Download Lifestyles and Self Protective Behaviours and more Papers Business Administration in PDF only on Docsity!
10.1177/0093854803252354CRIMINAL JUSTICE AND BEHAVIORTewksbury, Mustaine / LIFESTYLES AND GUARDIANSHIP ARTICLE
COLLEGE STUDENTS’
LIFESTYLES AND SELF-
PROTECTIVE BEHAVIORS
Further Considerations of
the Guardianship Concept
in Routine Activity Theory
RICHARD TEWKSBURY
University of Louisville
ELIZABETH EHRHARDT MUSTAINE
University of Central Florida
Routine activity theory has traditionally emphasized identifying victimization risks and suitable targets for crime. Assessments of the role of guardianship in criminal events are less emphasized. Explorations of who uses guardianship to attempt to reduce their chances for victimization have been developed only minimally, typically relying on demographics. This research goes further in assessing who uses self-protective strategies, considering lifestyles related to proximity to moti- vated offenders, the suitability of individuals as targets, and how these characteristics influence the use of self-protective devices. Results show the most influential lifestyle characteristics and behaviors on use of self-protective measures are exposure to potential offenders and neighbor- hood characteristics. Fear of crime, substance use, and individual demographics show only small relationships to guardianship.
Keywords: weapons; routine activities theory; guardianship
C
rime, as explained by routine activity theory, is a result of the con-
vergence in space and time of motivated or potential offenders,
suitable targets, and an absence of capable guardianship (see Cohen &
Felson, 1979; Cohen, Felson, & Land, 1980; Cohen, Kluegel, & Land,
CRIMINAL JUSTICE AND BEHAVIOR, Vol. 30 No. 3, June 2003 302- DOI: 10.1177/ © 2003 American Association for Correctional Psychology
1981). Although a large amount of information exists about potential
offenders (and “hot spots” where their numbers are high) and both
actual and potential victims or targets, the issue of guardianship lags in
theoretical development and empirical testing. Some research has
addressed what forms of self-protective behaviors are commonly
used, and other have attempted to assess the efficacy of guardianship
activities. However, little is known about the type of person who
employs self-protective measures, the type of lifestyles of those who
do and do not tend to use such measures, and how or if the use of self-
protection is related to individuals’proximity to potential offenders or
individuals’ suitability as a target.
ROUTINE ACTIVITY THEORY
Research inspired by routine activity theory has consistently shown
that criminal victimizations are not randomly distributed in society
but rather are associated with the lifestyles and daily routines of indi-
viduals as well as with demographics. Routine activities determine the
amount of exposure people have to potential offenders, how valuable
or vulnerable they or their property is as a target, and whether or how
well guarded they and their property are. When potential offenders,
suitable targets, and a lack of capable guardians converge, the likeli-
hood of a criminal event increases.
Routine activity theory specifically suggests that the social context
of crime is a central issue in understanding the incidence of criminal
events (Collins, Cox, & Langan, 1987; Lynch, 1987; Mustaine &
Tewksbury, 1998a, 1998b; Wooldredge, Cullen, & Latessa, 1992).
Certainly, a person may be willing to commit crime given sufficient
opportunity, but if that opportunity never comes, the crime will not
occur. Also possible is the circumstance of finding a suitable target but
one that is too well guarded to merit an attempt.
Although we know that guardianship exists at both the formal (i.e.,
official and institutional) and informal (personal behavior) levels, rel-
atively little is known about the use of most forms of individual-level
protective behaviors. Research examining offenders’ choices of tar-
gets, however, strongly suggests that the most effective guardianship
activities are accomplished on the individual level, not on official or
institutional levels (Buck, Hakim, & Rengert, 1993; Cromwell,
Tewksbury, Mustaine / LIFESTYLES AND GUARDIANSHIP 303
ianship measures’ influence on victimization risks. They argued,
“These items adequately represent variation in active crime-prevention
behavior undertaken to reduce victimization risks” (Miethe et al.,
1990, p. 364). However, we believe that changes in activities such as
these are not necessarily issues of guardianship as much as they are
simply changes in routine activities that may limit exposure to poten-
tial offenders. With the exception of increases in household members,
these measures do not play a role in actual protection of individual per-
son or property.
Likewise, Miethe and Meier (1990) used two measures of guard-
ianship in assessing risk of burglary, personal theft, and assault: living
with a household member older than the age of 16 (as a measure of
social guardianship) and whether someone either carries a weapon or
has a burglar alarm (as an indicator of physical guardianship). They
found that whereas their measure of social guardianship was associ-
ated with reduced victimization risks, physical guardianship was not
(see also Copes, 1999; Miethe & McDowall, 1993). However, their
measurements of guardianship, although perhaps the best available in
the literature to date, do not assess the wide variety of available guard-
ianship measures, how these relate to specific offenses, or what types
of individuals are likely to have such measures present.
GUARDIANSHIP ACTIVITIES
The specific guardianship activities individuals employ vary across
individuals’social status (as status increases, so too does likelihood of
acquiring a form of self-protection), age, gender, ethnicity, marital
status, education, and objective assessments of victimization risks
(Baumer, 1980; Rountree & Land, 1996; Ziegenhagen & Brosnan,
Possession of a gun, perhaps the clearest indication of an attempt to
protect oneself from victimization, has been shown by a number of
researchers to be associated with demographics, victimization experi-
ences, and perceived risk of victimization (Cao, Cullen, & Link, 1997;
Cook & Ludwig, 1996; Kleck & DeLeone, 1993; Lizotte & Bordua,
1980; Luxenburg, Cullen, Langworthy, & Kopache, 1994; Simon,
Dent, & Sussman, 1997; Smith & Uchida, 1988). In contrast, others
Tewksbury, Mustaine / LIFESTYLES AND GUARDIANSHIP 305
have found no or only minor significant relationships between victim-
ization or fear of victimization and possession of guns (DuRant, Getts,
Cadenhead, & Woods, 1995; Stinchcombe et al., 1980; Young, 1985).
Age is inversely related to possession of self-protective devices,
including guns (Baumer, 1980; Brosnan, 1992; DuRant et al., 1995;
Luxenburg et al., 1994; Ziegenhagen & Brosnan, 1990). For adoles-
cents and young adults, use of protective measures is common and
believed to often be an outgrowth of fear and safety concerns
(Kingery, Coggeshall, & Alford, 1999; Lizotte, Tesoriero,
Thornberry, & Krohn, 1994). Williams, Singh, and Singh (1994) have
reported that only 11% of urban high school students do not take any
defensive actions in response to concerns about their possible criminal
victimization. Among high school students and their families, posses-
sion of mace, dogs, extra lighting (see also Painter, 1996), and restric-
tions on out-of-home nighttime activities are fairly common means of
self-protection.
Other demographic variables have only been infrequently related
to weapon possession and/or carrying. In regard to ethnicity, African
Americans and Hispanics are more likely than Whites to carry self-
protective devices (U.S. Public Health Service, 1991; Ziegenhagen &
Brosnan, 1990; also see Kingery et al., 1999). Those with higher
incomes may be more likely to own guns (Luxenburg et al., 1994).
Also, married persons are more likely to own guns (Moriarty & Davis,
1992). However, other characteristics examined in the literature,
including family structure and religious behavior, have not been
shown to be associated with possessing and/or carrying a weapon
(DuRant et al., 1995).
Other factors that have been related to the use of self-protection are
fears about crime and actual victimization experiences. Fears about
possible victimization are common (Ferraro, 1995) for individuals in
both high- and low-crime neighborhoods (Greenberg, Rohe, & Wil-
liams, 1982). However, it is in highly disordered communities where
residents are most likely to have higher levels of fear of crime
(McGarrell, Giacomazzi, & Thurman, 1997). Therefore, it is apparent
that social structural variables or neighborhood characteristics may
be important, although perhaps indirect, predictors of the use of self-
protective behaviors.
306 CRIMINAL JUSTICE AND BEHAVIOR
associated with increased risks of criminal victimization. It is not only
students’social activities, however, that have been shown to be associ-
ated with increased victimization risks. Students who hold jobs in
addition to attending school may be at increased risk for some forms
of victimization (Mustaine & Tewksbury, 1998c).
So although researchers have examined how specific forms of life-
style and behaviors are related to students’ risks for victimization,
what has not been examined is how students’ lifestyles influence
whether they use self-protective devices for protection against victim-
ization. This is not unique to the study of campus crime, however.
Information about who uses self-protection and how or if this concept
is related to the other routine activity theory elements (potential
offenders and suitable targets) remains largely undeveloped. For
example, if a person frequently goes out at night for leisure (thereby
increasing her exposure to potential offenders), is she more likely to
carry self-protection? Is she more likely to carry self-protection than
someone who does not go out frequently? Is the place she goes influ-
ential in her decision to carry self-protection? Is someone who drinks
a lot or uses illegal drugs (thereby being an easier target) more likely to
carry self-protection than someone who does not? Or, is a person who
lives in a neighborhood with particular high-crime attributes more
likely to carry self-protection?
THIS STUDY
This study sought to fulfill two important theoretical purposes.
Because an important element in understanding criminal events is the
presence (or absence) of capable guardians and the literature on
guardianship has neglected the consideration of who uses self-
protection, we offered an assessment of lifestyle characteristics of
persons employing guardianship measures. Second, whereas the mea-
sures of individual characteristics (such as lifestyle) used in this study
were more specific than those previously used in the literature, this
research offered important advances in the specificity of routine activ-
ity theory and the comprehension of guardianship as an essential ele-
ment in the awareness of criminal events.
308 CRIMINAL JUSTICE AND BEHAVIOR
METHOD
PARTICIPANTS
The data in this study came from self-administered surveys col-
lected during the first 3 weeks of the fall 1996 academic term. A total
of 1,513 college and university students in nine postsecondary institu-
tions (including large and small universities, 4-year colleges, and
community colleges) in eight states
1
completed the survey. Institu-
tions and courses for data collection were selected for inclusion based
on geographic dispersion and the inclusion of a wide variety of
postsecondary educational institutions. The sample was drawn from
students in introductory-level sociology and criminal justice courses;
this approach was employed so as to maximize the diversity of stu-
dents’ majors and life experiences. All respondents were volunteers
and received no compensation for their participation.
As shown in Table 1, the sample appeared to be representative of
contemporary college students. The sample was overwhelmingly full-
time, heterosexual students and more than one half of the sample was
female, White, younger than the age of 21, never married, employed,
and living off campus.
PROCEDURE
When a department agreed to participate, introductory courses
were identified for solicitation and faculty members were provided
either a written script or an audiotaped set of solicitations for students,
instructions for instrument completion, and two copies of an informed
consent form for all students. If students elected to participate, they
read and signed both consent forms, retaining one countersigned
form.
The 95-item instrument assessed a wide range of individual demo-
graphics, experiences, daily routines, residential and social commu-
nity structural and contextual variables, fear of crime, perceptions of
safety, and self-protective behaviors. Also included was self-reported
substance use (including rates, varieties, and settings in which use
occurred).
Tewksbury, Mustaine / LIFESTYLES AND GUARDIANSHIP 309
less than half of the sample (42.6%) reported regularly using one or
more of these types of self-protection; of these, the majority used only
one form (see Table 1). The most common forms of self-protection
were the use of mace (21.5%) and guns (17%).
3
The independent vari-
ables for the analysis are discussed in the Results section and are high-
lighted in Table 2.
RESULTS AND DISCUSSION
To explore the element of guardianship in detail, we assessed the
personal and lifestyle characteristics of those college students who
used forms of self-protection. To accomplish this, the logistic regres-
sion analysis proceeded in two stages. First, we tested several theoreti-
cally driven groupings of variables to come up with the model that pro-
vided the best fit to the data for the use of self-protective behaviors. In
each case, those variables were grouped together, which measured
related routine activity concepts or similar lifestyle activities, includ-
ing respondents’ demographic characteristics, characteristics of their
neighborhoods or communities, structures in their neighborhoods or
communities, lifestyle behaviors influencing exposure to potential
offenders, measures of fear of crime; measures of respondents’ per-
ceived safety, and alcohol and drug use. These lifestyle groupings
were derived from important routine activity theory propositions
about proximity to potential offenders, perceptions about proximity to
potential offenders, and target suitability. Each theoretical grouping of
variables was regressed separately against the dependent variable of
the use of self-protective behaviors (the guardianship element in rou-
tine activity theory), and we kept the indicators with significant
explanatory powers ( p = .10). Although we are aware that measures of
significance are inapplicable for a nonrandom sample, they are
reported and used here as aids in the interpretation of the results and
for subsequent theoretical development. The initial results of these
groupings are displayed in Table 2 and are provided for informational
purposes only. They will be noted in general but will not be discussed
in detail.
Tewksbury, Mustaine / LIFESTYLES AND GUARDIANSHIP 311
312 CRIMINAL JUSTICE AND BEHAVIOR
TABLE 2: Theoretical Groupings and Initial Logistic Regression Results
Variable B Exponential (b) SETolerance
Demographic characteristics Age (older) .02 1.02 .01. Sex (male) –.09 .92 .11. Race (non–White) –.08 .92 .13. Married .30 1.35 .23. Social class (higher) .01 1.01 .06. Sexual orientation (gay, lesbian, or bisexual) –.07 .94 .38. Lives in a rural area –.35* .71 .18. Lives on campus –.13 .88 .14. Employed –.51* 1.66 .11. χ^2 = 39.485,p =. Neighborhood structures Lives near a park .17 1.18 .11. Lives near a fast food restaurant –.38* .69 .13. Lives near a liquor store –.01 1.00 .13. Lives near a night club –.19 .83 .15. Lives near a convenience store .05 1.05 .14. Lives near a place with homeless persons .14 1.15 .26. Lives near a bar –.05 .95 .12. Lives near a place with gang graffiti .07 1.08 .18. Lives near a police station .47* 1.60 .13. χ^2 = 26.719,p =. Neighborhood conditions Rarely encounters young people hanging out in neighborhood .02 1.02 .05. Lives in a neighborhood with litter .06 1.07 .11. Lives in a neighborhood with unattended dogs .28* 1.32 .11. Lives in a neighborhood with unsupervised youth .25* 1.29 .10. Lives near noisy and disruptive neighbors –.04 .96 .13. Lives in a neighborhood with vacant buildings –.10 .91 .17. Lives in a neighborhood with too much noise –.01 .99 .14. Lives in a neighborhood with people high on drugs or drunk in public –.10 .90 .13. Lives in a neighborhood with too much crime –.14 .87 .15. Lives in a neighborhood with abandoned cars or car parts lying around .30 1.35 .20. χ^2 = 23.293,p =.
(continued)
Next, these significant variables were considered together and any
variables that were not significant (assessed at p = .10) were removed
from the model. By comparing chi squares and degrees of freedom
across models, the model providing the most parsimonious fit to the
data was determined.
4
This best model is presented in Table 3.
In all models, we used logistic regression for the analysis of the
data. To check for multicollinearity, we used two common methods
for nonratio and/or interval data (Menard, 1995). First, as a rough
guideline, unstandardized logistic regression coefficients greater than
2 should be examined. Second, by running the logistic models using
linear regression, the variables’ tolerances can be calculated. Any tol-
erances of more than .60 indicate there is no serious problem with
multicollinearity, whereas variables with a tolerance of less than.
are cause for concern and a tolerance of less than .10 indicates a seri-
ous problem (Menard, 1995). With all models, multicollinearity was
not a problem (see Bs and tolerances reported in Tables 2 and 3).
314 CRIMINAL JUSTICE AND BEHAVIOR
Perceives residence to be safe .33* 1.40 .11. Feels vulnerable to robbers –.04* .96 .03. Neighbors would help in event of emergency .023 1.25 .16. Knows neighbors fairly or very well –.19* .83 .07. χ^2 = 36.210,p =. Drug or alcohol use Gets drunk frequently –.08 .92 .07. Drinks alcohol regularly .08 1.08 .06. Has been drunk in public recently –.02 .98 .14. Uses marijuana .02 1.02 .06. Uses cocaine –.51 .60 .33. Uses crack 1.02* 2.76 .49. Uses acid .15 1.16 .26. Uses other drugs .27 1.32 .26. Has been present when others were using or preparing to use drugs –.20 .82 .13. χ^2 = 12.564,p =.
*p = .10; variable carried forward.
TABLE 2 (continued)
Variable B Exponential (b) SETolerance
LIFESTYLE BEHAVIORS AND CHARACTERISTICS
ASSOCIATED WITH SELF-PROTECTIVE BEHAVIORS
An examination of Table 2 illustrates the myriad lifestyle charac-
teristics that are potentially associated with the use of self-protective
or guardianship behaviors. It is interesting to note what variables are
not significantly associated with the use of self-protection. With the
exception of the theoretical category measuring lifestyle and exposure
to criminal events, most theoretical categories have only a small num-
ber and proportion of variables that are significant. Thus, most catego-
ries do not provide robust explanations for individuals’ probability of
using self-protective behaviors. This is a significant and valuable dis-
covery because it suggests that rationales purported by routine activity
theory are important not only for understanding and delineating vic-
timization risks but also for determining choices about the use of
guardianship behaviors. Routine activity theory suggests that life-
styles that involve leaving the home more frequently and exposure to
strangers or persons in their crime-prone years are more dangerous
activities than staying at home and confining associations to more
familiar persons and those who do not share the characteristics of
criminals. This initial assessment of Table 2 suggests that for these
Tewksbury, Mustaine / LIFESTYLES AND GUARDIANSHIP 315
TABLE 3: Assessing Lifestyle Behaviors and Characteristics Associated With the Use of Self-Protection
Variable B SE Exponential (b) Tolerance
Employed –.52 .12 .60. Lives near a police station .51 .13 1.66. Lives near a fast food restaurant –.32 .12 .73. Lives in neighborhood with unsupervised youth .21 .09 1.23. Uses crack 1.21 .49 3.36. Perceives residence to be safe .30 .09 1.34. Percentage of time spent with strangers during week –.01 .00 .99. Percentage of time spent with strangers during weekend .01 .00 1.01. Is frequently on foot when one goes out for entertainment in the evening .18 .06 1.19. Constant –.71.
NOTE: All variables are significant atp ≤ .05; χ^2 = 88.995,p = .001; –2log likelihood = 1881.08;df = 9;N = 1,442.
safety and fear were the most influential over individuals’ use of per-
sonal protection.
Moving on to Table 3, we can see the variables that were signifi-
cantly related to the use of self-protection or guardianship. This table
presents the most parsimonious results from the logistic regression for
the use of self-protection. An examination of the variables in the
model reveals that nine measures of lifestyle stemming from six out of
the seven theoretical groupings are significantly associated with col-
lege students’ use of protective behaviors.
First, we find that one demographic characteristic, employment
status, contributes significantly to the probability of a college student
using protective behaviors. Students who are employed are 40% less
likely to use self-protection than those who are unemployed. It may be
that students who are unemployed are more suitable targets because
they have characteristics that are more similar to potential offenders.
In addition, it is conceivable that students who are unemployed are
more exposed (or perceive they are more exposed) to potential offend-
ers because they are not inside working or getting to and from work
but are outside and in locations where potential offenders may be.
These perceptions or circumstances may increase their likelihood of
behaving self-protectively. Again, we note that none of the other eight
demographic indicators are significant determinants of guardianship
behavior. This is in line with routine activity theory and its critics who
suggest that when proxies of lifestyle, such as demographics, are con-
sidered together with actual lifestyle behaviors and characteristics, the
proxies lose their significance in the face of better measures.
Turning to the theoretical categories of neighborhood structures
and characteristics, we find that three measures of social disorgani-
zation of communities are important in explaining the use of self-
protection among these students: living near police stations, near fast
food restaurants, and in neighborhoods with unsupervised youth. To
elaborate, students who live near police stations have 66% greater
odds of using self-protection than do persons who do not live near
police stations. These findings are contrary to some previous research
(Wycoff, Skogan, Pate, & Sherman, 1985). However, persons who
live near police stations have a relatively constant reminder of the
presence of crime and may therefore be more likely to take precau-
tions; considerations of crime may be more prevalent in communities
Tewksbury, Mustaine / LIFESTYLES AND GUARDIANSHIP 317
with physical and social signs of crime. In addition, police stations are
likely to be located in disordered or lower class communities. Not only
does nonresidential property use tend to lower property values, but
also neighborhood and community policing programs have increas-
ingly moved to locate police substations in such neighborhoods.
Therefore, it may be that crime rates are higher in these areas and resi-
dents objectively perceive themselves to be less safe. Finally, as sug-
gested by some scholars (Johnson, 1987), individuals may employ
types of self-protection because they believe the official agents of
community social control are unable to adequately deter potential
offenders.
Turning to the indicator living near fast food restaurants, we see
that students who live near fast food restaurants have 27% lower odds
of using guardianship to prevent crime than students who do not live
near fast food restaurants. Certainly, research has pointed out that fast
food restaurants are places where many people come and go (espe-
cially youth in their crime prone years); thus, living near fast food res-
taurants is an indication of being in closer proximity to potential
offenders by virtue of where individuals live. However, it appears that
these students do not perceive fast food restaurants to be places where
crime may occur or criminals congregate, thus explaining why they
are less likely to take measures to protect themselves.
Continuing, persons who live in communities with unsupervised
youth are 23% more likely to use forms of self-protection than persons
in communities without these characteristics. This was expected
because this neighborhood condition is likely to increase neighbor-
hood disorder and contribute to a negative atmosphere in the commu-
nity, thus increasing the frequency that crime and other social prob-
lems will occur (McGarrell et al., 1997; Skogan, 1990). It is also
possible that residents in neighborhoods with this condition feel the
need to protect themselves from the unsupervised youth, as these
same youth may be perceived as potential offenders. Drawing
together findings concerning neighborhood structures and neighbor-
hood conditions also shows support for the earlier conclusion that
actual proximity to potential offenders rather than fear of crime is an
important element in the use of self-protection.
Moving on, we find that only one lifestyle behavior encompassing
drug and alcohol use is a significant predictor of the use of self-protec-
318 CRIMINAL JUSTICE AND BEHAVIOR
sure. Specifically, the odds ratios suggest that persons who spend a
greater percentage of time during the week with strangers have 1%
lower odds of using self-protection than do those students who spend
less time with strangers during the week. Conversely, those students
who spend a greater percentage of time on the weekend with strangers
have a 1% greater odds of using self-protection than do students who
spend the weekend with people with whom they are more familiar.
These contrary associations are interesting. Previous research in rou-
tine activity theory finds that persons who are exposed to strangers
more of the time are more likely to be victimized. As such, the latter
finding is in accordance with expectations as this research suggests
that being in the proximity of potential offenders on the weekend (and
having heightened victimization risks) is also associated with the
increased use of self-protection or guardianship. However, this analy-
sis also finds that students who are in the company of strangers during
the week are not more likely to use self-protection but rather are less
likely. It may be that students do not perceive their associations during
the week as having a large impact on their likelihood for victimization.
Perhaps participants see these strangers at work, in class, or walking
around on campus and they do not perceive themselves to be at greater
risk for victimization. However, when they are in the company of peo-
ple they do not know on the weekend, they are more concerned with
their personal safety. Perhaps they are more likely to see these strang-
ers at places where perceptions about their risks are higher (bars,
entertainment clubs, parties, and so forth), or they may perceive that
crime is more of a problem on the weekend.
Finally, students who when going out for leisure are more fre-
quently on foot have 19% greater odds of using self-protection than do
students who use other forms of transportation when going out for lei-
sure. Similar to spending increased time with strangers, students who
are frequently out walking are exposing themselves to dangerous cir-
cumstances or potential offenders. Taking these relationships between
lifestyles that increase exposure and decisions to use guardianship
again we find even stronger support for the contention that it is objec-
tive measures of exposure and proximity to potential offenders that
influence persons’ use of self-protection rather than more indirect
measures of exposure such as fear and perceptions of safety.
320 CRIMINAL JUSTICE AND BEHAVIOR
CONCLUSIONS
The major conclusion to arise from this article is that when examin-
ing lifestyle characteristics and behaviors and their association with
persons’ use of self-protective measures, objective measures of prox-
imity and exposure are the most robust group of predictors. This offers
important implications for routine activity theory. Since inception, the
theory has been used to explain victimization risks and, more recently,
lifestyles associated with offending behaviors. What is proposed here
is an examination concerning whether this theory is equally dynamic
at explaining the lifestyles that are connected to the choice to use self-
protective behaviors to avoid victimization. Evidence presented here
suggests that this is a valid and fruitful endeavor. As such, taking this
and previous research together, routine activity theory can be used to
explain the lifestyles associated with each of the essential elements of
criminal incidents: potential offenders, suitable targets, and the pres-
ence (or absence) of guardianship.
To elaborate, this analysis found that the students’ use of self-
protective devices—mace, clubs or bats, knives, body alarms, guns—
is dependent on their employment status, transportation activities
(walking to leisure-related activities), frequency of associating with
strangers, living in disordered neighborhoods, use of crack, and per-
ceptions of the safety of their homes. These findings are highly infor-
mative for routine activity theory, which usually examines the pres-
ence (or absence) of guardianship and its impact on the criminal
incident without considering the status characteristics and lifestyle
behaviors of individuals who do or do not use it.
Previous research has examined individual characteristics and how
they relate to the use of self-protection. Most of this research, how-
ever, has focused on demographic characteristics. Although these
analyses are informative, it is likely that a person’s use of self-protec-
tion is based on more than just their socially defined statuses. Cer-
tainly, individuals are aware of the risks for victimization that are pres-
ent in society (and in their own communities) and must make
decisions about how to confront these risks. Theoretically, these deci-
sions are not made in a vacuum but rather are based on individuals’
lifestyles and how often they lead one into potentially dangerous situ-
ations. This analysis finds support for this proposition; when consid-
Tewksbury, Mustaine / LIFESTYLES AND GUARDIANSHIP 321
