generations in the Philippines. Journal of Cross-cultural Geronthology,
10, 21-51.

Eck, J. (1995). Examining Routine Activity Theory: A Review of Two Books. Justice
Quarterly, 12, 783- 797.

Enriquez, V. (1975). Mga Batayan ng Sikolohiyang Pilipino sa Kultura at Kasaysayan.
(The Bases of Philippine Psychology in Culture and History). General Education
Journal. 21, 69—88.

Felson, M. (1986). Linking Criminal Choices, Routine Activities, Informal Control,
and Criminal Outcomes. In The Reasoning Criminal: Rational Choice
Perspectives on Oﬂending. Cornish, D.B & Clarke, R.V. (eds.) New York:
Springer.

Garofalo, J., Siegel, L., & Laub, J. (1987). School-Related Victimizations Among

Adolescents: An Analysis of National Crime Survey (N CS) Narratives. Journal
of Quantitative Criminology, 3, 321-339.

93

Gerson, P. (1998). "Poverty, Income Distribution, and Economic Policy in the
Philippines”. IMF Working Paper No. 98/20 Available at SSRN: Accessed
February 11, 2007. llllj)I-".'SSI’11.C(llllf't’llﬁil'ittﬁi882248

Glaser, D. (1971). Social Deviance. Chicago: Markham.

Gonzalez, J .L. III (1998). Philippine Labour Migration: Critical Dimension of Public
Policy. Institute of Southeast Asian Studies.

Gottfredson, MR. (1981). On the Etiology of Criminal Victimization. Journal of
Criminal Law and Criminology, 72, 714-726.

Hough, M. (1987). Offenders’ Choice of Target: Findings from Victim Surveys. Journal
of Quantitative Criminology, 3, 355-368.

Hunter, AJ. (1985). Private, Parochial and Public School Orders: The Problem of
Crime and Incivility in Urban Communities. In Suttles, G.D. & Zald M.N. (eds).
The Challenge of Social Control: Citizenship and Instituion Building in Modern
Society. Northwood, N.J.: Ablex Publishing.

Hutchcroft, P. (1998). Booty Capitalism: The Politics of Banking in the Philippines. New
York: Cornell University Press.

Jocano, F. L. (1998). Filipino Prehistory: Anthropology of the Filipino People I. Quezon
City: F. Landa J ocano and Punlad Research House Inc.

Kennedy, L.W., & F orde, DR. (1990). Routine Activities and Crime: An Analysis of
Victimization in Canada. Criminology, 28, 137-152.

Komhauser, R. (1978). Social Sources of Delinquency. Chicago: University of Chicago
Press

Kubrin, C.E. & Weitzer, R. (2003). New Directions in Social Disorganization Theory.
Journal of Research in Crime and Delinquency, 4, 374-402.

Land, K.C., McCall, P.L., & Cohen, LE. (1990). Structural Covariates of Homicide
Rates: Are there any Invariances Across Time and Social Space? American
Journal of Sociology, 95, 922-963.

Lee, M. R. (2000). Community Cohesion and Violent Predatory Victimization: A
Theoretical Extension and Cross-National Test of Opportunity Theory. Social
Forces, 79 , 683-706.

Lipset, S. (1960). The Political Man: Social Bases of Politics. Garden City, N.Y.,
Doubleday,

94

Lynch, JP. (1987). Routine Activity and Victimization at Work. Journal of Quantitative
Criminology. 3, 283-291.

Massey, J .L., Krohn, M.D., & Bonati, L.M. (1989). Property Crime and the Routine
Activities of Individuals. Journal of Research in Crime and Delinquency, 26,
378-400.

Maxﬁeld, MG. (1987). Lifestyle and Routine Activity Theories of Crime: Empirical
Studies of Victimization, Delinquency, and Offender Decision-Making. Journal
of Quantitative Criminology, 3, 275-282.

Maxﬁeld, MG. (1987). Household Composition, Routine Activity, and Victimization: A
Comparative Analysis. Journal of Quantitative Criminology, 3, 301-320.

Messner, S.F., & Tardiff, K. (1985). The Social Ecology of Urban Homicide: An
Application of the Routine Activities Approach. Criminology, 23, 241-267.

Messner, S.F., & Blau, J .R. (1987). Routine Leisure Activities and Rates of Crime: A
Macro-Level Analysis. Social Forces, 65, 1035-1052.

Meier, R.F., & Miethe, T.D., (1993). Understanding Theories of Criminal Victimization.
Crime and Justice. 1993, 459-499.

Miethe, TD. (1995). Fear and Withdrawal from Urban Life. Annals of American
Academy. 539, 14-27.

Miethe, T.D., Stafford, M.C., & Long, J .S. (1987). Social Differentiation in Criminal
Victimization: A Test of Routine Activities/Lifestyle Theories. American
Sociological Review, 52, 184-194.

Miethe, T.D., Hughes, M., & McDowall, D. (1991). Social Change and Crime Rates: An
Evaluation of Alternative Theoretical Approaches. Social Forces, 70, 165-185.

Miethe, T.D., & McDowall, D. (1993). Contextual Effects in Models of Criminal
Victimizations. Social Forces, 71, 741 -759.

Morenoff, J .D., Sampson, R.J., & Raudenbush, SW. (2001). Neighborhood Inequality,
Collective Efﬁcacy, and the Spatial Dynamics of Urban Violence. Criminology,
39, 517-559.

Murakami, A., Zain, M., Takeuchi, K., Tsunekawa, A., & Yokota, S. (2005). Trends in
Urbanization and Patterns of Land Use in the Asian Mega-Cities Jakarta,
Bangkok and Metro Manila. Landscape and Urban Planning 70: 251-259.

Mustaine, E.E., & Tewksbury, R. (1998). Predicting Risks of Larceny Theft
Victimization: A Routine Activity Analysis Using Reﬁned Lifestyle Measures.

95

Criminology, 36, 829-857.
National Council on the Role of the Filipino Women ( 2005).

National Statistical Coordination Board (2000). Accessed August 20, 2006.

h Itp : .s’r’xvuiu' . use I). gin-41h secSt:’1t.—‘tl_pupn .asp

 

Pe-pua, R. & Protacio-Marcelino, E. (2000). Sikolohiyang Pilipino (Filipino
Psychology): A Legacy of Virgilio Enriquez. Asian Journal of Social Psychology,
3, 49-71.

Patillo, ME. (1998). Sweet Mothers and Gang-Bangers: Managing Crime in a Black
Middle-Class Neighborhood. Social Forces, 76, 747-774.

Rivera, T. (1994). Landlords and Capitalists: Class, Family and State in Philippine
Manufacturing. Quezon City: University of the Philippines Press.

Roundtree, P., Land, K.C., & Miethe, TD. (1994). Macro-micro Integration in the
Study of Victimization: A Hierarchical Logistic Model Analysis across Seattle
Neighborhoods. Criminology, 32, 387-414.

Roncek, D.W. & Maier, RA. (1991). Bars, Locks. And Crimes Revisited: Linking the
Theory of Routine Activities to the Empiricism of “Hot Spots”. Criminology. 29,
725-753.

Sampson, R.J. (1985). Neighborhood and Crime: The Structural Determinants of
Personal Victimization. Journal of Research in Crime and Delinquency, 22, 7-40.

Sampson, R.J. (1987). Personal Violence by Strangers: An Extension and Test of the
Opportunity Model of Predatory Victimization. Journal of Criminal Law and
Criminology, 1987, 327-356.

Sampson, R.J., & Groves, W.B. (1989). Community Structure and Crime: Testing Social
Disorganization Theory. American Journal of Sociology, 94, 774-802.

Sampson, R.J., & Lauritsen, J .L. (1990). Deviant Lifestyles, Proximity to Crime, and the
Offender-Victim Link in Personal Violence. Journal of Research in Crime and
Delinquency, 27, 110-139.

Sampson, R.J., Raudenbush, S.W., & Earls, F. (1997). Neighborhoods and Violent
Crime. Science. 277, 918-924.

Sampson, R.J., & Wooldredge, JD. (1987). Linking the Micro- and Macro-Level

Dimensions of Lifestyle-Routine Activity and Opportunity Models of Predatory
Victimization. Journal of Quantitative Criminology. 3. 371-393.

96

Shaw, C. & Mckay, J. (1942). Juvenile Delinquency and Urban Areas. Chicago:
University of Chicago Press.

Sherman, L.W., Gartin, P.R., & Buerger, M. E. (1989). Hot Spots of Predatory Crime:
Routine Activities and the Criminology of Place. Criminology, 27, 27—55.

Sidel, J. T. (1999). Capital, Coercion, and Crime Bossism in the Philippines. Palo Alto,
California: Stanford University Press.

Silliman, GS. (1985). A Political Analysis of the Philippines Katarungang Pambarangay
System of Informal Justice Through Mediation. Law and Society Review, 19, 297-
302.

Smith, D.A., & Jarjoura, GR. (1989). Household Characteristics, Neighborhood
Composition and Victimization Risk. Social Forces, 68, 621-640.

Smith, W.R., Frazee, S.G., & Davison, E. (2000). Furthering the Integration of Routine
Activity Theory and Social Disorganization Theories: Small Units of Analysis
and the Study of Street Robbery as a Diffusion Process. Criminology, 38, 489-
521.

Social Weather Stations. SOCIAL WEATHER STATIONS SURVEY [PHILIPPINES]:
QUARTER IV, 1995 [Computer File]. ICPSR Version. Quezon City, Philippines:
Social Weather Stations [producer], 1999. Ann Arbor, Mi: Inter-University
Consortium for Political and Social Research [distributor], 1999.

Stahura, J .M., & Sloan, J.J. (1988). Urban Stratiﬁcation of Places, Routine Activities
and Suburban Crime Rates. Social Forces, 1102-1118.

Stark, R. (1987) Deviant Places: A Theory of the Ecology of Crime. Criminology, 25,
893-909.

Velez, M. B. (2001). The Role of Public Social Control In Urban Neighborhoods: A
Multi-level Analysis of Victimization Risks. Criminology, 39, 837-864.

Veysey, B.M. & Messner, S. F. (1999). Further Testing of Social Disorganization
Theory: An Elaboration of Sampson and Groves’ “Community Structure and
Crime.” Journal of Research in Crime and Delinquency, 36, 156-174.

Villarreal, A., & Silva, B.F.A. (2006). Social Cohesion, Criminal Victimization and
Perceived Risk of Crime in Brazilian Neighborhoods. Social Forces, 84, 1725-
1753.

Vitug, M & Gloria, G. (2000). Under the Crescent Moon: Rebellion in Mindanao.
Quezon City: Ateneo Center for Social Policy and Public Affairs.

97

Williams, L., & Domingo, L]. (1993). The Social Status of Elderly Women and Men in
the Filipino Family. Journal of Marriage and the Family, 55, 415-426.

Wooldredge, J .D., Cullen, F.T., & Latessa, El. (1992). Research Note Victimization in
the Workplace: A Test of the Routine Activities Theory. Justice Quarterly, 9,
325-335.

Zhang, L., Welte, J .W. & Wieczorek, WP. (2001). Deviant Lifestyle and Crime
Victimization. Journal of Criminal Justice, 29, 133-143.

98

IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

llIlllllllllIllllllllllllllllIllllllllllllllllllllll