Summer crime spike seen, according to federal report

8/11/2019 Summer crime spike seen, according to federal report

1/22

JUNE 2014 NCJ 245959

U.S. Department of Justice

Ofce of Justice Programs

Bureau of Justice Statistics

BJS

S R

Janet L. Lauritsen, Ph.D., BJS Visiting Fellow, and Nicole White, Ph.D., University of Missouri St. Louis

Seasonal Patterns in Criminal

Victimization Trends

Introduction

Seasonal patterns exist when similar fluctuations in thevictimization rates tend to recur each year during the sameseason. Seasonal patterns are a long-standing topic in bothpopular and scholarly literatures on crime and show how

environmental actors, such as temperature changes anddaylight hours, might be associated with crime throughoutthe year. Seasonal patterns also are examined to learnwhether routine activity patterns, such as those associatedwith the beginning and end o the school year, are associatedwith variations in crime throughout the year.

Tis report examines seasonal patterns in violent andhousehold property victimization in the United States rom1993 to 2010. It describes seasonal patterns or householdproperty victimization (burglary, motor vehicle thef, andhousehold larceny) and types o violence (rape and sexualassault, robbery, aggravated assault, and simple assault).

rends in other orms o violence, including intimate partnerviolence, victimizations involving weapons, and thoseresulting in injury, also are examined to determine the extentto which seasonal fluctuations exist.

For household property crimes, the victimization dataindicate that there are seasonal patterns in householdburglary and larceny, with higher rates in the summer andlower rates during other seasons o the year. Motor vehiclethefs do not exhibit the same seasonal patterns.

For violent victimizations, seasonal patterns also varydepending on the type o crime. Rape and sexual assaultrates and aggravated assault rates are higher in the summerthan in most other seasons. In comparison, simple assaultrates are higher in the all than in other seasons, and robbery

victimization rates show no seasonal patterns. Intimatepartner violence exhibits regular seasonal fluctuations withrates highest during the summer and lowest during thewinter. Rates o violence involving weapons and violenceresulting in serious injury are higher in the summer than inthe winter and spring seasons; however, the all rates are notsignificantly different rom the summer rates.

HIGHLIGHTS

When seasonal variations in household property

victimization were found, the differences between the

highest and lowest seasonal rates were less than 11%.

Seasonal patterns existed in household larceny and burglary

victimization rates. Rates of these household crimes tendedto be higher in the summer than during other seasons of

the year.

Though rates of motor vehicle theft tended to be lower

in the spring than in the summer, there were few regular

differences between summer, fall, and winter rates.

When seasonal variations were found for violent

victimization, the differences between the rates of the

highest and lowest season were less than 12%.

Aggravated assault rates were higher during the summer

than during the winter, spring, and fall. In comparison,

simple assault rates were higher during the fall than during

other seasons of the year.

Robbery rates did not exhibit seasonal variations. Rape and sexual assault victimization rates tended to be

higher during the summer than during the fall and winter.

Rates of intimate partner violence were higher during the

summer than during the winter, spring, and fall.


2/22

2SEASONAL PATTERNS IN CRIMINA L VICTIMIZATION TRENDS | JUNE 2014

Prior BJS Research on Seasonality

BJS has previously examined seasonal patterns in

selected types of victimization using monthly data from

1973 to 1977 from the National Crime Survey (NCS), the

predecessor of the National Crime Victimization Survey

(NCVS).1Earlier research examined some of the same types

of victimizations assessed in this report.

The previous report found notable seasonal fluctuations

in household larceny and burglary, and rates were

highest during the summer months. Seasonal differences

were also found for motor vehicle theft and for assault

victimization. For these crimes, the differences between

the summer and winter month rates were relatively

smaller than those found for household burglary and

larceny. The earlier findings for these types of victimization

are similar to the patterns shown in this report; however,

the magnitude of the seasonal differences were somewhat

larger from 1973 to 1977 than from 1993 to 2010. During

the earlier period, burglaries were about 26% less frequent

in the winter than in the summer, while in the more recent

period, the winter versus summer difference in burglary

was about 11%. No significant seasonal fluctuations were

found for robbery in either study.

1Crime and Seasonality: National Crime Survey Report, NCJ 64818,BJS web, May 1980.

Te data in this report were developed rom the Bureauo Justice Statistics (BJS) National Crime VictimizationSurvey (NCVS), which annually collects inormation onnonatal victimizations against persons age 12 or olderrom a nationally representative sample o U.S. households.Te data included crimes reported and not reported to thepolice rom 1993 to 2010. Asking the victim in which monththe incident occurred makes it possible to assess how therisk or victimization varies throughout the year. For eacho the crime types, differences in the rates are comparedacross the our seasons o the year. Winter rates are basedon victimizations that occur in December, January, andFebruary; spring rates occur in March, April, and May;summer rates occur in June, July, and August; and all ratesoccur in September, October, and November. Seasonalpatterns in the 1993 to 2010 trends exist when statisticalanalyses reveal periodic fluctuations in the rates, which tendto recur each year during the same season SeeMethodologyor a description o the statistical analysis used in this report.

BJS Visiting Fellows

The BJS Visiting Fellows program is intended to facilitate

collaboration between academic scholars and governmentresearchers in survey methodology, statistics, economics,

and social sciences. Visiting Fellows have the unique

opportunity to address substantive, methodological, and

analytic issues relevant to BJS programs, and to further

knowledge and understanding of criminal justice systems

operation. Fellows conduct research at BJS or at their

home site, use BJS data and facilities, and interact with BJS

staff. Pending available funding, the program is normally

conducted each year through a competitive solicitation.

Applicants are limited to senior-level social science

researchers or statisticians in the fields of statistics, surveymethodology, mathematics, criminology, demography,

economics, behavioral science, and other related fields.

They must have an established research record in their

field, have considerable expertise in their area of proposed

research, and be willing to commit a substantial portion of

their time (typically 6 to 18 months) to undertake analyses

of existing BJS data. In addition, they must produce a

report that both summarizes their analyses and meets

BJS publication and data quality standards. For further

information, see http://www.bjs.gov/content/fellows.cfm.
http://www.bjs.gov/content/fellows.cfmhttp://www.bjs.gov/content/fellows.cfm


3/22


TABLE 1

Average percent difference in seasonal rate of householdvictimization, by type of crime, 19932010

Total property

crime

Household

burglary

Motor

vehicle theft

Household

larcenyWinter -6.9%* -10.5%* -2.2% -6.4%*

Spring -8.0* -8.8* -5.6* -8.0*

Summer ~ ~ ~ ~

Fall -3.1* -5.9* 0 -2.7*Note: See appendix table 1 for regression coefficients and standard errors.

*Difference is statistically significant (p < .05, two-tail test).

~Highest rate season used for percentage comparisons.

Source: Bureau of Justice Statistics, National Crime Victimization Survey,19932010.

0

20

40

60

80

100

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93

Per 1,000 households

Source: Bureau of Justice Statistics, National Crime Victimization Survey, 19932010.

FIGURE 1

Seasonal rates of household property victimization, 19932010

Seasonal patterns in householdproperty victimization

Total household property victimization

Rates o total household property victimization includeburglary, motor vehicle thef, and household larceny.Differences in the total household property victimization

rates across winter, spring, summer, and all exhibitedrecurrent seasonal patterns (figure 1). In all o the yearsexcept 1995, rates o household property victimizationappeared highest in summer and lowest in either winter orspring. Statistical analysis o the seasonal rates indicates thatthey were higher in the summer than in the other seasonso the year (table 1). Compared to summer rates, householdproperty victimization rates were an average o about

7% lower in winter, 8% lower in spring, and 3% lower in all.Tereore, the difference between rates in the summer andother seasons was relatively small.


4/22


0

5

10

15

20

S

ummer'10

Winter'10

S

ummer'09

Winter'09

S

ummer'08

Winter'08

S

ummer'07

Winter'07

S

ummer'06

Winter'06

S

ummer'05

Winter'05

S

ummer'04

Winter'04

S

ummer'03

Winter'03

S

ummer'02

Winter'02

S

ummer'01

Winter'01

S

ummer'00

Winter'00

S

ummer'99

Winter'99

S

ummer'98

Winter'98

S

ummer'97

Winter'97

S

ummer'96

Winter'96

S

ummer'95

Winter'95

S

ummer'94

Winter'94

S

ummer'93

Winter'93



FIGURE 2

Seasonal rates of burglary, 19932010

Burglary

Household burglary exhibited seasonal patterns, with rateso victimization typically highest in summer and lowest inwinter or spring (figure 2). In all but 3 years (1999, 2005,and 2008), rates o burglary appeared2highest in summer.

In some years, spring rates o burglary were lower thanwinter rates. On average, household burglary rates wereapproximately 11% lower in winter, 9% lower in spring, andabout 6% lower in all than in summer. Among householdproperty crimes, burglary showed the largest averagedifference (11%) between the peak (summer) and trough(winter) seasonal rates.

2Te term "appeared" reers to within year seasonal variations in pointestimates that are not statistically different.


5/22


0

2

4

6

8

10

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93



FIGURE 3

Seasonal rates of motor vehicle theft, 19932010

Motor vehicle theft

Compared to other household property crimes, rates omotor vehicle thef showed limited amounts o seasonal

variation rom 1993 to 2010 (figure 3). In approximatelyhal o the years, motor vehicle thef appeared highest in thesummer, while in other years it appeared higher during theall. Te lowest rates o motor vehicle thef seemed to be in

the spring, though this was not always the case. Statisticalanalysis confirmed the lack o seasonality in motor vehiclethef and revealed that the differences between summer, all,and winter rates were not statistically significant. However,spring rates were significantly lower than summer rateso motor vehicle thef, with spring rates on average about6% lower than summer rates.


6/22


0

20

40

60

80

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93



FIGURE 4

Seasonal rates of household larceny, 19932010

Household larceny

Rates o household larceny showed some consistent seasonalfluctuations throughout the 1993 to 2010 period (figure 4).Household larceny typically appeared highest in thesummer, though in some years (such as 1994 to 1996), allrates appeared equal to or slightly higher than the summer

rates. Like burglary, household larceny rates also appeared tobe lowest in the spring or winter. Statistical analyses revealedseasonality in household larceny. Compared to summerrates, household larceny rates were about 6% lower in thewinter, 8% lower in spring, and about 3% lower in all.


7/22


0

5

10

15

20

S

ummer'10

Winter'10

S

ummer'09

Winter'09

S

ummer'08

Winter'08

S

ummer'07

Winter'07

S

ummer'06

Winter'06

S

ummer'05

Winter'05

S

ummer'04

Winter'04

S

ummer'03

Winter'03

S

ummer'02

Winter'02

S

ummer'01

Winter'01

S

ummer'00

Winter'00

S

ummer'99

Winter'99

S

ummer'98

Winter'98

S

ummer'97

Winter'97

S

ummer'96

Winter'96

S

ummer'95

Winter'95

S

ummer'94

Winter'94

S

ummer'93

Winter'93

Per 1,000 persons age 12 or older


FIGURE 5

Seasonal rates of total violent victimization, 19932010

TABLE 2Average percent difference in seasonal rate of violentvictimization, by type of crime, 19932010

Totalviolence

Seriousviolence

Rape/sexualassault Robbery

Aggravatedassault

Simpleassault

Winter -4.4%* -5.6%* -9.0%* -2.5% -6.9%* -5.6%*

Spring -3.4* -4.7* -6.0 -1.9 -6.0* -4.7*

Summer -1.8 ~ ~ -0.6 ~ -5.0*

Fall ~ -3.2* -10.4* ~ -3.6* ~

Note: See appendix table 2 for regression coefficients and standard errors.




Seasonal patterns in violentvictimization

Total violent victimization

Summary indicators and specific types o violence wereexamined to assess the seasonal patterns or violent

victimization. otal violent victimization includes rape

and sexual assault, robbery, aggravated assault, and simpleassault. Differences in the total violent victimization ratesacross the winter, spring, summer, and all exhibited somerecurrent seasonal patterns rom 1993 to 2010 (figure 5).However, these patterns were different rom those observedor household property victimization.

In most years, rates o violence appeared highest in theall and lowest in the winter, though this was not alwaysthe case. On average, statistical analysis indicated that therates were higher in the all than in the winter and spring

and that summer rates were statistically similar to rates oviolence during the all (table 2). Tough the differencesbetween the all rates o violence and the winter and springrates were statistically significant, the seasonal variationswere small. On average, rates o violence during the all wereabout 4% higher than winter rates and about 3% higher thanspring rates.


8/22


0

2

4

6

8

10

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93



FIGURE 6

Seasonal rates of serious violent victimization, 19932010

Serious violent victimization

Serious violent victimization rates, which include rapeand sexual assault, robbery, and aggravated assault, variedsignificantly across the seasons in ways that differed romthose or total violent victimization (figure 6). Statisticalanalysis o seasonal rates showed that serious violence wassignificantly higher during the summer than during thewinter, spring, and all seasons. Winter rates o serious

violence were approximately 6% lower on average thansummer rates o serious violence, and rates or spring(5% lower) and all (3% lower) were also lower than summerrates. Because serious violent victimization excluded simpleassault, the differences between the seasonal patterns ortotal and serious violent victimization indicate a uniqueseasonal pattern or simple assault victimization comparedto other types o violence.


9/22


0.0

0.2

0.4

0.6

0.8

1.0

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93



FIGURE 7

Seasonal rates of rape and sexual assault, 19932010

Rape and sexual assault

When trends in specific types o violent crime wereexamined separately, differences in seasonal patternsemerged. Rape and sexual assault rates exhibited someregular fluctuation across some o the seasons; however,the patterns did not indicate a single peak season, as the

differences between the summer and spring rates were notstatistically significant (figure 7). Winter and all rates orape and sexual assault tended to be lower than summerrates, but they were not significantly lower than spring rates.From 1993 to 2010, winter rates o rape and sexual assaultwere approximately 9% lower on average than summer rates,and all rates were about 10% lower than summer rates.


10/22


0.0

0.4

0.8

1.2

1.6

2.0

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93



FIGURE 8

Seasonal rates of robbery, 19932010

Robbery

Unlike the other types o violent crime, robbery did notexhibit seasonal fluctuations rom year to year (figure 8).Robbery victimization rates were statistically similar during

winter, spring, summer, and all. Tis similarity in seasonalrates indicates that the periodic fluctuations that wereobserved in the rates did not usually recur each year duringthe same season.


11/22


0

1

2

3

4

5

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93



FIGURE 9

Seasonal aggravated assault rates, 19932010

Aggravated assault

Seasonal patterns in aggravated assault were statisticallysignificant and most comparable to those shown or serious

violent victimization (figure 9). Tis was because aggravatedassault rates were higher than the rates or the other serious

violence crime types (rape and sexual assault and robbery).In over hal o the years, aggravated assault appeared to peak

in the summer. Likewise, the seasonality analysis indicatedthat aggravated assault rates were significantly higher duringthe summer than during the winter, spring, and all seasons.More specifically, winter rates o aggravated assault wereabout 7% lower on average than summer rates, and rates orspring (6% lower) and all (4% lower) were also lower thansummer rates.


12/22


0

2

4

6

8

10

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93



FIGURE 10

Seasonal rates of simple assault, 19932010

Simple assault

As suggested by the differences in the seasonal patternsor total violence and serious violent victimization, simpleassault rates exhibited a unique pattern o seasonality(figure 10). Unlike any other type o violent victimization,simple assault rates were highest in the all. On average,

winter rates o simple assault were nearly 6% lower thanall rates, while spring and summer rates were about 5%lower than all rates. Tis unique seasonal pattern or simpleassault was due to the experiences o youth ages 12 to 17whose victimization experiences exhibited distinct seasonalpatterns compared to adults (see textbox on page 13).


13/22


0

5

10

15

20

25

30

35

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93


Age 1217

Age 18 and older


FIGURE 11

Seasonal rates of simple assault for youth and adults, 19932010

Seasonal patterns in simple assault differed for youth and adults

Simple assault victimization rates were higher among youth ages 12 to 17 than among adults age 18 and older and

exhibited different peak and trough seasons (figure 11). Simple assault rates among youth were lowest during the summer

when the school year ended, and highest in the fall when the school year began. Adult rates of simple assault exhibited

relatively less seasonal fluctuation and were highest in the summer.


14/22


0.0

0.4

0.8

1.2

1.6

2.0

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93



FIGURE 12

Seasonal rates of intimate partner violence, 19932010

TABLE 3

Average percent difference in the seasonal rate of violentvictimization involving intimate partners, weapon use, andinjury, 19932010

Intimate partnerviolence

Serious violenceinvolving weapons

Serious violenceresulting in injury

Winter -11.5%* -5.7%* -6.7**

Spring -5.8* -4.7* -6.7*

Summer ~ ~ ~

Fall -9.0* -3.2 -5.6Note: See appendix table 3 for regression coefficients and standard errors.




Intimate partner violence

o assess the degree o regular seasonal fluctuations inintimate partner violence, rates o all types o violent

victimization in which the victim reported that the offenderwas a current or ex-spouse, or a current or ex-boyriendor girlriend, were examined. Seasonal rates o intimatepartner violence appeared to peak in the summer in morethan hal o the years (figure 12). Statistical analysis o the

intimate partner violence rates indicated that there wereseasonal differences between summer rates and the ratesor other seasons (table 3). Compared to the summer,intimate partner violence was about 12% lower in the winter,6% lower in the spring, and 9% lower in the all (table 3).


15/22


0

2

4

6

8

10

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93



FIGURE 13

Seasonal rates of serious violent crime involving weapons, 19932010

Serious violence involving weapons

Rates o serious violence involving weapons also exhibitedsome statistically significant patterns o seasonal variation(figure 13). Violence involving weapon use tended to behigher during the summer than during the winter and

spring. On average, winter rates o this type o violencewere approximately 6% lower than summer rates, whilespring rates were about 5% lower than summer rates. Tedifferences between the summer and all rates were not largeenough across the years to be statistically significant.


16/22


0.0

0.5

1.0

1.5

2.0

Summer'10

Winter'10

Summer'09

Winter'09

Summer'08

Winter'08

Summer'07

Winter'07

Summer'06

Winter'06

Summer'05

Winter'05

Summer'04

Winter'04

Summer'03

Winter'03

Summer'02

Winter'02

Summer'01

Winter'01

Summer'00

Winter'00

Summer'99

Winter'99

Summer'98

Winter'98

Summer'97

Winter'97

Summer'96

Winter'96

Summer'95

Winter'95

Summer'94

Winter'94

Summer'93

Winter'93



FIGURE 14

Seasonal rates of serious violent crime resulting in injury, 19932010

Serious violence resulting in injury

Seasonal patterns in rates o serious violent victimizationresulting in injury were comparable to those ound orserious violence involving weapons (figure 14). Violent

victimizations resulting in serious injury tended to occur at

higher rates during the summer than during the winter andspring. Winter and spring rates o serious violence resultingin injury were about 7% lower than summer rates. Althoughall rates were also somewhat lower on average than summerrates, these differences were not statistically significant.


17/22


Summary

With the exception o robbery, each o the violent andhousehold property crimes exhibited a pattern o recurrentseasonal fluctuations in their trends rom 1993 to 2010.Te most common pattern or household property crimeswas that victimization occurred at higher rates duringthe summer than during other seasons. Motor vehiclethef rates exhibited less regular seasonal fluctuation than

household burglary and larceny. While summer rates omotor vehicle thef were higher than spring rates, theywere not significantly different rom winter and all rates.Across each type o household property victimization, thefluctuations between the seasonal peak and trough rates areless than 11%. In comparison, the variations in the annual

victimization rates rom 1993 to 2010 were much larger.For example, the household property victimization ratedeclined 62% rom 318.9 per 1,000 households in 1993 to120.2 per 1,000 households in 2010.3Tereore, the variationin household victimization rates across the years was largerthan the seasonal fluctuations within the calendar year.

For violent crimes, the most common pattern showed thatvictimization rates were higher during the summer thanduring other seasons o the year. However, this patternhad two important exceptions. First, unlike any other type

o violence, robbery rates exhibited no regular seasonalvariation over time. Second, simple assault rates were higherin the all than in the winter, spring, and summer. Tesefindings indicating little seasonality in robbery and peakrates o simple assault in the all were similar to those oundin prior BJS research based on victimization data rom themid-1970s.4

Intimate partner violence rates were higher during the

summer than during other seasons o the year. Tis patternwas more similar to the seasonal variations ound or serious

violence than or simple assault victimization. rends inassault resulting in serious injury and assault involvingweapons also indicated that summer rates were higher thanthe rates in other seasons; however, the differences betweenall and summer rates in these types o violence were notstatistically significant. When seasonal differences in varioustypes o violence were ound, the magnitude o the differencebetween the peak and trough rates was 12% or less. Likehousehold property crime, variations in the annual violent

victimization rates rom 1993 to 2010 were larger. Teviolent victimization rate declined 70% rom 49.9 per 1,000persons age 12 or older in 1993 to 14.9 per 1,000 in 2010,

varying more rom 1993 to 2010 than across the seasons othe year.5

3Criminal Victimization, 2010, NCJ 235508, BJS web, September 2011. 4Crime and Seasonality: National Crime Survey Report, NCJ 64818, BJS web,May 1980.

5Criminal Victimization, 2010, NCJ 235508, BJS web, September 2011.


18/22


Methodology

Survey coverage

Te National Crime Victimization Survey (NCVS) is anannual data collection conducted by the U.S. Census Bureauor the Bureau o Justice Statistics (BJS). Te NCVS is asel-report survey in which interviewed persons are asked

about the number and characteristics o victimizations theythey experienced during the prior 6 months. Te NCVScollects inormation on nonatal personal crimes (rape orsexual assault, robbery, aggravated and simple assault, andpersonal larceny) and household property crimes (burglary,motor vehicle thef, and other thef) both reported and notreported to police. In addition to providing annual-level andchange estimates on criminal victimization, the NCVS isthe primary source o inormation on the nature o criminal

victimization incidents. Survey respondents provideinormation about themselves (such as age, sex, race andethnicity, marital status, education level, and income) andwhether they experienced a victimization. Te survey also

collects inormation about the offender in each victimizationincident (such as age, race and ethnicity, sex, and victimoffender relationship), characteristics o the crime (includingtime and place, use o weapons, nature o injury, andeconomic consequences), whether the crime was reported topolice, reasons why the crime was or was not reported, andexperiences with the criminal justice system.

Te NCVS is administered to persons age 12 or older roma nationally representative sample o households in theUnited States. Te NCVS defines a household as a group omembers who all reside at a sampled address. Persons areconsidered household members when the sampled address is

their usual place o residence at the time o the interview andwhen they have no usual place o residence elsewhere. Onceselected, households remain in the sample or 3 years, andeligible persons in these households are interviewed every6 months or a total o seven interviews. New householdsrotate into the sample on an ongoing basis to replaceoutgoing households that have been in the sample or the3-year period. Te sample includes persons living in groupquarters, such as dormitories, rooming houses, and religiousgroup dwellings, and excludes persons living in militarybarracks and institutional settings, such as correctional orhospital acilities, and homeless shelters. (For more detail,see Survey Methodology inCriminal Victimization in the

United States, 2008, NCJ 231173, BJS web, May 2011.)

Nonresponse and weighting adjustments

In 2010, about 41,000 households and 73,300 individuals age12 or older were interviewed or the NCVS. Each householdwas interviewed twice during the year. Te response ratewas 92.3% o households and 87.5% o eligible individuals.Victimizations that occurred outside o the U.S. wereexcluded rom this report.

Estimates in this report use data rom the 1993 to 2010NCVS data files. Tese files can be weighted to produceannual estimates o victimization or persons age 12 or olderliving in U.S. households. Because the NCVS relies on asample rather than a census o the entire U.S. population,weights are designed to inflate sample point estimates toknown population totals and to compensate or surveynonresponse and other aspects o the sample design.

Te NCVS data files include both person and householdweights. Person weights provide an estimate o thepopulation represented by each person in the sample.Household weights provide an estimate o the U.S.

household population represented by each household in thesample. Afer proper adjustment, both household and personweights are also typically used to orm the denominator incalculations o crime rates.

Victimization weights used in this analysis account orthe number o persons present during an incident. High-requency repeat victimizations (or series victimizations)are handled differently in this report than in recent BJS

victimization reports because they present a special problemor seasonality analysis. Series victimizations are similar intype but occur with such requency that a victim is unableto recall each individual event or describe each event in

detail. Survey procedures allow NCVS interviewers toidentiy and classiy these similar victimizations as series

victimizations and to collect detailed inormation on onlythe most recent incident in the series. Because the monthin which a victimization occurred is recorded only or themost recent incident, series victimizations are treated asa single victimization. Te percentage o victimizationsreported as series victimizations declined rom 1993 to2010. In 2010, series incidents accounted or about 1%o household property victimizations and about 3% o

violent victimizations.


19/22


Methodological changes to the NCVS in 2006

Methodological changes implemented in 2006 may haveaffected the crime estimates or that year to such an extentthat they are not comparable to estimates rom other years.Evaluation o 2007 and later data rom the NCVS conductedby BJS and the Census Bureau ound a high degree oconfidence that estimates or 2007, 2008, 2009, and 2010 areconsistent with and comparable to estimates or 2005 and

previous years. Te reports, Criminal Victimization, 2006,NCJ 219413, December 2007; Criminal Victimization, 2007,NCJ 224390, December 2008; Criminal Victimization, 2008,NCJ 227777, September 2009; Criminal Victimization, 2009,NCJ 231327, October 2010; and Criminal Victimization,2010, NCJ 235508, September 2011, are available on theBJS website.

Estimating and analyzing seasonal victimization rates

Victims reports o the month in which the incidentoccurred are used to estimate seasonal rates o victimization.

Seasons o the year are defined according to the month inwhich the season begins. Tereore, winter rates include inDecember, January, and February incidents; spring ratesinclude March, April, and May incidents; summer ratesinclude June, July, and August incidents; and all ratesinclude September, October, and November incidents. Te1993 to 2010 seasonal victimization rates include incidentsoccurring rom December 1992 through November 2010.Seasonal household property victimization rates areestimated by dividing the weighted number o incidentsby the weighted number o households and multiplying by1,000. Violent victimization rates are estimated by dividingthe weighted number o incidents by the weighted number

o persons age 12 and older and multiplying by 1,000.Sampling weights are used in all population estimates.

Te NCVS has three important design eatures thatreduce potential errors in measuring the month in which

victimization occurred. First, the relatively short 6-monthreerence period reduces the chances that victims will orgetthe incident, underreport victimizations, and attributeincidents to the wrong month. Second, the boundingprocedure provides a reerence point or respondentsabout previously reported incidents, and thereby reducestelescoping and minimizes error in assigning dates toincidents. Tird, the rotating panel structure o the survey

is designed so that data are collected rom respondents ineach month o the year. Tis minimizes recency bias errorsin reporting the dates o incidents, which may occur i thesurvey was administered only once during a fixed periodeach year.

Te extent and statistical significance o seasonal patterns inhousehold property victimization and violent victimizationare estimated by developing regression models thatdecompose each o the victimization trends into three

components: the long-term trend, seasonal variationaround that trend, and error. o ensure that the results arerobust, two regression-based approaches were used alongwith various sensitivity tests. For each analysis, variablesrepresenting the seasons o the year were regressed on each

victimization trend to estimate the difference between thepeak seasonal rate and the rates in the other seasons.

Te first approach estimated ordinary least squares

regression models or each victimization series and includedyear fixed effects to account or the long-term trend.Te degree o autocorrelation in the error structure wasestimated using Breusch-Godrey tests or up to our laggedperiods. For victimization series in which autocorrelatederror was present, a lagged indicator was ound to removethe autocorrelation. In a ew instances, autocorrelated errorspersisted despite a variety o lag specifications; thereore,Newey-West standard errors were estimated to take intoaccount the potential influences o autocorrelation andheteroskedasticity in the error structure.

Te second approach re-estimated the regression models

or each victimization series using first differences in ratesto examine whether the seasonality results were replicatedwhen the models accounted or possible stochasticnonstationarity in the trends. Autocorrelated errors weresimilarly assessed and handled in the first-difference modelsand the conclusions about the statistical significance o eacho the seasonal coefficients were replicated in these results.Stationarity in the victimization trends in levels and firstdifferences was examined using Dickey-Fuller and Philips-Perron tests.

Additional diagnostic tests were perormed to ensurethat the models, and thereore estimates o the seasonal

differences, fit the observed seasonal rates. Tese includedvisual inspections o observed and predicted trends andmodel-fit improvement F-tests that compared the resultsbased on long-term trends to those based on long-termtrends plus seasonal variations. Further details about thisstatistical process using the burglary trend as an illustrationare shown in appendix table 4.

Te results rom the series o regression models shown inthis report are presented in levels and converted to relativepercentage differences because this inormation provides amore intuitive interpretation o the magnitude o seasonal

variation relative to the average rate across the time period.

Tese percentages are obtained by comparing the coefficientor each season against the estimated comparison rate(typically the summer) represented by the constant. Teregression results that resulted in the percentages are shownin appendix tables 14.


20/22


APPENDIX TABLE 1

Results of regression analyses of seasonal patterns inhousehold property crime victimization, by type of propertycrime, 19932010

Total property Motor vehicle

Crime Burglary Theft Larceny

Winter -5.91 (0.79)* -1.71 (0.25)* -0.11 (0.07) -4.10 (0.64)*

Spring -6.88 (0.71)* -1.44 (0.23)* -0.28 (0.07)* -5.15 (0.59)*

Summer ~ ~ ~ ~

Fall -2.69 (0.68)* -0.96 (0.25)* 0.00 (0.08) -1.74 (0.59)*

Constant 85.75 (1.98) 16.31 (0.78) 5.00 (0.15) 64.43 (1.49)

Note: All constants are statistically significant (p < .05, two-tail test). Standarderrors are shown in parentheses.

~Comparison season (season with highest average rate).

*Seasonal rate is significantly different from the summer rate (p < .05, two-tailtest).


APPENDIX TABLE 2

Results of regression analyses of seasonal patterns in violent victimization, by type of crime, 19932010Total violence Serious violence Rape/sexual assault Robbery Aggravated assault Simple assault

Winter -0.62 (0.21)* -0.31 (0.09)* -0.06 (0.03)* -0.04 (0.05) -0.23 (0.06)* -0.48 (0.16)*Spring -0.48 (0.16)* -0.26 (0.09)* -0.04 (.03) -0.03 (0.05) -0.20 (0.06)* -0.40 (0.11)*

Summer -0.25 (0.20) ~ ~ -0.01 (0.05) ~ -0.43 (0.15)*

Fall ~ -0.18 (0.09)* -0.07 (0.03)* ~ -0.12 (0.06)* ~

Constant 13.98 (0.61) 5.57 (0.19) 0.67 (0.07) 1.59 (0.08) 3.31 (0.15) 8.59 (0.45)

Note: All constants are statistically significant (p < .05, two-tail test). Standard errors are shown in parentheses.


*Seasonal rate is significantly different from the comparison rate (p < .05, two-tail test).


APPENDIX TABLE 3

Results of regression analyses of seasonal patterns in violentvictimization involving intimate partners, weapon use, and

injury, 19932010Intimate partnerviolence

Serious violenceinvolving weapons

Serious violenceresulting in injury

Winter -0.18 (0.05)* -0.23 (0.08)* -0.12 (0.04)*

Spring -0.09 (0.04)* -0.19 (0.07)* -0.12 (0.04)*

Summer ~ ~ ~

Fall -0.14 (0.04)* -0.13 (0.07) -0.10 (0.05)

Constant 1.56 (0.11) 4.04 (0.13) 1.78 (0.11)Note: All constants are statistically significant (p < .05, two-tail test). Standarderrors are shown in parentheses.


*Seasonal rate is significantly different from the summer rate (p < .05, two-tailtest).

Source: Bureau of Justice Statistics, National Crime Victimization Survey,

19932010.


21/22


22/22

Office of Justice Programs

Innovation Partnerships Safer Neighborhoods

www.ojp.usdoj.gov

Te Bureau o Justice Statistics, located in the Office o Justice Programs,U.S. Department o Justice, collects, analyzes, and disseminates statisticalinormation on crime, criminal offenders, victims o crime, and the operationo justice systems at all levels o government. William J. Sabol is acting director.

Tis report was written and verified by Janet L. Lauritsen and Nicole White,University o Missouri - St. Louis. Lynn Langton assisted with verification.

Jill Tomas edited the report. Barbara Quinn produced the report.

June 2014, NCJ 245959

Documents

Summer crime spike seen, according to federal report