Cohen and Felson 1979 Routine Activities

Social Change and Crime Rate Trends: A Routine Activity Approach
Author(s): Lawrence E. Cohen and Marcus Felson

Source: American Sociological Review, Vol. 44, No. 4 (Aug., 1979), pp. 588-608
Published by: American Sociological Association
Stable URL: https://www.jstor.org/stable/2094589
Accessed: 03-10-2019 01:18 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms
American Sociological Association is collaborating with JSTOR to digitize, preserve and

extend access to American Sociological Review
This content downloaded from 128.95.71.166 on Thu, 03 Oct 2019 01:18:09 UTC
All use subject to https://about.jstor.org/terms
SOCIAL CHANGE AND CRIME RATE TRENDS:
A ROUTINE ACTIVITY APPROACH*
LAWRENCE E. COHEN AND MARCUS FELSON

University of Illinois, Urbana
American Sociological Review 1979, Vol. 44 (August):588-608,
In this paper we present a "routine activity approach" for analyzing crime rate trends and
cycles. Rather than emphasizing the characteristics of offenders, with this approach we
concentrate upon the circumstances in which they carry out predatory criminal acts. Most
criminal acts require convergence in space and time of likely offenders, suitable targets and the
absence of capable guardians against crime. Human ecological theory facilitates an
investigation into the way in which social structure produces this convergence, hence allowing
illegal activities to feed upon the legal activities of everyday life. In particular, we hypothesize
that the dispersion of activities away from households and families increases the opportunity for
crime and thus generates higher crime rates. A variety of data is presented in support of the
hypothesis, which helps explain crime rate trends in the United States 1947-1974 as a byproduct
of changes in such variables as labor force participation and single-adult households.
INTRODUCTION family income during the same period. Also

during the same period the number of per-
In its summary report the National sons living below the legally-defined poverty
Commission on the Causes and Preven- level in cities declined from 11.3 million to
tion of Violence (1969: xxxvii) presents an 8.3 million.
important sociological paradox:
Why, we must ask, have urban violent Despite the general continuation of
crime rates increased substantially during these trends in social and economic condi-
the past decade when the conditions that are tions in the United States, the Uniform
supposed to cause violent crime have not Crime Report (FBI, 1975:49) indicates
worsened-have, indeed, generally im- that between 1960 and 1975 reported rates
proved? of robbery, aggravated assault, forcible
The Bureau of the Census, in its latest rape and homicide increased by 263%,
report on trends in social and economic con-
164%, 174%, and 188%, respectively.
ditions in metropolitan areas, states that
Similar property crime rate increases re-
most "indicators of well-being point toward
ported during this same period' (e.g.,
progress in the cities since 1960." Thus, for
example, the proportion of blacks in cities 200%o for burglary rate) suggest that the
who completed high school rose from 43 paradox noted by the Violence Commis-
percent in 1960 to 61 percent in 1968; unem- sion applies to nonviolent offenses as
ployment rates dropped significantly be- well.
tween 1959 and 1967 and the median family
income of blacks in cities increased from 61
I Though official data severely underestimate
percent to 68 percent of the median white
crime, they at least provide a rough indicator of
trends over time in the volume of several major
* Address all communications to: Lawrence felonies.
E. The possibility that these data also reflect
Cohen; Department of Sociology; University of trends
Illi-in rates at which offenses are reported to the
nois; Urbana, IL 61801. police has motivated extensive victimology research
For their comments, we thank David J. Bordua, (see Nettler, 1974; and Hindelang, 1976, for a re-
Ross M. Stolzenberg, Christopher S. Dunn, Kenneth view). This work consistently finds that seriousness
C. Land, Robert Schoen, Amos Hawley, and an of offense is the strongest determinant of citizen
anonymous reviewer. Funding for this study was reporting to law enforcement officials (Skogan, 1976:
provided by these United States Government grants: 145; Hindelang, 1976: 401). Hence the upward trend
National Institute for Mental Health 1-RO1- in official crime rates since 1960 in the U.S. may
MH31117-0l;National Science Foundation, SOC- reflect increases in both the volume and seriousness
77-13261; and United States Army RI/DAHC 19- of offenses. Though disaggregating these two com-
76-G-0016. The authors' name order is purely alpha- ponents may not be feasible, one may wish to inter-
betical. pret observed trends as generated largely by both.
588
SOCIAL CHANGE AND CRIME RATE TRENDS 589
In the present paper we consider these made little progress since Shaw and
paradoxical trends in crime rates in terms McKay and their colleagues (1929) pub-
of changes in the "routine activities" of lished their pathbreaking work, Delin-
everyday life. We believe the structure of quency Areas. Variations in crime rates
such activities influences criminal oppor- over space long have been recognized
tunity and therefore affects trends in a (e.g., see Guerry, 1833; Quetelet, 1842),
class of crimes we refer to as direct- and current evidence indicates that the
contact predatory violations. Predatory pattern of these relationships within met-
violations are defined here as illegal acts ropolitan communities has persisted
in which "someone definitely and inten- (Reiss, 1976). Although most spatial research
tionally takes or damages the person or is quite useful for describing crime rate
property of another" (Glaser, 1971:4). patterns and providing post hoc explana-
Further, this analysis is confined to those tions, these works seldom consider-
predatory violations involving direct phys- conceptually or empirically-the funda-
ical contact between at least one offender mental human ecological character of il-
and at least one person or object which legal acts as events which occur at specific
that offender attempts to take or damage. locations in space and time, involving
We argue that structural changes in specific persons and/or objects. These and
routine activity patterns can influence related concepts can help us to develop an
crime rates by affecting the convergence extension of the human ecological
in space and time of the three minimal analysis to the problem of explaining
elements of direct-contact predatory vio- changes in crime rates over time. Unlike
lations: (1) motivated offenders, (2) suita- many criminological inquiries, we do not
ble targets, and (3) the absence of capable examine why individuals or groups are in-
guardians against a violation. We further clined criminally, but rather we take crim-
argue that the lack of any one of these inal inclination as given and examine the
elements is sufficient to prevent the suc- manner in which the spatio-temporal
cessful completion of a direct-contact organization of social activities helps
predatory crime, and that the convergence people to translate their criminal inclina-
in time and space of suitable targets and tions into action. Criminal violations are
the absence of capable guardians may treated here as routine activities which
even lead to large increases in crime rates share many attributes of, and are interde-
without necessarily requiring any increase pendent with, other routine activities.
in the structural conditions that motivate This interdependence between the struc-
individuals to engage in crime. That is, if ture of illegal activities and the organiza-
the proportion of motivated offenders or tion of everyday sustenance activities
even suitable targets were to remain stable leads us to consider certain concepts from
in a community, changes in routine activi- human ecological literature.
ties could nonetheless alter the likelihood
of their convergence in space and time,
thereby creating more opportunities for
Selected Concepts froml Hawley's Human
crimes to occur. Control therefore be-
Ecological Theory
comes critical. If controls through routine
activities were to decrease, illegal pred- While criminologists traditionally have
atory activities could then be likely to in- concentrated on the spatial analysis of
crease. In the process of developing this crime rates within metropolitan com-
explanation and evaluating its consistency munities, they seldom have considered
with existing data, we relate our approach the temporal interdependence of these
to classical human ecological concepts acts. In his classic theory of human ecol-
and to several earlier studies. ogy, Amos Hawley (1950) treats the com-
munity not simply as a unit of territory but
rather as an organization of symbiotic and
The Structure of Criminll A ctivitY
commensalistic relationships as human
Sociological knowledge of how com- activities are performed over both space
munity structure generates illegal acts has and time.
590 AMERICAN SOCIOLOGICAL REVIEW
Hawley identified three important tem- lated social roles and relationships to the
poral components of community struc- occurrence or absence of illegal acts.
ture: (1) rhythm, the regular periodicity The conjunction of these minimal ele-
with which events occur, as with the ments can be used to assess how social
rhythm of travel activity; (2) tempo, the structure may affect the tempo of each
number of events per unit of time, such as type of violation. That is, the probability
the number of criminal violations per day that a violation will occur at any specific
on a given street; and (3) timing, the coor- time and place might be taken as a func-
dination among different activities which tion of the convergence of likely offenders
are more or less interdependent, such as and suitable targets in. the absence of
the coordination of an offender's rhythms capable guardians. Through consideration
with those of a victim (Hawley, 1950:289; of how trends and fluctuations in social
the examples are ours). These compo- conditions affect the frequency of this
nents of temporal organization, often ne- convergence of criminogenic circum-
glected in criminological research, prove stances, an explanation of temporal trends
useful in analyzing how illegal tasks are in crime rates can be constructed.
performed-a utility which becomes more
apparent after noting the spatio-temporal
The Ecological Nature of Illegal Acts
requirements of illegal activities.
This ecological analysis of direct-
The Minimal Elements of Direct-Contact contact predatory violations is intended to
Predatory Violations be more than metaphorical. In the context
of such violations, people, gaining and los-
As we previously stated, despite their
ing sustenance, struggle among them-
great diversity, direct-contact predatory
selves for property, safety, territorial
violations share some important require-
hegemony, sexual outlet, physical con-
ments which facilitate analysis of their
trol, and sometimes for survival itself. The
structure. Each successfully completed
interdependence between offenders and
violation minimally requires an offender
victims can be viewed as a predatory rela-
with both criminal inclinations and the
tionship between functionally dissimilar
ability to carry out those inclinations, a
individuals or groups. Since predatory
person or object providing a suitable
violations fail to yield any net gain in
target for the offender, and absence of
sustenance for the larger community, they
guardians capable of preventing viola-
can only be sustained by feeding upon
tions. We emphasize that the lack of any
other activities. As offenders cooperate to
one of these elements normally is suffi-
increase their efficiency at predatory vio-
cient to prevent such violations from
lations and as potential victims organize
occurring.2 Though guardianship is im-
their resistance to these violations, both
plicit in everyday life, it usually is marked
groups apply the symbiotic principle to
by the absence of violations; hence it is
improve their sustenance position. On the
easy to overlook. While police action 'is
other hand, potential victims of predatory
analyzed widely, guardianship by ordi-
crime may take evasive actions which en-
nary citizens of one another and of prop-
courage offenders to pursue targets other
erty as they go about routine activities
than their own. Since illegal activities
may be one of the most neglected ele-
must feed upon other activities, the spatial
ments in sociological research on crime,
and temporal structure of routine legal ac-
especially since it links seemingly unre-
tivities should play an important role in
determining the location, type and quan-
2 The analytical distinction between target and
guardian is not important in those cases where a
tity of illegal acts occurring in a given
personal target engages in self-protection from community or society. Moreover, one can
direct-contact predatory violations. We leave open analyze how the structure of community
for the present the question of whether a guardian isorganization as well as the level of
effective or ineffective in all situations. We also
technology in a society provide the cir-
allow that various guardians may primarily supervise
offenders, targets or both. These are questions for cumstances under which crime can thrive.
future examination. For example, technology and organization
affect the capacity of persons with crimi- be of central importance for explaining
nal inclinations to overcome their targets, crime rates.
as well as affecting the ability of guardians The ideas presented so far are not new,
to contend with potential offenders by but they frequently are overlooked in the
using whatever protective tools, weapons theoretical literature on crime. Although
and skills they have at their disposal. an investigation of the literature uncovers
Many technological advances designed for significant examples of descriptive and
legitimate purposes-including the auto- practical data related to the routine activi-
mobile, small power tools, hunting ties upon which illegal behavior feeds,
weapons, highways, telephones, etc.- these data seldom are treated within an
may enable offenders to carry out their analytical framework. The next section
own work more effectively or may assist reviews some of this literature.
people in protecting their own or someone
else's person or property.
RELATION OF THE ROUTINE ACTIVITY
Not only do routine legitimate activities
APPROACH TO EXTANT STUDIES
often provide the wherewithal to commit
offenses or to guard against others who do A major advantage of the routine ac-
so, but they also provide offenders with tivity approach presented here is that it
suitable targets. Target suitability is likely helps assemble some diverse and previ-
to reflect such things as value (i.e., the ously unconnected criminological
material or symbolic desirability of a per- analyses into a single substantive frame-
sonal or property target for offenders), work. This framework also serves to link
physical visibility, access, and the inertia illegal and legal activities, as illustrated by
of a target against illegal treatment by of- a few examples of descriptive accounts of
fenders (including the weight, size, and criminal activity.
attached or locked features of property
inhibiting its illegal removal and the phys-
Descriptive A analyses
ical capacity of personal victims to resist
attackers with or without weapons). There are several descriptive analyses
Routine production activities probably af- of criminal acts in criminological litera-
fect the suitability of consumer goods for ture. For example, Thomas Reppetto's
illegal removal by determining their value (1974) study, Residential Crime, considers
and weight. Daily activities may affect the how residents supervise their neighbor-
location of property and personal targets hoods and streets and limit access of
in visible and accessible places at particu- possible offenders. He also considers how
lar times. These activities also may cause distance of households from the central
people to have on hand objects that can be city reduces risks of criminal victimiza-
used as weapons for criminal acts or self- tion. Reppetto's evidence-consisting of
protection or to be preoccupied with tasks criminal justice records, observations of
which reduce their capacity to discourage comparative features of geographic areas,
or resist offenders. victimization survey data and offender
While little is known about conditions interviews-indicates that offenders are
that affect the convergence of potential very likely to use burglary tools and to
offenders, targets and guardians, this is a have at least minimal technical skills, that
potentially rich source of propositions physical characteristics of dwellings affect
about crime rates. For example, daily their victimization rates, that the rhythms
work activities separate many people from of residential crime rate patterns are
those they trust and the property they marked (often related to travel and work
value. Routine activities also bring to- patterns of residents), and that visibility of
gether at various times of day or night potential sites of crime affects the risk that
persons of different background, some- crimes will occur there. Similar findings
times in the presence of facilities, tools or are reported by Pope's (1977a; 1977b)
weapons which influence the commission study of burglary in California and by
or avoidance of illegal acts. Hence, the Scarr's (1972) study of burglary in and
timing of work, schooling and leisure may around the District of Columbia. In addi-
tion, many studies report that architec- view of the literature on ecological studies
tural and environmental design as well as of crimes, see Wilks, 1967). Recently, Al-
community crime programs serve to de- bert Reiss (1976) has argued convincingly
crease target suitability and increase that these spatial variations (despite some
capable guardianship (see, for example, claims to the contrary) have been sup-
Newman, 1973; Jeffery, 1971; Washnis, ported consistently by both official and
1976), while many biographical or au- unofficial sources of data. Reiss further
tobiographical descriptions of illegal ac- cites victimization studies which indicate
tivities note that lawbreakers take into ac- that offenders are very likely to select
count the nature of property and/or the targets not far from their own residence
structure of human activities as they go (see USDJ, 1974a; 1974b; 1974c).
about their illegal work (see, e.g.,
Chambliss, 1972; Klockars, 1974; Suther-
Macrolevel Analyses of Crime Trends
land, 1937; Letkemann, 1973; Jackson,
and Cycles
1969; Martin, 1952; Maurer, 1964; Came-
ron, 1964; Williamson, 1968). Although details about how crime oc-
Evidence that the spatio-temporal orga- curs are intrinsically interesting, the im-
nization of society affects patterns of portant analytical task is to learn from
crime can be found in several sources. these details how illegal activities carve
Strong variations in specific predatory their niche within the larger system of ac-
crime rates from hour to hour, day to day, tivities. This task is not an easy one. For
and month to month are reported often example, attempts by Bonger (1916),
(e.g., Wolfgang, 1958; Amir, 1971; Rep- Durkheim (1951; 1966), Henry and Short
petto, 1974; Scarr, 1972; FBI, 1975; (1954), and Fleisher (1966) to link the rate
1976), and these variations appear to cor- of illegal activities to the economic condi-
respond to the various tempos of the re- tion of a society have not been completely
lated legitimate activities upon which they successful. Empirical tests of the relation-
feed. Also at a microsociological level, ships postulated in the above studies have
Short and Strodtbeck (1965: chaps. 5 and produced inconsistent results which some
11) describe opportunities for violent con- observers view as an indication that the
frontations of gang boys and other com- level of crime is not related systematically
munity residents which arise in the con- to the economic conditions of a society
text of community leisure patterns, such (Mansfield et al., 1974: 463; Cohen and
as "quarter parties" in black com- Felson, 1979).
munities, and the importance, in the cal- It is possible that the wrong economic
culus of decision making employed by and social factors have been employed in
participants in such episodes, of low these macro studies of crime. Other re-
probabilities of legal intervention. In addi- searchers have provided stimulating alter-
tion, a wealth of empirical evidence indi- native descriptions of how social change
cates strong spatial variations over com- affects the criminal opportunity structure,
munity areas in crime and delinquency thereby influencing crime rates in particu-
rates3 (for an excellent discussion and re- lar societies. For example, at the begin-
ning of the nineteenth century, Patrick
3 One such ecological study by Sarah Boggs (1965)
Colquhoun (1800) presented a detailed,
presents some similar ideas in distinguishing farni/- lucid description and analysis of crime in
iarity of offenders with their targets and profitabilitythe London metropolitan area and sugges-
of targets as two elements of crime occurrence. tions for its control. He assembled sub-
Boggs's work stands apart from much research on
stantial evidence that London was experi-
the ecology of crime in its consideration of crime
occurrence rates separately from offender rates. The encing a massive crime wave attributable
former consist of the number of offenses committed to a great increment in the assemblage and
in a given area per number of suitable targets within
that area (as estimated by various indicators). The
latter considers the residence of offenders in comput- senses in St. Louis and shows that the two are often
ing the number of offenders per unit of population. independent. It appears from her analysis that both
Boggs examines the correlations between crime target and offender characteristics play a central role
occurrence rates and offender rates for several of- in the location of illegal activity.
movement of valuable goods through its To do so requires a conceptual framework

ports and terminals. such as that sketched in the preceding sec-
A similar examination of crime in the tion. Before attempting to demonstrate
period of the English industrial expansion the feasibility of this approach with ma-
was carried out by a modem historian, crolevel data, we examine available mi-
J. J. Tobias (1967), whose work on the crolevel data for its consistency with the
history of crime in nineteenth century En- major assumptions of this approach.
gland is perhaps the most comprehensive
effort to isolate those elements of social
Microlevel Assumptions of the Routine
change affecting crime in an expanding
Activity Approach
industrial nation. Tobias details how far-
reaching changes in transportation, cur- The theoretical approach taken here
rency, technology, commerce, merchan- specifies that crime rate trends in the
dising, poverty, housing, and the like, had post-World War II United States are re-
tremendous repercussions on the amount lated to patterns of what we have called
and type of illegal activities committed in routine activities. We define these as any
the nineteenth century. His thesis is that recurrent and prevalent activities which
structural transformations either provide for basic population and indi-
facilitated or impeded the opportunities to vidual needs, whatever their biological or
engage in illegal activities. In one of the cultural origins. Thus routine activities
few empirical studies of how recent social would include formalized work, as well as
change affects the opportunity structure the provision of standard food, shelter,
for crime in the United States, Leroy sexual outlet, leisure, social interaction,
Gould (1969) demonstrated that the in- learning and childrearing. These activities
crease in the circulation of money and the may go well beyond the minimal levels
availability of automobiles between 1921 needed to prevent a population's extinc-
and 1965 apparently led to an increase in tion, so long as their prevalence and re-
the rate of bank robberies and auto thefts, currence makes them a part of everyday
respectively. Gould's data suggest that life.
these relationships are due more to the Routine activities may occur (1) at
abundance of opportunities to perpetrate home, (2) in jobs away from home, and (3)
the crimes than to short-term fluctuations in other activities away from home. The
in economic activities. latter may involve primarily household
Although the sociological and historical members or others. We shall argue that,
studies cited in this section have provided since World War II, the United States has
some useful empirical generalizations and experienced a major shift of routine activi-
important insights into the incidence of ties away from the first category into the
crime, it is fair to say that they have not remaining ones, especially those
articulated systematically the theoretical nonhousehold activities involving
linkages between routine legal activities nonhousehold members. In particular, we
and illegal endeavors. Thus, these studies shall argue that this shift in the structure
cannot explain how changes in the larger of routine activities increases the probabil-
social structure generate changes in the ity that motivated offenders will converge
opportunity to engage in predatory crime in space and time with suitable targets in
and hence account for crime rate trends.4 the absence of capable guardians, hence
contributing to significant increases in the
4 The concept of the opportunity for crime con-

given at different points in the social structure (Clo-
tained in the above research and in this study differs
ward and Ohlin, 1960: 15 1). Rather than circumstan-
considerably from the traditional sociological usage
tial determinants of crime, they use differential
of the cfifferential opportunitiA concept. For example,
Cloward and Ohlin (1960) employed this term in dis- opportunity to emphasize structural features which
cussing how legitimate and illegitimate opportunities motivate offenders to perpetrate certain types of
affect the resolution of adjustment problems leading crimes. Cloward and Ohlin are largely silent on the
to gang delinquency. From their viewpoint, this reso- interaction of this motivation with target suitability
lution depends upon the kind of social support for and guardianship as this interaction influences crime
one or another type of illegitimate activity that is rates.
direct-contact predatory crime rates over National estimates from large-scale

these years. government victimization surveys in 1973
If the routine activity approach is valid, and 1974 support this generalization (see
then we should expect to find evidence for methodological information in Hindelang
a number of empirical relationships re- et al., 1976: Appendix 6). Table I presents
garding the nature and distribution of several incident-victimization rates per
predatory violations. For example, we 100,000 population ages 12 and older.
would expect routine activities performed Clearly, the rates in Panels A and B are far
within or near the home and among family lower at or near home than elsewhere and
or other primary groups to entail lower far lower among relatives than others. The
risk of criminal victimization because they data indicate that risk of victimization var-
enhance guardianship capabilities. We ies directly with social distance between
should also expect that routine daily ac- offender and victim. Panel C of this table
tivities affect the location of property and indicates, furthermore, that risk of lone
personal targets in visible and accessible victimization far exceeds the risk of vic-
places at particular times, thereby in- timization for groups. These relationships
fluencing their risk of victimization. Fur- are strengthened by considering time
thermore, by determining their size and budget evidence that, on the average,
weight and in some cases their value, Americans spend 16.26 hours per day at
routine production activities should affect home, 1.38 hours on streets, in parks,
the suitability of consumer goods for il- etc., and 6.36 hours in other places
legal removal. Finally, if the routine ac- (Szalai, 1972:795). Panel D of Table 1
tivity approach is useful for explaining the presents our estimates of victimization per
paradox presented earlier, we should find billion person-hours spent in such loca-
that the circulation of people and prop- tions.6 For example, personal larceny
erty, the size and weight of consumer
items etc., will parallel changes in crime indicate that nonfamily members are usually much
rate trends for the post-World War II more dangerous than family members are to each
other (see text). Also, when family violence leads to
United States.
death, its suppression becomes quite difficult. The
The veracity of the routine activity ap- murder circumstances data indicate that about two-
proach can be assessed by analyses of thirds of killings involve nonrelatives. Without deny-
both microlevel and macrolevel inter- ing the evidence that the level of family violence is
dependencies of human activities. While far greater than police reports would indicate, avail-
able data also suggest that time spent in family activi-
consistency at the former level may ap-
ties within households incurs less risk of victimiza-
pear noncontroversial, or even obvious, tion than many alternative activities in other places.
one nonetheless needs to show that the In addition, many of the most common offenses
approach does not contradict existing data (such as robbery and burglary) always have been
recognized as usually involving nonfamily members.
before proceeding to investigate the latter
6 Billion person-hours can easily be concep-
level.
tualized as 1,000,000 persons spending 1,000 hours
EMPIRICAL ASSESSMENT
each (or about 42 days) in a given location (Szalai,
1972:795). Fox obtained these data from a 1966 time
Circumstances and Location of Offenses budget study in 44 American cities. The study was
carried out by the Survey Research Center, the Uni-
The routine activity approach specifies versity of Michigan. We combined four subsamples
that household and family activities entail in computing our figures. We combined activities
lower risk of criminal victimization than into three locations, as follows: (1) at or just outside
home; (2) at another's home, restaurants or bars, or
nonhousehold-nonfamily activities, de-
indoor leisure; (3) in streets, parks, or outdoor lei-
spite the problems in measuring the sure. Our computing formula was
former.5
Q = [(R -105) (A .365)] * 109,
5 Recent research indicates the existence of sub- where Q is the risk per billion person-hours; R is the
stantial quantities of family violence which remains victimization rate, reported per 10- persons in Hin-
outside of UCR data (see annotated bibliography of delang et al. (1976: Table 318); A is the hours spent
family violence in Lystad, 1974). While we cannot per location calculated from Szalai (1972: 795); 365 is
rule out the likelihood that much family violence is the multiplier to cover a year's exposure to risk; and
concealed from victimization surveys, the latter cap- 109 converts risk per person-hour to billion person-
ture information absent from police data and still hours.
Table 1. Incident-Specific Risk Rates for Rape, Robbery, Assault and Personal Larceny with Contact,
United States, 1974
Personal
Larceny
A.* Rape Robbery Assault with Contact Total

PLACE OF In or near home 63 129 572 75 839
RESIDENCE Elsewhere 119 584 1,897 1,010 3,610
B.
VICTIM- (Lone Offender)
OFFENDER Relative 7 13 158 5 183
RELATIONSHIP Well Known 23 30 333 30 416
Casual Acquaintance 11 26 308 25 370
Don't Know/Sight Only 106 227 888 616 1,837
(Multiple Offender)
Any known 10*** 68 252 43 373
All strangers 25*** 349 530 366 1,270
C.*
NUMBER one 179 647 2,116 1,062 4,004
OF Two 3 47 257 19 326
VICTIMS Three 0 13 53 3 09
Four Plus 0 6 43 1 50
D.**
LOCATION AND Home, Stranger 61 147 345 103 654
RELATIONSHIP Home, Nonstranger 45 74 620 22 761
(sole Street, Stranger 1,370 7,743 15,684 7,802 32,460
offender Street, Nonstranger 179 735 5,777 496 7,167
only) Elsewhere, Stranger 129 513 1,934 2,455 4,988
Elsewhere, Nonstranger 47 155 1,544 99 1,874
* Calculated from Handelang et al., 1977: Tables 3.16

ages 12 and over.
** See fn. 6 for source. Rates are per billion person-hours in stated locations.
*** Based on white data only due to lack of suitable sample size for nonwhites as victims of rape with
multiple offenders.
rates (with contact) are 350 times higher at to disaggregate all major components of
the hands of strangers in streets than at target suitability (i.e., value, visibility, ac-
the hands of nonstrangers at home. Sepa- cessibility and inertia), together they
rate computations from 1973 victimization imply that expensive and movable dura-
data (USDJ, 1976: Table 48) indicate that bles, such as vehicles and electronic
there were two motor vehicle thefts per appliances, have the highest risk of illegal
million vehicle-hours parked at or near removal.
home, 55 per million vehicle-hours in As a specific case in point, we com-
streets, parks, playgrounds, school pared the 1975 composition of stolen
grounds or parking lots, and 12 per million property reported in the Uniform Crime
vehicle-hours elsewhere. While the direc- Report (FBI, 1976: Tables 26-7) with na-
tion of these relationships is not surpris- tional data on personal consumer expendi-
ing, their magnitudes should be noted. It tures for goods (CEA, 1976: Tables 13-16)
appears that risk of criminal victimization and to appliance industry estimates of the
varies dramatically among the circum- value of shipments the same year (Mer-
stances and locations in which people chandising Week, 1976). We calculated
place themselves and their property. that $26.44 in motor vehicles and parts
were stolen for each $100 of these goods
consumed in 1975, while $6.82 worth of
Target Suitability
electronic appliances were stolen per $100
Another assumption of the routine ac- consumed. Though these estimates are
tivity approach is that target suitability in- subject to error in citizen and police esti-
fluences the occurrence of direct-contact mation, what is important here is their size
predatory violations. Though we lack data relative to other rates. For example, only
8V worth of nondurables andcriminal

predatory 12Vvictimization.
worth We also
of furniture and nonelectronic household expect that adolescents and young adults
durables were stolen per $100 of each who are perhaps more likely to engage in
category consumed, the motor vehicle peer group activities rather than family
risk being, respectively, 330 and 220 times activities will have higher rates of criminal
as great. Though we lack data on the victimization. Finally' married persons
"stocks" of goods subject to risk, these should have lower rates than others. Ta-
"flow" data clearly support our assump- bles 2 and 3 largely confirm these expecta-
tion that vehicles and electronic tions (with the exception of personal lar-
appliances are greatly overrepresented in ceny with contact). Examining these ta-
thefts. bles, we note that victimization rates ap-
The 1976 Buying Guide issue of Con- pear to be related inversely to age and are
sumer Reports (1975) indicates why elec- lower for persons in "less active" statuses
tronic appliances are an excellent retail (e.g., keeping house, unable to work, re-
value for a thief. For example, a tired) and persons in intact marriages. A
Panasonic car tape player is worth $30 per notable exception is indicated in Table 2,
lb., and a Phillips phonograph cartridge is where persons unable to work appear
valued at over $5,000 per lb., while large more likely to be victimized by rape, rob-
appliances such as refrigerators and wash- bery and personal larceny with contact
ing machines are only worth $1 to $3 per than are other "inactive persons." Un-
lb. Not surprisingly, burglary data for the employed persons also have unusually
District of Columbia in 1969 (Scarr, 1972: high rates of victimization. However,
Table 9) indicate that home entertainment these rates are consistent with the routine
items alone constituted nearly four times activity approach offered here: the high
as many stolen items as clothing, food, rates of victimization suffered by the un-
drugs, liquor, and tobacco combined and employed may reflect their residential
nearly eight times as many stolen items as proximity to high concentrations of poten-
office supplies and equipment. In addi- tial offenders as well as their age and ra-
tion, 69o of national thefts classified in cial composition, while handicapped per-
1975 (FBI, 1976: Tables 1, 26) involve sons have high risk of personal victimiza-
automobiles, their parts or accessories, tion because they are less able to resist
and thefts from automobiles or thefts of motivated offenders. Nonetheless, per-
bicycles. Yet radio and television sets plus sons who keep house have noticeably
electronic components and accessories to- lower rates of victimization than those
taled only 0.10%o of the total truckload who are employed, unemployed, in school
tonnage terminated in 1973 by intercity or in the armed forces.
motor carriers, while passenger cars, As Table 3 indicates, burglary and rob-
motor vehicle parts and accessories, bery victimization rates are about twice as
motorcycles, bicycles, and their parts, to- high for persons living in single-adult
taled only 5.5% of the 410 million households as for other persons in each
truckload tons terminated (ICC, 1974). age group examined. Other victimization
Clearly, portable and movable durables data (USDJ, 1976: Table 21) indicate that,
are reported stolen in great disproprotion while household victimization rates tend
to their share of the value and weight of to vary directly with household size,
goods circulating in the United States. larger households have lower rates per
person. For example, the total household
victimization rates (including burglary,
Family Activities and Crime Rates
household larceny, and motor vehicle
One would expect that persons living in theft) per 1,000 households were 168 for
single-adult households and those em- single-person households and 326 for
ployed outside the home are less obligated households containing six or more per-
to confine their time to family activities sons. Hence, six people distributed over
within households. From a routine ac- six single-person households experience
tivity perspective, these persons and their an average of 1,008 household victimiza-
households should have higher rates of tions, more than three times as many as
Table 2. Selected Status-Specific Personal Victimization Rates for the United States (per 100,000 Persons in
Each Category)
Personal Personal
Variables Larceny Larceny
and Victim with without
Sources Category Rape Robbery Assault Contact Contact
A. AGE 12-15 147 1,267 3,848 311 16,355

(Source: 16-19 248 1,127 5,411 370 15,606
Hindelang, et al., 1977: 20-24 209 1,072 4,829 337 14,295
Table 310, 1974 25-34 135 703 3,023 263 10,354
rates 35-49 21 547 1,515 256 7,667
50-64 33 411 731 347 4,588
65+ 20 388 492 344 1,845
B. MAJOR (Male 16+)

ACTIVITY OF Armed Forces 1,388 4,153 118 16,274
VICTIM Employed 807 3,285 252 10,318
(Source: Unemployed 2,179 7,984 594 15,905
Hindelang, et al., 1977: Keep house 0 2,475 463 3,998
Table 313, 1974 In school 1,362 5,984 493 17,133
rates) Unable to work 1,520 2,556 623 3,648
Retired 578 662 205 2,080
(Female 16+
Keep house 116 271 978 285 4,433
Employed 156 529 1,576 355 9,419
Unemployed 798 772 5,065 461 12,338
In School 417 430 2,035 298 12,810
Unable to work 287 842 741 326 1,003
Retired 120 172 438 831 1,571
C. MARITAL STATUS (Male 12+)

(Source:USDJ: Never Married 1,800 5,870 450 16,450
1977, Table 5, Married 550 2,170 170 7,660
1973 rates) Separated/Divorced 2,270 5,640 1,040 12,960
Widowed 1,150 1,500 4,120
(Female 12+)
Never Married 360 580 2,560 400 12,880
Married 70 270 910 220 6,570
Separated/Divorced 540 1,090 4,560 640 9,130
Widowed 450 590 480 2,460
Line indicates too few offenses for accurate estimates of rate. However, rates in these cells are usually
small.
one six-person household. Moreover, age by persons under 20, the motor vehicle
of household head has a strong relation- theft rate is nine times as high, and the
ship to a household's victimization rate burglary and household larceny rates four
for these crimes. For households headed times as high as those for households
headed by persons 65 and over (USDJ,
1976: Table 9).
Table 3. Robbery-Burglary Victimization Rates by
Ages and Number of Adults in Household, While the data presented in this section
1974 and 1976 General Social Survey were not collected originally for the pur-
pose of testing the routine activity ap-
Number of Adults in Household
proach, our efforts to rework them for
Age One Two or More Ratio
these purposes have proven fruitful. The
18-35 0.200 (140) 0.095 (985) 2.11 routine activity approach is consistent
36-55 0.161 (112) 0.079 (826) 2.04
with the data examined and, in addition,
56 and over 0.107 (262) 0.061 (640) 1.76
helps to accommodate within a rather
All Ages 0.144 (514) 0.081 (2451) 1.78
simple and coherent analytical framework
(Numbers in parentheses are the base for comput- certain findings which, though not neces-
ing risk rates.)
sarily new, might otherwise be attributed
Source: Calculated from 1974 and 1976 General
Social Survey, National Opinion Research Center, only "descriptive" significance. In the
University of Chicago. next section, we examine macrosocial
trends as they relate to trends in crime Table 4. Proportion of Households Unattended by

rates. Anyone 14 Years Old or Over by Time of
Day during First Visit by Census Bureau
Interviewer, 1960 and 1971
CHANGING TRENDS IN ROUTINE ACTIVITY
November, 1971
STRUCTURE AND PARALLEL TRENDS IN 1960 Current Pop. Percent
CRIME RATES Time of day Census Survey Change
8:00- 8:59 a.m. 29% 43 +48.9o

The main thesis presented here is that
9:00- 9:59a.m. 29 44 +58
the dramatic increase in the reported 10:00-10:59 a.m. 31 42 +36
crime rates in the U.S. since 1960 is linked 11:00-11:59 a.m. 32 41 +28
to changes in the routine activity structure 12:00-12:59 p.m. 32 41 +28
1:00- 1:59 p.m. 31 43 +39
of American society and to a correspond-
2:00- 2:59 p.m. 33 43 +30
ing increase in target suitability and de- 3:00- 3:59 p.m. 30 33 +10
crease in guardian presence. If such a 4:00- 4:59 p.m. 28 30 + 7
thesis has validity, then we should be able 5:00- 5:59p.m. 22 26 +18
to identify these social trends and show 6:00- 6:59p.m. 22 25 +14
7:00- 7:50 p.m. 20 29 +45
how they relate to predatory criminal vic-
8:00- 8:59p.m. 24 22 - 8
timization rates.
Source: Calculated from USBC (1973b: Table A).
Trends in Human Activity Patterns

trends, tallying an 81% increase in the
The decade 1960-1970 experienced number of vacations taken by Americans
noteworthy trends in the activities of the from 1967 to 1972, a five-year period
American population. For example, the (USBC, 1973a: Introduction).
percent of the population consisting of The dispersion of activities away from
female college students increased 118% households appears to be a major recent
(USBC, 1975: Table 225). Married female social change. Although this decade also
labor force participant rates increased experienced an important 31% increase in
31% (USBC, 1975: Table 563), while the the percent of the population ages 15-24,
percent of the population living as primary age structure change was only one of
individuals increased by 34% (USBC, many social trends occurring during the
1975: Table 51; see also Kobrin, 1976). We period, especially trends in the circulation
gain some further insight into changing of people and property in American soci-
routine activity patterns by comparing ety.7
hourly data for 1960 and 1971 on house- The importance of the changing activity
holds unattended by persons ages 14 or structure is underscored by taking a brief
over when U.S. census interviewers first look at demographic changes between the
called (see Table 4). These data suggest years 1970 and 1975, a period of continu-
that the proportion of households un- ing crime rate increments. Most of the
attended at 8 A.M. increased by almost recent changes in age structure relevant to
half between 1960 and 1971. One also crime rates already had occurred by 1970;
finds increases in rates of out-of-town indeed, the proportion of the population
travel, which provides greater opportunity ages 15-24 increased by only 6% between
for both daytime and nighttime burglary of 1970 and 1975, compared with a 15% in-
residences. Between 1960 and 1970, there crease during the five years 1965 to 1970.
was a 72% increase in state and national On the other hand, major changes in the
park visits per capita (USBC, 1975), an structure of routine activities continued
144% increase in the percent of plant
workers eligible for three weeks vacation 7 While the more sophisticated treatments of the
(BLS, 1975: Table 116), and an 184% in- topic have varied somewhat in their findings, most
crease in overseas travellers per 100,000 recent studies attempting to link crime rate increases
to the changing age structure of the American popu-
population (USBC, 1975: Table 366). The
lation have found that the latter account for a rela-
National Travel Survey, conducted as tively limited proportion of the general crime trend
part of the U.S. Census Bureau's Census (see, for example, Sagi and Wellford, 1968; Fer-
of Transportation, confirms the general dinand, 1970; and Wellford, 1973).
during these years. For example, in only on the weight of many consumer durable
five years, the estimated proportion of the goods were examined. Sears is the na-
population consisting of husband-present, tion's largest retailer and its policy of pur-
married women in the labor force house- chasing and relabeling standard manufac-
holds increased by I11%, while the esti- tured goods makes its catalogs a good
mated number of non-husband-wife source of data on widely merchandised
households per 100,000 population in- consumer goods. The lightest television
creased from 9,150 to 11,420, a 25% in- listed for sale in 1960 weighed 38 lbs.,
crease (USBC, 1976: Tables 50, 276; compared with 15 lbs. for 1970. Thus, the
USBC, 1970-1975). At the same time, the lightest televisions were 21/2 times as
percent of population enrolled in higher heavy in 1960 as 1970. Similar trends are
education increased 16% between 1970 observed for dozens of other goods listed
and 1975. in the Sears catalog. Data from Consumer
Reports Buying Guide, published in De-
cember of 1959 and 1969, show similar
Related Property Trends and Their
changes for radios, record players, slide
Relation to Human Activity Patterns
projectors, tape recorders, televisions,
Many of the activity trends mentioned toasters and many other goods. Hence,
above normally involve significant in- major declines in weight between 1960 and
vestments in durable goods. For example, 1970 were quite significant for these and
the dispersion of population across rela- other goods, which suggests that the con-
tively more households (especially sumer goods market may be producing
non-husband-wife households) enlarges many more targets suitable for theft. In
the market for durable goods such as tele- general, one finds rapid growth in prop-
vision sets and automobiles. Women par- erty suitable for illegal removal and in
ticipating in the labor force and both men household and individual exposure to at-
and women enrolled in college provide a tack during the years 1960-1975.
market for automobiles. Both work and
travel often involve the purchase of major
Related Trends in Business
movable or portable durables and their
Establishments
use away from home.
Considerable data are available which Of course, as households and individu-
indicate that sales of consumer goods als increased their ownership of small
changed dramatically between 1960 and durables, businesses also increased the
1970 (as did their size and weight), hence value of the merchandise which they
providing more suitable property available transport and sell as well as the money
for theft. For example, during this decade, involved in these transactions. Yet the
constant-dollar personal consumer ex- Census of Business conducted in 1958,
penditures in the United States for motor 1963, 1967, and 1972 indicate that the
vehicles and parts increased by 71%, number of wholesale, retail, service, and
while constant-dollar expenditures for public warehouse establishments (includ-
other durables increased by 105% (calcu- ing establishments owned by large orga-
lated from CEA, 1976: Table B- 16). In nizations) was a nearly constant ratio of
addition, electronic household appliances one for every 16 persons in the United
and small houseware shipments increased States. Since more goods and money were
from 56.2 to 119.7 million units (Electrical distributed over a relatively fixed number
Merchandising Week, 1964; Merchandis- of business establishments, the tempo of
ing Week, 1973). During the same decade, business activity per establishment appar-
appliance imports increased in value by ently was increasing. At the same time,
681% (USBC, 1975: Table 1368). the percent of the population employed as
This same period appears to have sales clerks or salesmen in retail trade de-
spawned a revolution in small durable clined from 1.48% to 1.27%, between 1960
product design which further feeds the and 1970, a 14.7% decline (USBC, 1975:
opportunity for crime to occur. Relevant Table 589).
data from the 1960 and 1970 Sears catalogs Though both business and personal
- 600 AMERICAN SOCIOLOGICAL REVIEW
property increased, the changing pace of Table 5. Offense Analysis Trends for Robbery,
activities appears to have exposed the lat- Burglary, Larceny and Murder; United
States, 1960-1975
ter to greater relative risk of attack,
whether at home or elsewhere, due to the A. ROBBERIESa 1960 1965 1970
dispersion of goods among many more Highway Robbery 52.6 57.0 59.8
Residential Robbery 8.0 10.1 13.1
households, while concentrating goods in
Commercial Robbery 39.4 32.9 27.1
business establishments. However, mer-
Totals 100.0 100.0 100.0
chandise in retail establishments with
heavy volume and few employees to guard B. BURGLARIES 1960 1965 1970 1975
Residential 15.6 24.5 31.7 33.2
it probably is exposed to major increments
Residential Nightime 24.4 25.2 25.8 30.5
in risk of illegal removal than is most other Commercial 60.0 50.2 42.5 36.3
business property.
Totals 100.0 99.9 100.0 100.0
C. LARCENIES 1960 1965 1970 1975

Composition of Crime Trends Shoplifting 6.0 7.8 9.2 11.3
Other 94.0 92.2 90.8 88.7
If these changes in the circulation of
Totals 100.0 100.0 100.0 100.0
people and property are in fact related to
D. MURDERS 1963 1965 1970 1975
crime trends, the composition of the latter
Relative Killings 31.0 31.0 23.3 22.4
should reflect this. We expect relatively Romance, Argumentsb 51.0 48.0 47.9 45.2
greater increases in personal and house- Felon Typesc 17.0 21.0 28.8 32.4
hold victimization as compared with most Totals 100.0 100.0 100.0 100.0
business victimizations, while shoplifting
Source: Offense Analysis from UCR, various
should increase more rapidly than other years.
types of thefts from businesses. We ex- a Excluding miscellaneous robberies. The 1975
pect personal offenses at the hands of distribution omitted due to apparent instability of
post-1970 data.
strangers to manifest greater increases
b Includes romantic triangles, lovers' quarrels and
than such offenses at the hands of arguments.
nonstrangers. Finally, residential burglary e Includes both known and suspected felon types.
rates should increase more in daytime
than nighttime.
The available time series on the com-
THE RELATIONSHIP OF THE HOUSEHOLD
position of offenses confirm these expec-
ACTIVITY RATIO TO FIVE ANNUAL
tations. For example, Table 5 shows that
OFFICIAL INDEX CRIME RATES IN THE
commercial burglaries declined from 60Wo
UNITED STATES, 1947-1974
to 36% of the total, while daytime residen-
tial burglaries increased from 16% to 33%. In this section, we test the hypothesis
Unlike the other crimes against business, that aggregate official crime rate trends in
shoplifting increased its share. Though we the United States vary directly over time
lack trend data on the circumstances of with the dispersion of activities away from
other violent offenses, murder data con- family and household. The limitations of
firm our expectations. Between 1963 and annual time series data do not allow con-
1975, felon-type murders increased from struction of direct measures of changes in
17% to 32% of the total. Compared with a hourly activity patterns, or quantities,
47% increase in the rate of relative killings qualities and movements of exact stocks
in this period, we calculated a 294% in- of household durable goods, but the Cur-
crease in the murder rate at the hands of rent Population Survey does provide re-
known or suspected felon types. lated time series on labor force and
Thus the trends in the composition of household structure. From these data, we
recorded crime rates appear to be highly calculate annually (beginning in 1947) a
consistent with the activity structure household activity ratio by adding the
trends noted earlier. In the next section number of married, husband-present
we apply the routine activity approach in female labor force participants (source:
order to model crime rate trends and so- BLS, 1975: Table 5) to the number of
cial change in the post-World War II non-husband-wife households (source:
United States. USBC, 1947-1976), dividing this sum by
the total number of households in the U.S. homicide mortality rate taken from the
(source: USBC, 1947-1976). This calcula- vital statistics data collected by the
tion provides an estimate of the propor- Bureau of the Census (various years). The
tion of American households in year t ex- latter rate has the advantage of being col-
pected to be most highly exposed to risk lected separately from the standard crime
of personal and property victimization due reporting system and is thought to contain
to the dispersion of their activities away less measurement error (see Bowers and
from family and household and/or their Pierce, 1975). Hence, this analysis of offi-
likelihood of owning extra sets of durables cial index crime rates includes three vio-
subject to high risk of attack. Hence, the lent offenses (homicide, forcible rape, and
household activity ratio should vary di- aggravated assault), one property offense
rectly with official index crime rates. (burglary), and one offense which in-
Our empirical goal in this section is to volves both the removal of property and
test this relationship, with controls for the threat of violence (robbery). The
those variables which other researchers analysis thus includes one offense thought
have linked empirically to crime rate to have relatively low reporting reliability
trends in the United States. Since various (forcible rape), one thought to have rela-
researchers have found such trends to in- tively high reliability (homicide), and
crease with the proportion of the popula- three others having relatively intermediate
tion in teen and young adult years (Fox, levels of reporting quality (Ennis, 1967).
1976; Land and Felson, 1976; Sagi and Since official crime rates in year t are
Wellford, 1968; Wellford, 1973), we in- likely to reflect some accumulation of
clude the population ages 15-24 per criminal opportunity and inclinations over
100,000 resident population in year t as several years, one should not expect these
our first control variable (source: USBC, rates to respond solely to the level of the
various years). Others (e.g., Brenner, independent variables for year t. A useful
1976a; 1976b) have found unemployment model of cumulative social change in cir-
rates to vary directly with official crime cumstances such as this is the difference
rates over time, although this relationship equation, which can be estimated in two
elsewhere has been shown to be empiri- forms (see Goldberg, 1958). One form
cally questionable (see Mansfield et al., takes the first difference (yt - yt-1) as the
1974: 463; Cohen and Felson, 1979). Thus, dependent variable-in this case, the
as our second, control variable, we take change in the official crime rate per
the standard annual unemployment rate 100,000 population between year t- 1 and
(per 100 persons ages 16 and over) as a year t. Alternatively, one can estimate the
measure of the business cycle (source: difference equation in autoregressive form
BLS, 1975). by taking the official crime rate in year t as
Four of the five crime rates that we a function of the exogenous predictors
utilize here (forcible rape, aggravated as- plus the official crime rate in year t - 1 on
sault, robbery and burglary) are taken the right-hand side of the equation. (See
from FBI estimates of offenses per Land, 1978, for a review of these and
100,000 U.S. population (as revised and other methods and for references to re-
reported in OMB, 1973). We exclude lated literature.) Both forms are estimable
larceny-theft due to a major definitional with ordinary least squares methods,
change in 1960 and auto theft due to ex- which we employ for the years 1947
cessive multicollinearity in the analysis.8 through 1974. The N is 28 years for all but
For our homicide indicator we employ the the homicide rate, for which publication
lags reduce our N to 26.
8 The auto theft rate lagged one year correlated
Even if a positive relationship between
quite strongly with the predictor variables. This mul- the household activity ratio and the offi-
ticollinearity impaired our difference equation cial crime rates is observed, with controls
analysis, although we again found consistently posi- for age and unemployment, we are open to
tive coefficients for the household activity ratio. We
the charge that this may be a spurious
were able to remove autocorrelation by logging all
variables and including the unemployment as a con- consequence of autocorrelation of distur-
trol, but do not report these equations. bances, that is, the possibility that residu-
als are systematically related for nearby lation significantly (and also reduced
time points. While spurious relationships variance explained to between 35% and
are a risk one also takes in cross-sectional 77%). These equations also consistently
regression analysis, time-series analysts produce significant positive coefficients
have devised a variety of methods for for the household activity variable. When
monitoring and adjusting for spuriousness unemployment is included in these equa-
due to this autocorrelation, including the tions, its coefficients are all negative and
Durbin and Watson (1951) statistic, Dur- near zero.
bin's h statistic (Durbin, 1970), the The top panel of Table 6 presents re-
Griliches (1967) criterion, as well as Coch- gression estimates of first differences for
rane and Orcutt (1949) corrections. We five official crime rates, with the age
employ (but do not report in detail) these structure and household activity variables
methods to check for the likelihood that in year t as the only predictors. Again, the
the observed relationship is spurious. (See household activity coefficients are consis-
Land, 1978, for a review of such tests and tently positive, with t ratios always sig-
the related literature on their applicability nificant with a one-tailed test. Except for
and robustness; see Theil, 1971, for a the aggravated assault equation, the
methodological review.) household activity variable has a t ratio
and standardized coefficient greater than
that of the age structure variable. The
Findings
standardized coefficients for the house-
Our time-series analysis for the years hold activity variable range from .42 to
1947-1974 consistently revealed positive .72, while the age structure coefficients
and statistically significant relationships are consistently positive. In general, the
between the household activity ratio and household activity variable is a stronger
each official crime rate change predictor of official crime rate trends than
Whichever official crime rate is employed, the age structure.
this finding occurs-whether we take the The equations in the top panel of Table
first difference for each crime rate as 6 generally have lower variance explained
exogenous or estimate the equation in au- but also lower risk of autocorrelation of
toregressive form (with the lagged depen- disturbances than those reported above.
dent variable on the right-hand side of the For all five equations, the Durbin-Watson
equation); whether we include or exclude statistic allows acceptance of the null hy-
the unemployment variable; whether we pothesis that autocorrelation is absent at
take the current scales of variables or the 1% level. A 5% level (which increases
convert them to natural log values; the likelihood of proving the statistic non-
whether we employ the age structure vari- zero) allows us neither to accept nor reject
able as described or alter the ages exam- the null hypothesis that autocorrelation is
ined (e.g., 14-24, 15-19, etc.). In short, absent in the homicide and robbery equa-
the relationship is positive and significant tions.
in each case. Though autocorrelation has not been
Before calculating the difference equa- proven to exist in these five equations, its
tions, we regressed each crime rate in year risk may be sufficient in two to motivate
t on the three independent variables for further efforts at equation estimation (see
year t. This ordinary structural equation bottom panel of Table 6). We estimated
also produced consistent positive and sig- the equations in autoregressive form to
nificant coefficients for the routine ac- see if the risk abates. Since the Durbin-
tivity coefficient, the total variance ex- Watson statistic was not designed for
plained ranges from 84% to 97%. How- evaluating autocorrelation in these equa-
ever, the Durbin-Watson statistics for tions, we calculated Durbin's h, a statistic
these equations indicated high risk of au- specifically designed for equations esti-
tocorrelation, which is hardly surprising mated with a lagged dependent variable
since they ignore lagged effects. Reesti- (Durbin, 1970), and recently found to be
mated equations taking first differences as robust for small samples (Maddala and
endogenous reduced the risk of autocorre- Rao, 1973). This statistic allows ac-
6, o oo ffi ? , -tb nc 3o o $Q _ -
0 - O 't N VC
en ON "m N C l
_- D .0
.t en r- cq 0 f 11 en W) X O
_ - -< ClCO _ u u ,
CZ C
-o
C1 r- . .
a)
ca ON VI V- en 0? 0 ro CZ CZoWt =,QX fi>

c u W r- o 4 0 0 'It W) W) 00 00 0 QQ
*1-I o1N1 N- N ><11<<

-o D
n 11e' 1~~O~ O 1 N"1 NoV~V ~ X I I S S

UQ 00 N ClC r r
0 -C
Q C) CL o so t ? o ~~~ db u vo oo ??W W)V u 0

a) ~~ 'rC ~0 (A NN', O
C~~~~~~~~~~ZC
? I 1 s I C l
s : = ? e~0 x u W n , =n88=
ceptance of the null hypothesis (at both percent of foreign stock declined 0.1%,
1% and 5% levels) that autocorrelation is etc. (see USBC, 1975: 654, 19, 39).
absent for all five equations. Application On the other hand, the convergence in
of the Griliches (1967) criterion further time and space of three elements (moti-
allows acceptance of each equation as vated offenders, suitable targets, and the
manifesting distributing lags rather than absence of capable guardians) appears
serial correlation. We also employed the useful for understanding crime rate
Cochrane-Orcutt (1949) iterative proce- trends. The lack of any of these elements
dure to calculate a correction estimate for is sufficient to prevent the occurrence of a
any autocorrelation present. The resulting successful direct-contact predatory crime.
correction for the household activity coef- The convergence in time and space of
ficient proves minimal in all five cases. suitable targets and the absence of capable
Finally, we calculated each of the above guardians can lead to large increases in
equations for natural log values of the crime rates without any increase or
relevant variables, finding again that the change in the structural conditions that
household activity coefficient was consis- motivate individuals to engage in crime.
tently positive and statistically significant Presumably, had the social indicators of
and the risk of autocorrelation reduced the variables hypothesized to be the
still further. causes of crime in conventional theories
The positive and significant relationship changed in the direction of favoring in-
between the household activity variable creased crime in the post-World War II
and the official crime rates is robust and United States, the increases in crime rates
appears to hold for both macro- and mi- likely would have been even more stagger-
crolevel data; it explains five crime rate ing than those which were observed. In
trends, as well as the changing composi- any event, it is our belief that
tion of official crime rates reported in criminologists have underemphasized the
Table 5. These results suggest that routine importance of the convergence of suitable
activities may indeed provide the oppor- targets and the absence of capable guard-
tunity for many illegal activities to occur. ians in explaining recent increases in the
crime rate. Furthermore, the effects of the
convergence in time and space of these
elements may be multiplicative rather
DISCUSSION
than additive. That is, their convergence
In our judgment many conventional by a fixed percentage may produce in-
theories of crime (the adequacy of which creases in crime rates far greater than that
usually is evaluated by cross-sectional fixed percentage, demonstrating how
data, or no data at all) have difficulty ac- some relatively modest social trends can
counting for the annual changes in crime contribute to some relatively large
rate trends in the post-World War II changes in crime rate trends. The fact that
United States. These theories may prove logged variables improved our equations
useful in explaining crime trends during (moving Durbin-Watson values closer to
other periods, within specific com- "ideal" levels) lends support to the argu-
munities, or in particular subgroups of the ment that such an interaction occurs.
population. Longitudinal aggregate data Those few investigations of cross-
for the United States, however, indicate sectional data which include household
that the trends for many of the presumed indicators produce results similar to ours.
causal variables in these theoretical struc- For example, Roncek (1975) and Choldin
tures are in a direction opposite to those and Roncek (1976) report on block-level
hypothesized to be the causes of crime. data for San Diego, Cleveland and Peoria
For example, during the decade 1960- and indicate that the proportion of a
1970, the percent of the population below block's households which are primary in-
the low-income level declined 44% and the dividual households consistently offers
unemployment rate declined 186o. Cen- the best or nearly the best predictor of a
tral city population as a share of the whole block's crime rate. This relationship per-
population declined slightly, while the sisted after they controlled for numerous
social variables, including race, density, victimization. Indeed, the opportunity for
age and poverty. Thus the association be- predatory crime appears to be enmeshed
tween household structure and risk of in the opportunity structure for legitimate
criminal victimization has been observed activities to such an extent that it might be
in individual-level and block-level cross- very difficult to root out substantial
sectional data, as well as aggregate na- amounts of crime without modifying much
tional time-series data. of our way of life. Rather than assuming
Without denying the importance of fac- that predatory crime is simply an indicator
tors motivating offenders to engage in of social breakdown, one might take it as a
crime, we have focused specific attention byproduct of freedom and prosperity as
upon violations themselves and the pre- they manifest themselves in the routine
requisites for their occurrence. However, activities of everyday life.
the routine activity approach might in the
future be applied to the analysis of offend-
ers and their inclinations as well. For
example, the structure of primary group REFERENCES
activity may affect the likelihood that cul-
Amir, Menachem
tural transmission or social control of 1971 Patterns of Forcible Rape. Chicago: Uni-
criminal inclinations will occur, while the versity of Chicago Press.
structure of the community may affect the Boggs, Sarah
1%5 "Urban crime patterns." American
tempo of criminogenic peer group ac-
Sociological Review 30:899-905.
tivity. We also may expect that circum- Bonger, W. A.
stances favorable for carrying out viola- 1916 Criminality and Economic Conditions. Bos-
tions contribute to criminal inclinations in ton: Little, Brown.
the long run by rewarding these inclina- Bowers, W. J. and Glen L. Pierce
1975 "The illusion of deterrence of Isaac
tions.
Ehrlich's research on capital punishment."
We further suggest that the routine ac- Yale Law Journal 85:187-208.
tivity framework may prove useful in ex- Brenner, Harvey
plaining why the criminal justice system, 1976a Estimating the Social Costs of National
Economic Policy: Implications for Mental
the community and the family have ap-
and Physical Health and Criminal Aggres-
peared so ineffective in exerting social sion. Paper no. 5, Joint Economic Commit-
control since 1960. Substantial increases tee, Congress of the United States. Wash-
in the opportunity to carry out predatory ington, D.C.: U.S. Government Printing
Office.
violations may have undermined society's
1976b Effects of the National Economy on Crimi-
mechanisms for social control. For exam-
nal Aggression II. Final Report to National
ple, it may be difficult for institutions Institute of Mental Health. Contract #282-
seeking to increase the certainty, celerity 76-0355FS.
and severity of punishment to compete Bureau of Labor Statistics (BLS)
1975 Handbook of Labor Statistics 1975-
with structural changes resulting in vast
Reference Edition. Washington, D.C.: U.S.
increases in the certainty, celerity and Government Printing Office.
value of rewards to be gained from illegal Cameron, Mary Owen
predatory acts. 1964 The Booster and the Snitch. New York:
It is ironic that the very factors which Free Press.
Chambliss, William J.
increase the opportunity to enjoy the
1972 Boxman: A Professional Thief's Journey.
benefits of life also may increase the New York: Harper and Row.
opportunity for predatory violations. For Choldin, Harvey M. and Dennis W. Roncek
example, automobiles provide freedom of 1976 "Density, population potential and path-
ology: a block-level analysis." Public Data
movement to offenders as well as average
Use 4: 19-30.
citizens and offer vulnerable targets for Cloward, Richard and Lloyd Ohlin
theft. College enrollment, female labor 1960 Delinquency and Opportunity. New York:
force participation, urbanization, subur- Free Press.
banization, vacations and new electronic Cochrane, D., and G. H. Orcutt
1949 "Application of least squares regression to
durables provide various opportunities to
relationships containing autocorrelated
escape the confines of the household error terms." Journal of the American
while they increase the risk of predatory Statistical Association 44:32-61.
Cohen, Lawrence E. and Marcus Felson Gould, Leroy

1979 "On estimating the social costs of national 1969 "The changing structure of property crime
economic policy: a critical examination of in an affluent society." Social Forces
the Brenner study." Social Indicators Re- 48:50-9.
search. In press. Griliches, Z.
Colquhoun, Patrick 1967 "Distributed lags: a survey." Economet-
1800 Treatise on the Police of the Metropolis. rica 35:16-49.
London: Baldwin. Guerry, A. M.
Consumer Reports Buying Guide 1833 "Essai sur la statistique morale de la
1959 Consumer Reports (December). Mt. Ver- France." Westminister Review 18:357.
non: Consumers Union. Hawley, Amos
1969 Consumer Reports (December). Mt. Ver- 1950 Human Ecology: A Theory of Community
non: Consumers Union. Structure. New York: Ronald.
1975 Consumer Reports (December). Mt. Ver- Henry, A. F., and J. F. Short
non: Consumers Union. 1954 Suicide and Homicide. New York: Free
Council of Economic Advisors (CEA) Press.
1976 The Economic Repprt of the President. Hindelang, Michael J.
Washington, D.C.: U.S. Government Print- 1976 Criminal Victimization in Eight American
ing Office. Cities: A Descriptive Analysis of Common
Durbin, J. Theft and Assault. Cambridge: Ballinger.
1970 Testing for serial correlation when least Hindelang, Michael J., Christopher S. Dunn, Paul
squares regressors are lagged dependent Sutton and Alison L. Aumick
variables." Econometrica 38:410-21. 1976 Sourcebook of Criminal Justice
Durbin, J., and G. S. Watson Statistics-1975. U.S. Dept. of Justice,
1951 "Testing for serial correlation in least Law Enforcement Assistance Administra-
squares regression, II.- Biometrika tion. Washington, D.C.: U.S. Government
38:159-78. Printing Office.
Durkheim, Emile 1977 Sourcebook of Criminal Justice
1951 Suicide: A Study in Sociology. New York: Statistics-1976. U.S. Dept. of Justice,
Free Press. Law Enforcement Assistance Administra-
1966 The Division of Labor in Society. New tion. Washington, D.C.: U.S. Government
York: Free Press. Printing Office.
Electrical Merchandising Week
Interstate Commerce Commission (ICC)
1964 Statistical and Marketing Report (January). 1974 Annual Report: Freight Commodity Statis-
tics of Class I Motor Carriers of Property
New York: Billboard Publications.
Operative in Intercity Service. Washington,
Ennis, Philip H.
D.C.: U.S. Government Printing Office.
1967 "Criminal victimization in the U.S.: a re-
Jackson, Bruce
port of a national survey, field surveys II."
1969 A Thief's Primer. New York: Macmillan.
The President's Commission on Law En-
Jeffery, C. R.
forcement and the Administration of Jus-
1971 Crime Prevention Through Environmental
tice. Washington, D.C.: U.S. Government
Design. Beverly Hills: Sage.
Printing Office.
Klockars, Carl B.
Federal Bureau of Investigation (FBI) 1974 The Professional Fence. New York: Free
1975 Crime in the U.S.: Uniform Crime Report. Press.
Washington, D.C.: U.S. Government Kobrin, Frances E.
Printing Office. 1976 "The primary individual and the family:
1976 Crime in the U.S.: Uniform Crime Report. changes in living arrangements in the U.S.
Washington, D.C.: U.S. Government since 1940." Journal of Marriage and the
Printing Office. Family 38:233-9.
Ferdinand, Theodore N. Land, Kenneth C.
1970 "Demographic shifts and criminality." 1978 "Modelling macro social change." Paper
British Journal of Criminology 10:169-75. presented at annual meeting of the Ameri-
can Sociological Association, San Fran-
Fleisher, Belton M.
cisco.
1966 The Economics of Delinquency. Chicago:
Land, Kenneth C. and Marcus Felson
Quadrangle.
1976 "A general framework for building dynamic
Fox, James A.
macro social indicator models: including an
1976 An Econometric Analysis of Crime Data.
analysis of changes in crime rates and
Ph.D. dissertation, Department of Sociol-
police expenditures." American Journal of
ogy, University of Pennsylvania. Ann Ar-
Sociology 82:565-604.
bor: University Microfilms.
Letkemann, Peter
Glaser, Daniel 1973 Crime As Work. Englewood Cliffs:
1971 Social Deviance. Chicago: Markham. Prentice-Hall.
Goldberg, Samuel Lystad, Mary
1958 Introduction to Difference Equations. New 1974 An Annotated Bibliography: Violence at
York: Wiley. Home. DHEW Publication No. (ADM 75-
136). Washington, D.C.: U.S. Government criminal statistics." Journal of Criminal

Printing Office. Law, Criminology and Police Science
Maddala, G. S., and A. S. Rao 59:29-36.
1973 'Tests for serial correlation in regression Scarr, Harry A.
models with lagged dependent variables and 1972 Patterns of Burglary. U.S. Dept. of Justice,
serially correlated errors." Econometrica Law Enforcement Assistance Administra-
41:761-74. tion. Washington, D.C.: U.S. Government
Mansfield, Roger, Leroy Gould, and J. Zvi Namen- Printing Office.
wirth Sears Catalogue
1974 'A socioeconomic model for the prediction 1960 Chicago: Sears.
of societal rates of property theft." Social 1970 Chicago: Sears.
Forces 52:462-72. Shaw, Clifford R., Henry D. McKay, Frederick Zor-
Martin, John Bower baugh and Leonard S. Cottrell
1952 My Life in Crime. New York: Harper. 1929 Delinquency Areas. Chicago: University of
Maurer, David W. Chicago Press.
1964 Whiz Mob. New Haven: College and Uni- Short, James F., and Fred Strodtbeck
versity Press. 1965 Group Process and Gang Delinquency.
Merchandising Week Chicago: University of Chicago Press.
1973 Statistical and Marketing Report (Febru- Skogan, Wesley G.
ary). New York: Billboard Publications. 1976 "The victims of crime: some material find-
1976 Statistical and Marketing Report (March). ings." Pp. 131-48 in Anthony L. Guenther
New York: Billboard Publications. (ed.), Criminal Behavior in Social Systems.
National Commission on the Causes and Prevention Chicago: Rand McNally.
of Violence Sutherland, Edwin H.
1969 Crimes of Violence. Vol. 13. Washington, 1937 The Professional Thief. Chicago: Univer-
D.C.: U.S. Government Printing Office. sity of Chicago Press.
Nettler, Gwynn Szalai, Alexander (ed.)
1974 Explaining Crime. New York: McGraw- 1972 The Use of Time: Daily Activities of Urban
Hill. and Suburban Populations in Twelve Coun-
Newman, Oscar tries. The Hague: Mouton.
1973 Defensible Space: Crime Prevention Theil, Henri
Through Urban Design. New York: Mac- 1971 Principles of Econometrics. New York:
millan. Wiley.
Office of Management and the Budget (OMB) Tobias, J. J.
1973 Social Indicators 1973. Washington, D.C.: 1%7 Crime and Industrial Society in the
U.S. Government Printing Office. Nineteenth Century. New York: Schocken
Pope, Carl E. Books.
1977a Crime-Specific Analysis: The Char- U.S. Bureau of the Census (USBC)
acteristics of Burglary Incidents. U.S. 1973a Census of Transportation, 1972. U.S.
Dept. of Justice, Law Enforcement Assist- Summary. Washington, D.C.: U.S. Gov-
ance Administration. Analytic Report 10. ernment Printing Office.
Washington, D.C.: U.S. Government Print- 1973b Who's Home When. Working Paper 37.
ing Office. Washington, D.C.: U.S. Government Print-
1977b Crime-Specific Analysis: An Empirical ing Office.
Examination of Burglary Offense and Of- 1975- Statistical Abstract of the U.S. Washing-
fender Characteristics. U.S. Dept. of Jus- 1976 ton, D.C.: U.S. Government Printing
tice, Law Enforcement Assistance Ad- Office.
ministration. Analytical Report 12. Wash- 1947- Current Population Studies. P-25 Ser.
ington, D.C.: U.S. Government Printing 1976 Washington, D.C.: U.S. Government Print-
Office. ing Office.
Quetelet, Adolphe U.S. Department of Justice (USDJ)
1842 A Treatise on Man. Edinburgh: Chambers. 1974a Preliminary Report of the Impact Cities,
Reiss, Albert J. Crime Survey Results. Washington, D.C.:
1976 "Settling the frontiers of a pioneer in Law Enforcement Assistance Administra-
American criminology: Henry McKay." tion (NCJISS).
Pp. 64-88 in James F. Short, Jr. (ed.), De- 1974b Crime in the Nation's Five Largest Cities:
linquency, Crime, and Society. Chicago: Advance Report. Washington, D.C.: Law
University of Chicago Press. Enforcement Assistance Administration
Reppetto, Thomas J. (NCJISS).
1974 Residential Crime. Cambridge: Ballinger. 1974c Crimes and Victims: A Report on the
Roncek, Dennis Dayton-San Jose Pilot Survey of Victimiza-
1975 Crime Rates and Residential Densities in tion. Washington, D.C.: Law Enforcement
Two Large Cities. Ph.D. dissertation, De- Assistance Administration.
partment of Sociology, University of Il- 1976 Criminal Victimizations in the U.S., 1973.
linois, Urbana. Washington, D.C.: Law Enforcement
Sagi, Phillip C. and Charles E. Wellford Assistance Administration (NCJISS).
1968 "Age composition and patterns of change in 1977 Criminal Victimizations in the U.S.: A
Comparison of 1974 and 1975 Findings. quency." Pp. 138-56 in President's Com-
Washington, D.C.: Law Enforcement mission on Law Enforcement and the Ad-
Assistance Administration (NCJISS). ministration of Justice Task Force Report:
Washnis, George J. Crime and Its Impact-An Assessment.
1976 Citizen Involvement in Crime Prevention. Appendix A. Washington, D.C.: U.S. Gov-
Lexington: Heath. ernment Printing Office.
Wellford, Charles F. Williamson, Henry
1973 "Age composition and the increase in re- 1968 Hustler! New York: Doubleday.
corded crime." Criminology 11:61-70. Wolfgang, Marvin E.
Wilks, Judith A. 1958 Patterns of Criminal Homicide. Philadel-
1967 "Ecological correlates of crime and delin- phia: University of Pennsylvania Press.
INTERGENERATIONAL OCCUPATIONAL MOBILITY AND

FERTILITY: A REASSESSMENT*
FRANK D. BEAN AND GRAY SWICEGOOD

University of Texas, Austin
American Sociological Review 1979, Vol. 44 (August):608-619
This paper examines the relationship between intergenerational occupational mobility and
family size; it focuses on the partition of cumulative fertility into two components-intended
and unintended births-that may respond differently to social and economic sources of
influence. Four theoretical perspectives are drawn upon to set forth alternative predictions of
the relationship between unintended and intended births for upwardly and downwardly mobile
couples. The results of regression analyses of 1970 National Fertility Survey data lend support
to Easterlin's relative economic status hypothesis, which advances the expectation of a positive
relationship between direction of mobility and fertility, although this pattern is observed only
for unintended births. The implications of this finding for certain theories of fertility, as well as
for the inconclusive nature of previous research into the mobility-fertility relationship, are
discussed.
Social mobility provides one of the Westoff, 1953; Westoff et al., 1961; Wes-
major points of articulation between social toff et al., 1963; Boyd, 1973). The cumula-
stratification and demography. Not only is tive research evidence, however, has been
social mobility an integral feature of all inconclusive, if not totally negative.
systems of social stratification (e.g., Lo- Scarcely any substantial basis has been
preato et al., 1976; Tyree and Hodge, 1978), found for the conclusion that social mobil-
it also has noticeable demographic corre- ity accounts for additional variation in fer-
lates, both for societies and individuals tility above and beyond that which is
(Blau and Duncan, 1967:361; Featherman associated with measures of origin and
and Hauser, 1978). One of the most impor- destination position themselves (e.g.,
tant would seem to be fertility, as evi- Duncan, 1966; Boyd, 1973).
denced by numerous empirical studies For the most part, explanations of these
conducted during the past 25 years (e.g., negative results, as distinct from theoreti-
Berent, 1952; Goldberg, 1959; Tien, 1961; cal interpretations that are consonant with
the idea that knowledge of origin and des-
*Direct all communications to: Frank D. Bean; tination positions suffices to explain the
Department of Sociology; University of Texas; Au-
fertility of the mobile (e.g., Blau, 1956:29),
stin, TX 78712.
The data upon which this paper is based were
have taken either of two tacks. On the one
collected pursuant to contract #PH-43-65 1048 with hand, they have suggested that the
the National Institute of Health, Public Health Ser- methodological basis for assessing rela-
vice, Department of Health, Education and Welfare. tionships between mobility and fertility is
We would like to express our appreciation to Caro-
biased against the discovery of such rela-
lyn Boyd, Norval Glenn, Joseph Lopreato, Alberto
Palloni, and two anonymous reviewers for helpful tionships (e.g., Hope, 1971; 1975; Lo-
comments on an earlier draft of the paper. preato et al., 1976). On the other hand, they

Cohen and Felson 1979 Routine Activities

Uploaded by

Copyright:

Available Formats

Cohen and Felson 1979 Routine Activities

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cohen and Felson 1979 Routine Activities

Uploaded by

Copyright:

Available Formats

Social Change and Crime Rate Trends: A Routine Activity Approach

Author(s): Lawrence E. Cohen and Marcus Felson

American Sociological Association is collaborating with JSTOR to digitize, preserve and

LAWRENCE E. COHEN AND MARCUS FELSON

American Sociological Review 1979, Vol. 44 (August):588-608,

INTRODUCTION family income during the same period. Also

movement of valuable goods through its To do so requires a conceptual framework

4 The concept of the opportunity for crime con-

direct-contact predatory crime rates over National estimates from large-scale

A.* Rape Robbery Assault with Contact Total

* Calculated from Handelang et al., 1977: Tables 3.16

8V worth of nondurables andcriminal

A. AGE 12-15 147 1,267 3,848 311 16,355

B. MAJOR (Male 16+)

C. MARITAL STATUS (Male 12+)

trends as they relate to trends in crime Table 4. Proportion of Households Unattended by

8:00- 8:59 a.m. 29% 43 +48.9o

Trends in Human Activity Patterns

C. LARCENIES 1960 1965 1970 1975

ca ON VI V- en 0? 0 ro CZ CZoWt =,QX fi>

*1-I o1N1 N- N ><11<<

n 11e' 1~~O~ O 1 N"1 NoV~V ~ X I I S S

Q C) CL o so t ? o ~~~ db u vo oo ??W W)V u 0

Cohen, Lawrence E. and Marcus Felson Gould, Leroy

136). Washington, D.C.: U.S. Government criminal statistics." Journal of Criminal

INTERGENERATIONAL OCCUPATIONAL MOBILITY AND

FRANK D. BEAN AND GRAY SWICEGOOD

American Sociological Review 1979, Vol. 44 (August):608-619

You might also like