Geographic Crime Analysis

O R I G I N A L P A P E R

The Geography of Citizen Crime Reporting

Elise Wisnieski • Stephanie Bologeorges •

Tina Johnson • David B. Henry

Published online: 19 September 2013

� Society for Community Research and Action 2013

Abstract Research has shown variable conceptualiza-

tions of neighborhood, often inconsistent with administra-

tive boundaries. The present investigation seeks to quantify

the geographic area encompassed by citizens’ reporting of

crime. Two Chicago violence prevention organizations

gathered near real-time citizen reports of crime and other

precursors of violence in a south side community. Over the

course of 6 months, 48 community residents participated in

a weekly telephone survey about incidents occurring in

their community, including crime, incivilities, and disor-

der. For each incident reported in the study community,

respondents were asked to specify its location, whether it

was witnessed or heard about, and if it occurred within one

block of their residence. Incident locations were geocoded

and used to compute distance from residence. Incident

reporting radii were calculated for all types of incidents.

Calculated distances of events reported within a block

revealed discrepancies between resident perceptions and

geographic apportionments. On average, incident reports

spanned just over a half-mile geographic radius from

respondents’ residences. Reporting radii were greater for

more violent incidents and shorter for incidents witnessed

directly. There was no effect of age, gender, length of

residence, or length of participation in the study on

reporting radii. Descriptions of reporting radii and impli-

cations for crime prevention efforts and research are

discussed.

Keywords Crime reporting � Crime prevention � Neighborhood

Introduction

Time and time again, neighborhood-level research has put

forth competing definitions and conceptualizations of

neighborhood and community (Coulton et al. 2001).

Scholars have debated the utility and efficacy of pre-

defined administrative boundaries compared to more fluid

boundaries framed by resident perceptions, the census tract

versus the block group, and various concepts of what

constitutes a block (Taylor 1997; Taylor et al. 1984;

Coulton et al. 2001; Cromartie and Swanson 1996;

O’Campo 2003; Sharkey and Horel 2008; Chilensky 2011).

Others have suggested using radius criteria, such as the

Bayesian model of critical acceptable pedestrian walking

distances or the area surrounding one’s residence (Lee and

Moundon 2008; Seneviratne 1985).

A body of research has been dedicated to deciphering how

best to measure and evaluate neighborhood and other vari-

ables likely to influence and be influenced by neighborhood.

In a study of perceived neighborhood advantage, Cantillon

(2006) concluded that administratively defined boundaries

do not best serve the field of neighborhood effects; instead,

researchers should consider smaller neighborhood concep-

tualizations, such as the street block, to direct community

development and organization efforts. Other studies have

suggested a .25 mile radius as a reliable definition for

E. Wisnieski (&) � T. Johnson CeaseFire/Cure Violence, University of Illinois at Chicago,

1603 W. Taylor St., Chicago, IL, USA

e-mail: [email protected]

S. Bologeorges

School of Public Health, University of Illinois at Chicago,

1603 W. Taylor St., Chicago, IL, USA

D. B. Henry

Institute for Health Research and Policy, University of Illinois at

Chicago, 1747 W. Roosevelt Rd., Chicago, IL, USA

123

Am J Community Psychol (2013) 52:324–332

DOI 10.1007/s10464-013-9597-z

neighborhood based on measures of social contact with

neighbors, neighborhood social perceptions, fear of neigh-

borhood crime, effects of built environment, and satisfaction

with neighborhood quality of life (Kruger 2008).

Additional research advocates the use of multiple defini-

tions and sources of data in the same analysis to accommo-

date the study of various neighborhood processes (O’Campo

2003). For example, Weiss et al. (2007) propose the use of

both direct observation and respondent input; however, such

techniques are often deemed too subjective, inconsistent, and

labor- and time-intensive. In a pilot study comparison of

researcher and resident-defined neighborhoods, Coulton

et al. (2001) overlaid resident-drawn neighborhood maps

with census maps—ultimately concluding that there are

discrepancies between conceptualizations. Although resi-

dent-reported neighborhoods were similar in square mileage

(a mean area of .32 square miles) to corresponding census

tracts, the area did not map directly onto the census tract,

implying that while each conceptualization may be close in

size, boundaries do not align. The authors further emphasize

that such varying geographic apportionments likely result in

different values of neighborhood constructs.

Past neighborhood-level analyses have examined

neighborhood geography and its effects on a range of

variables, including but not limited to: social cohesion,

place attachment, collective efficacy, and perceived crime.

An examination of Baltimore neighborhoods concluded

that, ‘‘… people perceive common boundaries for their neighborhoods (that is, people define their environment

using a common set of blocks, a larger area or a city) and

have common perceptions of the quality of life and safety

of the environment in these neighborhoods’’ (Wilson et al.

2009). Although researchers have experimented with sur-

veying residents to gauge perceptions of crime and safety

in neighborhoods, knowledge of the geography encom-

passed by resident reports of actual crime in their neigh-

borhoods is sparse (Sampson et al. 1997; Hipp 2007). The

present study zeroes in on the geography of near real-time

crime reporting across one Chicago community.

Crime has traditionally been measured using victim-

ization surveys and officially collected statistics. Similar to

more novel attempts to survey residents on perceived

incivilities and disorder, the present study utilizes data

collected from a survey designed to gather citizen reports

of incivilities, disorder, and crime. However, unlike pre-

vious studies, where reports tend to be gathered retro-

spectively, this study is distinctive in that citizen crime

observations are collected in near real-time and are used to

understand resident perceptions of neighborhood bound-

aries (Sampson and Groves 1989; Slocum et al. 2010).

Specifically, the present investigation seeks to quantify the

geography encompassed by citizens’ surveillance of crime

and examine how findings may impact future community-

based crime prevention and research. We attempt to answer

the following questions:

1. What is the average radius of citizen crime reports?

2. To what extent do the distances of citizen crime

reporting vary with the incident-level predictors of

crime type, time, and whether the respondent wit-

nessed or heard about the incident?

3. To what extent do observer characteristics (gender,

age, and length of residence) influence the distances of

citizen crime reports?

4. To what extent do resident perceptions agree with

administrative definitions of community boundaries

and blocks?

This paper first provides the relevant study background

followed by a rationale for the selection of the pilot com-

munity. Then, the process behind reporting radii calcula-

tions and results are discussed, with special emphasis on

variations by crime type, event salience, and how the

respondent found out about the incident. Next, a compar-

ison is drawn between resident perceptions of community

geography against official geographic apportionments.

Finally, implications for neighborhood violence prevention

and research are considered.

Methods

Study Background

This study stems from an ongoing partnership between

CeaseFire, a violence prevention program, and the Chicago

Center for Youth Violence Prevention (CCYVP).

In 2005, the CCYVP and CeaseFire formed a working

research group to enrich the field of violence prevention by

improving the recruitment, training, and deployment of

violence prevention practitioners. The data for this study

were gathered as part of a project that attempted to assess

informative signatures for identifying communities likely

to experience increased youth violence (Henry et al. 2013).

Selecting a Pilot Community

The pilot community selected was the south side Chicago

community area of Englewood. According to the 2010

census, Englewood has a population of 37,403, spanning 3.1

square miles and 11 census tracts. A number of decisions

went into choosing the pilot community. First, existing ties to

CeaseFire were required to ease recruiting community

respondents, explaining the study purpose, and garnering

and sustaining community involvement. Second, high

numbers of crime incidents assured sufficient data to sustain

resident involvement. In 2011 alone, 526 robberies, 704

Am J Community Psychol (2013) 52:324–332 325

123

assaults, and 80 shootings were reported in Englewood.

Englewood is situated in the police district with the highest

shooting and homicide rate in all of Chicago (City of Chicago

2011). Table 1 provides a brief snapshot of the target com-

munity’s demographics. The population is primarily African

American with low educational attainment, low median

household income, and a low rate of home ownership.

Participants

Participant Recruitment

The aim was to achieve geographic representativeness

across the 11 census tracts to maximize geographically

dispersed reporting. Business owners, residents, employees,

and stakeholders were recruited across the 11 Englewood

census tracts. Researchers took two thorough inventories of

community businesses prior to recruitment and then part-

nered with CeaseFire staff to identify recruitment territories.

In keeping with previous research on recruiting ethnic

minorities, potential respondents were approached by

workers familiar with Englewood to overcome respondent

fear and distrust (Arean and Gallager-Thompson 1996;

Thompson et al. 1996). Study flyers were distributed in

mailboxes, and several area non-profit organizations

advertised the study. In the 8 weeks of recruitment,

respondents were invited to participate in a brief weekly

telephone survey about select incidents they may have

observed in their neighborhood over a 6-month period (June

2011–November 2011). Respondents were informed they

would receive a $100 gift-card for their participation. After

informing respondents about confidentiality, privacy, and

their right to withdraw at any time without penalty, partic-

ipants consented to participation. Researchers then recorded

a telephone number, interview availability times, and resi-

dent home address for each respondent.

Study Sample

Participant recruitment yielded 59 potential respondents,

48 of whom actually participated in the study. Table 2

depicts demographic characteristics of the sample. Partic-

ipants were predominantly African–American (97.9 %), a

proportion consistent with the neighborhood population.

The median age was 37 years, slightly older than the

population median age (30.7 years). Figure 1 is a dot

density map depicting the geographic distribution of

respondents. As shown, respondents were successfully

recruited in 9 of the 11 census tracts and were clustered

near the center of Englewood. Of the 48 respondents, 81 %

Table 1 Englewood community profile (US Census Bureau 2010)

Community characteristic Englewood

Population 37,403

Area square mileage 3.07

% African American 97.30 %

% Bachelor’s degree 4.50 %

Median household income $24,308

% Owner occupied housing unit 36.70 %

Median sale price (single family detached) in

2009

$10,000

% Household income under $25 K 54.00 %

2011 Homicide rate 28 (74 per 100,000)

2011 Shooting rate 80 (213 per

100,000) Table 2 Sample demographic characteristics

Sample characteristics Active sample (n = 48)

Male 45.8 % (n = 22)

Female 54.2 % (n = 26)

Age M = 38.76, SD = 11.142

Length of time living in target

community

M = 15.56 years,

SD = 11.92

Fig. 1 Geographic distribution of active respondents. This figure is a dot density map depicting the geographic distribution of active

respondents. To protect confidentiality, dots do not represent actual

respondent addresses, but a random distribution of respondents per

census tract

326 Am J Community Psychol (2013) 52:324–332

123

reported home addresses within Englewood boundaries and

19 % in the surrounding areas.

Measures

A five-item questionnaire was administered to participants

on a weekly basis via telephone interview. Interviewers

asked a series of closed-ended questions (yes/no) about five

categories of incidents they might have witnessed or heard

about in the preceding week in their community, Engle-

wood. The five incident categories were:

1. Threats/bullying/intimidation

2. Fights/beatings

3. Shootings/stabbings/other use of weapons

4. Other incidents, including robberies, sexual assaults,

and vandalism

5. New graffiti.

If a participant responded ‘‘yes’’ to any of the five

incident categories, they were then asked to specify: (1) if

the incident happened within a block of their reported

home address, and (2) if they witnessed the incident

themselves or heard about it from another source.

Qualitative, open-ended items regarding incident

descriptions and locations were then asked to ascertain

exact location and incident type. Interviewers used probing

techniques to aid respondents in pinpointing cross streets,

landmarks, and businesses proximal to the incident loca-

tion. Any additional information given about a particular

incident was recorded (e.g. age of the victim, injuries

sustained by parties involved, approximate date and/or time

of occurrence). This question pattern was repeated for each

incident reported, and participants were able to report more

than one incident per category.

Procedure

Over the course of 24 weeks, beginning in June 2011,

community respondents gave weekly reports of minor

incidents perpetrated in the community via telephone

interview. Each reported incident was geocoded using

ArcGIS software.

Computing Incident Reporting Radii

All participant home addresses and incidents reported were

mapped using ArcGIS software and geocoded to obtain X

and Y coordinate location data. Coordinate data were then

entered into SPSS version 20.0 for analyses. To obtain the

average radius of citizen incident reports, the distance

between each respondent’s home address and the location

of the incident reported was calculated by first converting

X and Y coordinates from degrees to radians. The Haver-

sine formula was used to obtain the distance (in miles)

from location of residence (coordinate pair 1) to location of

each incident (coordinate pair 2):

a ¼ sin dlat=2ð Þð Þ2þcos lat 1ð Þ � cos lat 2ð Þ � sin dlon=2ð Þð Þ2

c ¼ 2 � arctan 2 p

a; p

1 � að Þð Þ d ¼ 3961 � c

where dlon is the difference in longitudes between the

individual’s residence and incident locations, dlat is the

difference in latitudes, lat_1 is the residence latitude and

lat_2 is the incident latitude, all in radians.

The individual reporting radius for each participant was

computed by taking the mean of the distances between the

residence and the incidents. Each incident was coded

according to whether it was witnessed or heard about.

Incidents were also coded by whether they were perceived

as having occurred ‘‘within a block’’ of the residence.

Finally, incidents were coded according to the type of

incident reported (threats/bullying/intimidation, fights/

beatings, stabbings/shootings/other use of weapons, other

incidents, new graffiti).

Examining Predictors of Reporting Radii

To determine if the distances of citizen crime reports

varied by the incident-level predictors of crime type, time,

whether the respondent witnessed or heard about the inci-

dent, and/or respondent-level predictors of gender, age, and

length of residence in the community, a generalized linear

mixed model was employed using a person-period data set.

The unit of analysis was the incident, and the dependent

variable was the distance from the respondent’s home

address to the incident location. Other predictors were

incident type, week of report, and whether the incident was

witnessed or heard about, gender, age, and length of resi-

dence in Englewood. The model included random inter-

cepts for the individual.

Calculating Block-Level Perceptions

Resident perception of a block was calculated using the

mean of all distances from residences to incidents for

incidents reported to have occurred within one block of

residence. The distances of incidents reported as having

occurred within one block were compared to the adminis-

trative definition of one Chicago city block, .125 miles (8

city blocks = 1 mile). The number of incidents reported as

within one block that were actually within .125 miles was

divided by the total number of incidents reported to be

Am J Community Psychol (2013) 52:324–332 327

123

within one block to obtain the block reporting accuracy rate

by incident. As an additional exploration of block-level

accuracy, a second accuracy rate was calculated by

respondent. In this calculation, the percentage of incidents

correctly reported to actually be within 1 block (within

.125 miles of residence) was calculated as the average

accuracy rate by survey participant.

To examine if block reporting accuracy rates were more

likely to coincide with the administrative definition of a

Chicago city block if the event was witnessed by the

respondent, separate accuracy percentages were calculated

for incidents reported to have been witnessed within one

block of residence. The number of incidents reported to

have been witnessed within one block of residence that

were actually within a .125 mile radius of residence were

divided by the total number of incidents reported as wit-

nessed within one block. Block reporting accuracy rates by

incident were calculated separately for each of the five

categories of incidents.

To determine if distances of citizen crime reports that

were reported as having occurred within one block of the

respondent’s place of residence (block-level perceptions)

varied by the incident-level predictors of crime type, time,

whether it was witnessed or heard about, and/or respondent-

level predictors of gender, age and length of residence in the

community, a person-period dataset was created in SAS.

Each incident reported was considered its own case in this

dataset. Crime type was coded categorically (1–5), time was

reported in intervals by week of data collection (1–24), and

how the respondent found out about the incident was coded

in binary (witnessed = 0, heard about = 1).

A generalized linear mixed model was employed for the

binary outcome of whether or not the incident was reported

as being within one block of the respondent’s address. The

model used a binomial distribution with a logit link func-

tion. The predictor variables were identical to the model for

the distances.

Results

The study yielded a total of 644 completed surveys and 459

incident reports by 48 active respondents. Of the 48 active

respondents, 47 of the 48 reported home residences that

could be geocoded (97.9 %). Of the 459 incident reports,

415 could be geocoded (90.4 %). Given that the home

address of one respondent could not be geocoded (and that

respondent reported 3 incidents), 412 incident pairings

could be mapped using ArcGIS for both incident and res-

idence. These 412 incidents were selected for calculations.

Distances from home residence to reported incident loca-

tion ranged from .0004 miles to 7.6650 miles. The number of

incidents reported by participant that could be geocoded for

both residence and incident location ranged from 0 to 36

incidents. The individual reporting radii of individual

respondents ranged from .008 to 1.941 miles, with a mean

individual reporting radius of .55733 miles. The incident

reporting radius for all 412 incidents (weighted mean) was

.54640 miles, indicating citizen crime reports span just over a

half-mile geographic radius from a residence location.

For seven incidents, the respondent did not provide

further information regarding how they had learned of the

incident (witnessed, heard about, both) or whether the

incident had occurred within 1 block of their residence.

Only 405 incidents were used in further calculations. The

reporting radii for witnessed incidents (n = 245) was

0.43885 miles and .69617 miles for incidents heard about

(n = 149).

A total of 252 incidents were reported to have occurred

within one block of a respondent’s residence. For these

incidents, the overall reporting radius was 0.29972 miles,

which indicates the geographic distance residents perceive

to encompass one block. The reporting radius for witnessed

incidents that occurred ‘‘within 1 block’’ (N = 174) was

0.28763 miles, for incidents heard about (N = 70) was

0.31059 miles, and for incidents both witnessed and heard

about (N = 8) was 0.30750 miles.

Reporting radii by survey question for each of the five

incident types were calculated similarly, first for the

aggregate total by question and then for whether incidents

were witnessed or heard about or both, and whether or not

incidents occurred within one block of respondent resi-

dence. A total of 68 reporting radii were calculated. A

summary of all reporting radii are presented in Table 3.

As expected, reporting radii differed significantly by

whether the respondent witnessed the incident or heard

about it, with shorter distances for witnessed incidents

(B = .23, t(395) = 2.84, p \ .01), and longer distances for shootings and stabbings (B = .22, t(380) = 1.96, p = .05).

There were no significant effects by type of incident, age,

gender, length of residence, or the week of the study. Per-

ceptions that an incident had occurred on the respondent’s

block varied significantly by the distance from the respon-

dent (B = -1.53, t(394) = 6.20, p \ .01, OR = 0.21) and by whether the respondent witnessed or heard about the

incident (B = -1.10, t(394) = 6.20, p \ .01, OR = 0.33). Although the omnibus test for incident type was not sig-

nificant, the parameter for fights/beatings was (B = 1.06,

SE = 0.45, t(394) = 2.34, p \ .05), suggesting that respondents were more likely to perceive an incident as

having occurred within a block of their homes if the incident

was a fight or beating. Perceptions of incidents occurring

within a block of the respondent’s home did not vary by any

other predictor.

Approximately half of all incidents reported to have

occurred in Englewood (n = 233) fell within the

328 Am J Community Psychol (2013) 52:324–332

123

Table 3 Reporting radii calculations for all reported and geocoded incidents

N Distance (in miles)

All incident reporting radii 412 0.54640

Witnessed 245 0.43885

Heard about 149 0.69617

Both 11 0.35164

* missing data for 7 incidents

Incidents reported ‘‘Within 1 Block’’

Within 1 block reporting radii 252 0.29972

Witnessed 174 0.28763

Heard about 70 0.31059

Both 8 0.30750

Total (N) Witnessed (N) Heard about (N) Both (N)

Question 1: threats/bullying/intimidation

All incident radii 0.27467 (60) 0.16031 (32) 0.40833 (27) 0.32500 (1)

Reported within 1 block radii 0.14372 (43) 0.10965 (26) 0.19582 (17) N/A

Not within 1 block radii 0.60588 (17) 0.37983 (6) 0.76960 (10) 0.32500 (1)

Question 2: fights/beatings

All incident radii 0.49700 (93) 0.46774 (61) 0.59603 (29) 0.13100 (3)

Reported within 1 block radii 0.34100 (68) 0.35453 (47) 0.33450 (18) 0.13100 (3)

Not within 1 block radii 0.92184 (25) 0.84779 (14) 1.01609 (11) N/A

Question 3: stabbings/shootings

All incident radii 0.72589 (149) 0.56470 (67) 0.88600 (76) 0.49783 (6)

Reported within 1 block radii 0.42111 (79) 0.38620 (51) 0.48617 (24) 0.47600 (4)

Not within 1 block radii 1.06986 (70) 1.13369 (16) 1.07054 (52) 0.54150 (2)

Question 4: other incidents

All incident radii 0.37135 (48) 0.31930 (30) 0.47547 (17) 0.16300 (1)

Reported within 1 block radii 0.06420 (30) 0.06261 (18) 0.05782 (11) 0.16300 (1)

Not within 1 block radii 0.88378 (18) 0.70433 (12) 1.24117 (6) N/A

Question 5: new graffiti

All incident radii 0.48078 (55) 0.48078 (55) N/A N/A

Reported within 1 block radii 0.30347 (32) 0.30347 (32) N/A N/A

Not within 1 block radii 0.59323 (22) 0.59323 (22) N/A N/A

* 1 incident not indicated if within 1 block

Table 4 Within one block accuracy rates

Percentage of incidents

reported ‘‘within a block’’

actually within .125 miles (%)

Average distance of

incidents reported

‘‘within one block’’

(in miles)

Percentage of incidents reported

as both witnessed and ‘‘within

a block’’ actually within

.125 miles (%)

Q1: threats/bullying 60.47 0.14372 61.54

Q2: fights/beatings 54.41 0.34100 54.00

Q3: stabbings/shootings/

other use of weapons

41.77 0.42111 50.91

Q4: other incidents 83.33 0.06420 78.95

Q5: new graffiti 78.13 0.30347 78.13

All incidents 58.40 0.29972 61.00

Am J Community Psychol (2013) 52:324–332 329

123

administrative boundaries of Englewood. Another 27.5 %

mapped onto the community immediately west, West

Englewood. The remaining 22 % were spread across

neighboring communities.

Within-block accuracy rates are presented in aggregate

form as well as by survey question in Table 4. Of the 252

incidents reported to be within one block of residence, only

146 (57.94 %) had calculated distances from residences that

were B0.125 miles, the definition of a Chicago city block.

The average distance of incidents reported as within one

block from residence was 0.29972, suggesting resident

perceptions of a block span this geographic distance. Of all

incidents reported in the study, 149 actually fell within .125

miles of respondent residences. When divided by the num-

ber of study participants (n = 48), this shows an average of

3.10 violent incidents occurred within one administrative

block of participant’s homes over the 24 weeks of the study.

Further, dividing these 149 incidents by the total number of

incidents that could be calculated for distance from resi-

dence (n = 412) reveals that 36.17 % of all incidents

reported occurred within .125 miles of respondents’ homes.

Discussion

This study sheds light on the geography of citizen crime

reporting in a number of ways. First, this report adds to this

line of research by quantifying the area community

respondents perceive themselves as both belonging to and

being in their sphere of awareness. Moreover, the range in

reporting radii exemplifies the inherent difficulties in

assessing resident perceptions of neighborhood geography.

The geography of citizen crime reporting is variable,

influenced both by event severity and whether the

respondent witnessed or heard about an incident. Notably,

individual-level respondent characteristics (including age,

gender, and length of residence) and time do not have an

effect on reporting radii. Results indicate that there are

discrepancies between administrative and resident defini-

tions of neighborhood.

The radius of citizen crime reporting suggests that, when

asked to place crime in physical space, it appears residents

perceive the half-mile radius surrounding their residence as

their neighborhood. Other research has substantiated a .25

mile radius as an acceptable definition for examining both

the effects of built environment and pedestrian walking

distance (Kruger 2008; Seneviratne 1985). It seems that in

the case of reporting crime, radii are nearly doubled—

perhaps reflecting the salience of crime events.

That citizen crime reporting radii vary greatly depend-

ing on the type of incident reported suggests event severity

is a contributing factor to this phenomenon. In the report-

ing distance linear mixed model, there was an overall effect

of the type of incident. Residents were more likely to report

stabbings and shootings that occurred at greater distances

than more minor violent crimes, suggesting an interaction

between event severity and physical proximity. More

severe events may truly ‘‘hit close to home.’’ Respondents

were more likely to recount severe events at a greater

distance from home than less serious events such as bul-

lying or new graffiti, which are more likely to go unnoticed

unless in close proximity to the respondent’s home.

The models of distance and within-a-block perceptions

showed expected effects of whether an incident was wit-

nessed or heard about from others. Not surprisingly, radii

were greater for incidents that were heard about. Common

in child and adolescent developmental research, the con-

cept of ‘home range,’ or the distance individuals travel

away from their residence in the course of their daily

routines and pursuits, may prove useful in understanding

the implications of these results. ‘Home range’ spans pri-

vate and public spaces and has been quantified by using

both an area and a perimeter (Spilsbury 2005). Previous

studies have found an inverse effect of neighborhood vio-

lence on size of ‘home range’ (Matthews 1992). With

respect to this study, the radius encompassing respondent-

witnessed incidents may constitute their ‘home range.’ On

the other hand, if community observers heard about an

event from another source, the reporting radii increased by

roughly 60 %. It seems this extended radius may be due, in

part, to the breadth of social network ties and information

exchange related to violent crime and minor offenses.

There were no effects of age, gender, or length of

respondent’s residence in the community for either the

distance model or the model of within-block perceptions.

These results may have encouraging implications for

efforts to recruit community members for monitoring

activities, either for research or in connection with com-

munity crime prevention efforts. Recruitment may be

rendered less cumbersome because individual-level pre-

dictors showed no effects on either distance or within-

block perceptions, and thus may not need to be taken into

consideration when recruiting community members for

similar efforts. However, given that this was a small, non-

random sample not intended to be representative of any

group other than the neighborhood, future research should

examine how individual-level predictors matter with dif-

ferent populations.

Despite having enrolled respondents from only 9 of the

11 tracts, incident reports were spread across a total of 61

census tracts in Englewood and contiguous areas. As

expected, respondents did not confine incident reports to

administratively defined areas. The fact that roughly half of

the incidents reported actually fell into the administratively

defined boundaries of Englewood provides further evi-

dence of the high level of geographic discordance between

330 Am J Community Psychol (2013) 52:324–332

123

resident perceptions and administrative demarcations of

neighborhood (Cantillon 2006; Coulton et al. 1996).

Yet another indicator of discrepant resident and

administrative definitions involves city block-level esti-

mations. Interestingly enough, resident perceptions of a

block span a far greater distance than administratively

defined city blocks (0.29972 miles vs 0.125 miles). Overall,

only 58 % of resident-defined blocks aligned with the

administrative block designation of .125 miles. Although

some research has lauded the block as a more objective and

identifiable neighborhood boundary, the crime reporting

radii results suggest that the block may not be as objective

as previously thought (Brown et al. 2004). In fact, asking

respondents to report on an administratively defined

neighborhood block may lead to ‘‘spatial mismatch,’’

whereby a respondent recalls an area encased by bound-

aries different from the area where he or she truly lives

(Sampson and Raudenbush 2004). In turn, such mismatch

can pose a threat to the validity of neighborhood-level

analyses and lead to serious information biases (Lebel et al.

2007). The findings of the present investigation lend cre-

dence to these concerns.

Past research has employed measures of neighborhood

social organization based on constructs such as friendship

networks, neighboring, social ties, sense of community,

and civic participation (Leventhal and Brooks-Gunn 2000;

Sampson et al. 2002; Shinn and Toohey 2003). Studies

have also documented the association between neighbor-

hood attachment and heightened vigilance and protective-

ness over fellow neighbors and their residences (Felson

1987). Wilson et al. (2009) contend that the connection

between geography and crime needs take a forefront when

designing community-based crime prevention initiatives,

such as neighborhood watches. If violence prevention

practitioners are able to gauge the area encompassed by

citizen crime reporting, results can then be used in forming

and supporting block clubs and neighborhood watch

groups. Although the present investigation is limited in its

ability to draw conclusions about the relation between the

geography of crime reporting and neighborhood awareness,

social cohesion, and perceptions of safety, this line of

inquiry is recommended for future research.

Limitations

These findings should be reviewed with some caution.

First, the small sample size of this study is not represen-

tative of other populations or neighborhoods, so general-

izations should not be made until the study can be

replicated with a larger sample. Since the population was

entirely African-American and in a high-violence neigh-

borhood, it is unknown how well the results would

generalize to lower-crime neighborhoods with different

racial/ethnic compositions. Also, along the same lines, very

few respondent demographics were collected. This too

makes generalizations difficult to establish. In addition, the

questionnaire items were left open to interpretation of the

respondents (e.g., other violent incidents). Future research

should zero in on specific incident types to see if different

types influence resident perceptions of geography. One

additional limitation is that respondents did not have a

visual reference source to determine when and where

incidents occurred. This could be improved in future

studies by using recall aids (e.g., calendars and community

maps) to obtain more accurate data.

Conclusion

Not only do reporting radii vary considerably based on

crime severity and whether respondents witness or hear

about incidents, resident-based definitions of community

differ from administrative boundaries. Importantly, indi-

vidual-level characteristics of respondents do not have an

effect on reporting radii. Results are useful in launching

new discussion on the geography of crime, neighborhood

definitions, and community-based crime prevention.

Acknowledgments The authors gratefully acknowledge the contri- bution of Shango Johnson for his insights and help with study

implementation. This study was funded by the National Center for

Injury Prevention and Control, Centers for Disease Control and Pre-

vention U81/CCU517816 (University of Chicago, Illinois). The

findings and conclusions in this report are those of the authors and do

not necessarily represent the official position of the Centers for Dis-

ease Control and Prevention.

References

Areán, P. A., & Gallagher-Thompson, D. (1996). Issues and

recommendations for the recruitment and retention of older

ethnic minority adults into clinical research. Journal of Con-

sulting and Clinical Psychology, 64(5), 875–880. doi:10.1037/

0022-006X.64.5.875.

Brown, B. B., Perkins, D. D., & Brown, G. (2004). Incivilities, place

attachment and crime: Block and individual effects. Journal of

Environmental Psychology, 24(3), 359–371. doi:10.1016/j.jenvp.

2004.01.001.

Cantillon, D. (2006). Community social organization, parents, and

peers as mediators of perceived neighborhood block character-

istics on delinquent and prosocial activities. American Journal of

Community Psychology, 37(1), 111–127. doi:10.1007/s10464-

005-9000-9.

Chilenski, S. (2011). From the macro to the micro: A geographic

examination of the community context and early adolescent

problem behaviors. American Journal of Community Psychol-

ogy, 48(1), 352–364. doi:10.1007/s10464-011-9428-z.

City of Chicago. (2011). Data portal. Retrieved from https://data.

cityofchicago.org/.

Am J Community Psychol (2013) 52:324–332 331

123

Coulton, C., Korbin, J., & Su, M. (1996). Measuring neighborhood

context for young children in an urban area. American Journal of

Community Psychology, 24(1), 5–32. doi:10.1007/BF02511881.

Coulton, C., Korbin, J., Chan, T., & Su, M. (2001). Mapping

residents’ perceptions of neighborhood boundaries: A method-

ological note. American Journal of Community Psychology,

29(2), 371–383. doi:10.1023/A:1010303419034.

Cromartie, J. B., & Swanson, L. L. (1996). Census tracts more

precisely define rural populations and areas. Rural Development

Perspectives, 11(3), 31–39. Retrieved from http://permanent.

access.gpo.gov/lps2554/www.ers.usda.gov/Publications/rdp/.

Felson, M. (1987). Routine activities and crime prevention in the

developing metropolis. Criminology, 25, 911–931. doi:10.1111/j.

1745-9125.1987.tb00825.x.

Henry, D. H., Dymnicki, A., Kane, C., Quintana, E., Cartland, J.,

Bromann, K., Bhatia, S., & Wisnieski, E. (2013). Community

monitoring for youth violence surveillance: Testing a prediction

model. Prevention Science, pp. 1–11. doi:10.1007/s11121-013-

0374-x.

Hipp, J. R. (2007). Block, tract, and levels of aggregation: Neigh-

borhood structure and crime and disorder as a case in point.

American Sociological Review, 72(5), 659–680. doi:10.1177/

000312240707200501.

Kruger, D. J. (2008). Verifying the operational definition of

neighborhood for the psychosocial impact of structural deteri-

oration. Journal of Community Psychology, 36(1), 53–60. doi:10.

1002/jcop.20216.

Lebel, A., Pampalon, R., & Villeneuve, P. Y. (2007). A multi-

perspective approach for defining neighbourhood units in the

context of a study on health inequalities in the Quebec City

region. International Journal of Health Geographics, 6(1), 27.

doi:10.1186/1476-072X-6-27.

Lee, C., & Moudon, A. V. (2008). Neighbourhood design and

physical activity. Building Research and Information, 36(5),

395–411. doi:10.1080/09613210802045547.

Leventhal, T., & Brooks-Gunn, J. (2000). The neighborhoods they

live in: The effects of neighborhood residence on child and

adolescent outcomes. Psychological Bulletin, 126(2), 309–337.

doi:10.1037/0033-2909.126.2.309.

Matthews, M. H. (1992). Making sense of place. Hertfordshire,

United Kingdom: Harvester Wheatsheaf.

O’Campo, P. (2003). Invited commentary: Advancing theory and

methods for multilevel models of residential neighborhoods and

health. American Journal of Epidemiology, 157(1), 9–13. doi:10.

1093/aje/kwf171.

Sampson, R. J., Morenoff, J. D., & Gannon-Rowley, T. (2002).

Assessing ‘‘neighborhood effects’’: Social processes and new

directions in research. Annual Review of Sociology, 28(1),

443–478. doi:10.1146/annurev.soc.28.110601.141114.

Sampson, R. J., & Raudenbush, S. W. (2004). Seeing disorder:

Neighborhood stigma and the social construction of ‘‘broken

windows’’. Social Psychology Quarterly, 67(4), 319–342.

doi:10.1177/019027250406700401.

Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighbor-

hoods and violent crime: A multilevel study of collective

efficacy. Science, 277(5328), 918–924. doi:10.1126/science.277.

5328.918.

Sampson, R. J., & W. Byron Groves. (1989). Community structure

and crime: Testing social-disorganization theory. American

Journal of Sociology, 94(4), pp. 774–802. Retrieved from

http://www.jstor.org/stable/2780858.

Seneviratne, P. (1985). Acceptable walking distances in central areas.

Journal of Transportation Engineering, 111(4), 365–376. doi:10.

1061/(ASCE)0733-947X.

Sharkey, J. R., & Horel, S. (2008). Neighborhood socioeconomic

deprivation and minority composition are associated with better

potential spatial access to the ground-truthed food environment

in a large rural area. The Journal of Nutrition, 138(3), 620–627.

Retrieved from http://jn.nutrition.org/content/138/3/620.full.

Shinn, M., & Toohey, S. M. (2003). Community contexts of human

welfare. Annual Review of Psychology, 54(1), 427. doi:10.1146/

annurev.psych.54.101601.145052.

Slocum, L. A., Taylor, T. J., Brick, B. T., & Esbensen, F. (2010).

Neighborhood structural characteristics, individual-level atti-

tudes, and youths’ crime reporting intentions. Criminology,

48(4), 1063–1100. doi:10.1111/j.1745-9125.2010.00212.x.

Spilsbury, J. C. (2005). ‘We don’t really get to go out in the front

Yard’—children’s home range and neighborhood violence.

Children’s Geographies, 3(1), 79–99. doi:10.1080/1473328050

0037281.

Taylor, R. B. (1997). Social order and disorder of street blocks and

neighborhoods: Ecology, microecology, and the systemic model

of social disorganization. Journal of Research in Crime and

Delinquency, 34(1), 113–155. doi:10.1177/002242789703400

1006.

Taylor, R. B., Gottfredson, S. D., & Brower, S. (1984). Block crime

and fear: Defensible space, local social ties, and territorial

functioning. Journal of Research in Crime and Delinquency,

21(4), 303–331. doi:10.1177/0022427884021004003.

Thompson, E. E., Neighbors, H. W., Munday, C., & Jackson, J. S.

(1996). Recruitment and retention of African American patients

for clinical research: An exploration of response rates in an

urban psychiatric hospital. Journal of Consulting and Clinical

Psychology, 64(5), 861–867. doi:10.1037/0022-006X.64.5.861.

Weiss, L., Ompad, D., Galea, S., & Vlahov, D. (2007). Defining

neighborhood boundaries for urban health research. American

Journal of Preventive Medicine, 32(6, Supplement), S154–S159.

doi:10.1016/j.amepre.2007.02.034.

Wilson, R. E., Brown, T. H., & Schuster, B. (2009). Preventing

neighborhood crime: Geography matters. NIJ Journal, 263, NCJ

226875, Retrieved on 07/22/13 from http://www.nij.gov/

journals/263/neighborhood-crime.htm#note7.

332 Am J Community Psychol (2013) 52:324–332

123

Copyright of American Journal of Community Psychology is the property of Springer Science & Business Media B.V. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.