ASSAULTS WITHIN ADJC SECURE
CARE FACILITIES
SUBMITTED TO:

Michael D. Branham
Lou Goodman
Tom Gronski
Jim Hillyard
Esteban Veloz

SUBMITTED BY:
Stella Vasquez
John Vivian, Ph.D.
Gopal Chengalath, Ph.D.
Jennifer Grimes
Research and Development
June 17, 2004

Assaults within correctional facilities can cause major problems. Assaults can disrupt the
operation of correctional programming efforts and create an environment of fear and
intimidation. They can result in serious injuries requiring transportation of staff or
juveniles to a hospital. Research conducted in adult correctional facilities has shown that:
a well controlled institution is in the best interests of prisoners…there is a
tradition among many social scientists to write about disorder and control in
prison as a battle between oppressed prisoners and an authoritarian regime…the
fact is that the greatest risk faced by a prisoner is victimization from other
prisoners. It is now argued by many commentators that one of the unintended
consequences of the development of more liberal regimes in the 1970s and 1980s
was a power shift from staff to prisoner elites and an accompanying increase in
prison violence and disorder. (Wortley, 2002, 11)
Director Michael Branham is concerned about assaults within Arizona Department of
Juvenile Corrections (ADJC) facilities and he convened a committee to examine this
problem and to develop appropriate recommendations. The committee is composed of
Lou Goodman, Division Director Legal Systems, Tom Gronski, Assistant Director,
Youth Management, Jim Hillyard, Assistant Director, Operations, Esteban Veloz,
Administrator, Policy Unit and John Vivian, Administrator, Research and Development.
The purpose of this report is to 1) provide the Director and the committee with
background data on assaults within ADJC, 2) provide the Director and the committee
with an overview of requirements contained in the existing ADJC procedure(s) and
information on how some ADJC staff report they are complying with those requirements,
and 3) provide the Director and committee with background on three recommendations
designed to address the assault dilemma within ADJC.

1

The main findings of this study are:
•

While the number of assaults decreased between 2002 and 2003, the number of
staff assaulted by juveniles and the number of assaults with injuries increased.

•

The assault and assault with injury rates (per 100 juveniles) increased between
2002 and 2003.

•

As the number of assaults committed by a juvenile increases, the probability of
future assaults by that juvenile also increases.

•

Two-thirds (67.6%) of all assaults occurred in ADJC dayrooms, lounges,
hallways, youth rooms or classrooms.

•

Mondays and Tuesdays were the most common days for assaults.

•

More than half (51%) of the assaults were during the day shift, 48% were during
the swing shift and 1% were at night.

•

Approximately half (47.6%) of the assaults in 2003 resulted in injuries

•

Adobe Mountain School (AMS) had fewer assaults with injuries than expected.
Catalina Mountain School (CMS) and Encanto (ENC) had more assaults with
injuries than expected.

•

More assaults with injuries occurred than expected in the following locations:
lounges, hallways or facility grounds; fewer assaults occurred than expected in
juvenile rooms and education.

•

The commission of an assault within the first month of commitment did represent
a significant indicator of a juvenile’s future likelihood of becoming a chronic,
serious assailant.

•

Among a selected random sample of serious assaults that were examined within
CMS education, we found that the typical assailant was a young, minority gang
member with a history of violent behavior. The CMS serious assaults examined
overwhelmingly occurred between youth who were from different housing units,
thus suggesting that Education/Recreation may have provided an opportunity for
motivated juveniles to assault others.

•

Moderately strong statistical support was found for the notion that larger housing
units and higher staff vacancy rates spawn more assaults.

•

Juvenile age, emotional stability, race and public safety risk level were found to
be statistically significant indicators of being a, chronic, serious juvenile assailant.
Public safety risk level, while statistically significant, may not be a good predictor
2

of assaults. While all the factors considered in the study are statistically
significant, none of them individually, nor all of them collectively are
substantively significant.
•

ADJC Procedure 1190.02 requires Housing Unit Managers/Security Captains to
hold incident debriefings after a juvenile or staff assault.

•

All 4 Superintendents, 30 ADJC Housing Unit Managers and 4 Security Captains
were sent a questionnaire regarding debriefings held at their facilities. Only
eleven Housing Unit Managers reported that they conducted debriefings after a
juvenile or staff assault within the last 90 days. We were unable to obtain
information from Superintendents regarding the required assault reports to the
Assistant Director. The debriefings that were held reportedly complied with
almost all of the debriefing requirements.

•

An analysis of several different incident/assault reporting and debriefing systems
revealed that the design of our system compared favorably to others.

The study presents three suggestions to address the issue of assaults within ADJC:
•

Develop a better behavior management program for assaultive and non-assaultive
juveniles;

•

Fully implement the existing procedural requirements to conduct assault
debriefings; and

•

Consider adopting Dr. Wortley’s recommendations to provide more structure and
routines for times and places where assaults occur.

The report is organized into six sections. First, some basic information is presented on
ADJC assaults during 2002 and 2003. The second section presents information on
assaults with injuries. Third, the results of a survey of ADJC Superintendents, Housing
Unit Mangers and Security Captains are presented. Fourth, three suggestions to address
the ADJC assault problem are presented. The fifth section contains study conclusions.
Finally, the appendix contains flow charts of assault reporting and debriefing
requirements contained in ADJC procedure 1190.02, Incident Debriefing.

3

Table of Contents
ADJC ASSAULTS IN 2002 AND 2003 ........................................................................... 5
SERIOUS ASSAULTS ................................................................................................... 11
DETERMINANTS OF A SERIOUS CHRONIC ASSAILANT................................. 21
SURVEY OF SUPERINTENDENTS, HOUSING UNIT MANAGERS AND
SECURITY CAPTAINS ................................................................................................ 24
CONCLUSIONS ............................................................................................................. 28
APPENDIX...................................................................................................................... 31

4

1.

ADJC Assaults in 2002 and 2003
Figure 1: Assaults within ADJC, August 2002 through December 2003, by Type
250

200

150

100

50

Mutually Instigated Fight
Staff Assaulted by Juveniles

em
be

em
be

r

r

er

D

ec

ov
N

O

ct

ob

be
r

st

pt
em
Se

Au
gu

Ju
ly

Ju
ne

M
ay

Ap
ril

ar
ch
M

ry
ru
a

y
Fe
b

ua
r
Ja
n

ob
er
N
ov
em
be
r
D
ec
em
be
r

O

ct

be
r
pt
em

Se

Au
gu

st

0

Juvenile Assaulted by Juveniles
Total

Assaults1 are decreasing within ADJC secure care facilities. On average, the ADJC had
172.6 assaults per month from August to December2 of 2002 and 166.2 assaults per
month in 2003. As shown in Figure 1, the total number of assaults per month increased to
209 in April of 2003 but then declined after that. Most assaults3 in 2003 involved
juveniles assaulting juveniles (927 or 46.5%); followed by mutually instigated fights (581
or 29.1%) and staff assaulted by juveniles (484 or 24.3%).

Two (Juveniles Assaulted by Juveniles and Mutually Instigated Fights) of the three
assault categories decreased between 2002 and 2003. Meanwhile, there was an increase
in Staff Assaulted by Juveniles. On average, ADJC secure facilities had 35.2 Staff
Assaults by Juveniles per month in 2002 and 40.3 per month in 2003.
1

The terms assault and serious assault are used throughout this report to refer to assaults and serious
assaults that are reported through the ADJC incident reporting system. An unknown percentage of assaults
that occur are unreported by the incident reporting system (Maitland and Sluder, 1998, p. 55.)
2
The ADJC incident reporting system was audited in August of 2002. Assault data recorded previous to the
audit were not reliable.
3
Allegations of juveniles assaulted by staff are forwarded to and investigated by ADJC’s Inspections and
Investigations Section.

5

Figure 2: ADJC Assault Rates, August 2002 through November 2003
25

20.0
20

19.5
18.9
18.2
16.7

15

13.5

13.1

14.2

13.7

15.5

15.4

15.3

14.5

14.0
12.9

13.1

Feb

March

10

5

0
August

Sept

Oct

Nov

Dec

Jan

April

May

June

July

August

Sept

Oct

Nov

While the population of juveniles in ADJC secure care has decreased, the assault rate has
increased. The ADJC secure population decreased from an average of 852 in August of
2002, to 697 in November of 2003. Despite this decrease in population, the assault rate
increased. As shown in Figure 2, the assault rate (per 100 juveniles) increased from 13.5
in August of 2002 to 15.5 in November of 2003. This change represents a 14.8% increase
in the ADJC assault rate.

Some support was found for the notion that larger housing units spawn more assaults. In
fact, a moderate (r=.30, p< .08, N=37) correlation was found between housing unit
average daily population and the number of assaults. A stronger correlation between
housing unit average daily population and the number of assaults was found at the Adobe
Mountain School (r=.57, p< .05, N=15).

6

Figure 3: Chronic Assailants by Location
Sunrise
3%
Eagle Point
17%

Adobe Mountain
28%

Encanto
15%

Catalina Mountain
25%

Black Canyon
12%

An in-depth analysis was conducted of the 284 juveniles (31.2% of all assailants) who
were responsible for 81% of the assaults in 2003. Labeled Chronic Assailants, almost all
(87%) were male and many were either Hispanic (44.4%) or White (33%). Chronic
Assailants were somewhat older in that 66% were either 16 or 17 years old. More than
half (56.3%) of the Chronic Assailants had gang affiliations4. Two thirds (66.5%) were
committed on a felony level offense, and 33.5% were committed on a misdemeanor level
offense. Almost all (91.2%) of them had three or more adjudications, and violent
behavior, school adjustment, substance abuse, peer relationship, mental health or
intellectual problems. Over a third of the Chronic Assailants were from Maricopa County
(37.3%); a third (31.3%) were from Pima County and the remaining third (31.3%) were
from a rural Arizona county. As shown in Figure 3, over one quarter (28%) of the

4

58.14% of the male Chronic Assailants had a gang affiliation compared to 44.7% of the females.

7

Chronic Assailants were housed at the Adobe Mountain School; the Catalina Mountain
School had the second highest concentration of Chronic Assailants (25%).

Figure 4: Location of Assaults

All Other
13%
Dayroom/Lounge/Hall
36%
Recreation
6%
Dining Area
7%

Grounds
7%

Classroom
13%
Youth Room
18%

Two-thirds (67%) of all assaults occurred in ADJC dayrooms, lounges, hallways, youth
rooms or classrooms (see Figure 4). Approximately one-fifth (20%) occurred on facility
grounds, dining or recreation areas. The remaining 13% of the assaults occurred in one of
13 other locations recorded in Youthbase. Female assaults followed a similar pattern:
73% of the female assaults occurred in dayrooms, lounges, hallways, youth rooms or
classrooms.

8

Figure 5: Assaults by Day of Week

327
306
285
269

261

260

212

Monday

Tuesday

Wednesday

Thursday

Friday

Saturday

Sunday

Mondays and Tuesdays were the most common days for assaults, and assaults rarely
occurred on Sundays (see Figure 5). Tuesday (not shown) was the most common day for
female assaults (17.5%).

Figure 6: Assaults by Time of Day
152
146
139

135
125

123

141

135

124

122
114

110

106
99

66
54

16
6

7

1 to 6 6:00
AM
AM

7:00
AM

8:00
AM

9:00
AM

10:00 11:00 12:00
AM
AM
AM

1:00
PM

2:00
PM

3:00
PM

9

4:00
PM

5:00
PM

6:00
PM

7:00
PM

8:00
PM

9:00
PM

10:00 12:00
PM
AM

More than half (51%) of the assaults were during the day shift, 48% were during the
swing shift and 1% were at night. Female assaults had a slightly different pattern, in that
more than half (51%) were during the swing shift, 48% were during the day shift and 1%
were at night. As shown in Figure 6, the 4 p.m. hour had the most assaults, followed by 1
p.m. and 8 p.m.

The pattern of most assaults occurring during the day shift was true only for Mutually
Instigated Fights. In fact, 55% of the Mutually Instigated Fights were during the day
shift, 44% were during the swing shift and 1% were at night. In the case of Juveniles
Assaulted by Juveniles and Staff Assaulted by Juveniles, half of the assaults (49.7% and
49.5% respectively) were during the swing shift and most of the remaining assaults were
during the day shift. Two percent of the Staff Assaults by Juveniles occurred at night.

The likelihood of a juvenile committing future assaults given a certain number of
previous assaults may be useful information to assist ADJC management efforts to
control assaults within our facilities. In order to explore this matter, conditional
probabilities were calculated for future assaults using the formula: P (A/B) = P (A and
B)/P(B).
Table 1
Assault Probabilities for Juveniles Committed to ADJC in 2002
Number of Assaults
Probabilities
First
50%
Second given a first assault
61%
Third given a second assault
66%
Fourth given a third assault
76%
Fifth given a fourth assault
75%
Sixth given a fifth assault
77%

10

Table 1 shows that as the number of assaults increases, the probability of future assaults
also increases. In fact, there is a 50% chance that a randomly chosen juvenile will commit
an assault. The probability of that juvenile committing a second or third assault increases
to 61% and 66% respectively, while the probabilities for a fourth, fifth and sixth assault
are even higher. These results indicate that the appropriate time to present interventions
to a juvenile is after the first assault, because the likelihood of future assaults is high and
grows as the number of assaults grows.

2.

Serious Assaults

Approximately half (47.6%) of the assaults in 2003 resulted in injuries. These assaults
have been defined as serious assaults, and this section reviews the key findings from our
analysis of serious assaults.

Serious assaults increased by 19.7% between 2002 and 2003. Between August and
December of 2002, ADJC had an average of 56.4 serious assaults per month, and this
average increased to 76.1 per month in 2003. The largest increase in category of serious
assaults occurred among juveniles assaulting staff (69.2%), followed by mutually
instigated fights (55%) and then juveniles assaulted by juveniles (20.5%).

11

Figure: 7 ADJC Serious Assault Rates, August 2002 - November 2003
25

20

15

13.5

10.7
10.1
10

8.6

8.4

8.0

10.4

10.2

Jul

Aug

12.1

12.2

Sep

Oct

11.8

12.4

7.9

7.0

6.7

6.1
4.3

5

0
Aug

Sep

Oct

Nov

Dec

Jan

Feb

Mar

Apr

May

Jun

Nov

Dec

As can be seen from Figure 7, ADJC’s serious assault rates (per 100 juveniles) also
increased during this time. ADJC had a serious assault rate of 6.7 in August of 2002, and
that rate increased to 12.4 in December of 2003. The serious assault rates increased by
85% during the time period we are examining.

12

Figure 8: Comparison Between Percent of ADJC Population at
Each Facility and Percent of Serious Staff Assaults

45.2%
41.5%

17.0%

17.6%

16.0%
14.2%

15.7%

9.1%

8.5%
5.4%

7.1%
2.8%

AMS

CMS

BCS

ENC

Average Daily Population 2003

EP

PV

Serious Staff Assaults

Figure 8 displays a comparison between the percentage of juveniles housed in each
ADJC secure care facility to the percentage of serious staff assaults. Figure 8 shows that
AMS had fewer serious staff assaults than would be expected using this methodology,
while CMS and ENC had more than would be expected.

Serious assaults had a similar daily pattern as all assaults in that they tended to occur
earlier in the week. Serious assaults were more common on Mondays (17.6%) and
Tuesdays (16.8%), than Saturdays (13.3%) and Sundays (12.2%).

Serious assaults tended to exhibit the same location pattern as we found for all assaults.
In fact, serious assaults tended to occur most often in dayrooms/lounges/halls (36%)

13

followed by youth rooms (21%), classrooms (12%), facility grounds (7%), recreation
(7%), dining areas (6%) and other (10%).

Certain locations within ADJC secure care had more assaults than would be expected
given the estimated amount of time juveniles spend at those locations. As a result of
committee interest in comparing the percentage of assaults that occurred by location to
the estimated time juveniles spend at each location, we obtained corresponding figures
for the four ADJC facilities. Potential trouble spots are indicated by large differences
between the estimated amount of time spent at a particular location and the percentage of
serious assaults that occurred there. As shown in Table 2, more serious assaults occurred
while juveniles were in lounges, hallways, or grounds5 than would be expected. Fewer
serious assaults occurred than would be expected while juveniles were in their rooms or
in education6.
Table 2

Location
Room
Education
Lounge/Hall
Dining Area
Grounds
Movement
Recreation
Other
Total

Comparison Between Time Juveniles Spend at Different Locations
to the Serious Assaults That Occurred There, August 2002- April 2004
Adobe
Catalina
Time Spent*
Black Canyon
Eagle Point
Mountain
Mountain
48.3%
15.5%
28%
30%
25.9%
24.3%
12.9%
17%
21%
30.5%
16.3%
44%
35%
28.2%
26.8%
5.2%
6.7%
8%
3%
6.1%
2.1%
7%
7%
7.8%
11%
3.8%
1.8%
0%
2%
1.2%
unknown
8.8%
3%
5%
8.5%
0%
3.3%
2%
3%
0%
100%
100%
100%
100%
100%
(n=705)

(n= 145)

(n=450)

* Time spent estimates were provided by Eagle Point School.
5

The amount of time juveniles spend at recreation during the week was unavailable. For comparison sake,
it was estimated that juveniles spend 4.2% of their time on the weekends at recreation.
6
More serious assaults occurred in education than expected at the Eagle Point School.

14

(n=193)

Because of the high proportion of serious assaults that occurred at BCS lounges and
hallways, a closer look was taken at the serious assaults that occurred there. In large part,
the serious assaults that occurred in BCS lounges and hallways involved juveniles
assaulting juveniles, they tended to occur in the fall months of September or October.
Sunday was found to have a disproportionate number of serious assaults and the serious
assaults commonly occurred in the early evenings. In fact, more than half (59.1%) of the
BCS serious assaults in lounges, etc. involved juveniles assaulting juveniles, one quarter
(24.5%) involved mutually instigated fights and the remainder (16.3%) involved
juveniles assaulting staff. The three highest months in 2003 for BCS serious assaults in
lounges, etc were September (6), October (5) and December (4). In fact, the six that
occurred at BCS in September of 2003 represented more than double the annual monthly
average of 2.8. Sunday was the single most common day for BCS serious assaults. In
fact, for the relevant time period, 11 or 22.4% of the serious assaults that occurred in
BCS lounges etc. occurred on a Sunday. The only other above-average day for BCS
serious assaults in lounges etc. was Tuesday with nine of them. Almost three-quarters
(73.5%) of the BCS serious assaults that occurred in lounges etc. happened during the
swing shift, and two-thirds (67.3%) occurred between 3 and 8 pm.

A closer look also was taken of the pattern of serious assaults at CMS. In large part, the
serious assaults that occurred in CMS lounges etc. involved juveniles assaulting
juveniles, tended to occur in the summer months of May, June or July, midweek was
found to have a disproportionate number of serious assaults and the CMS serious assaults

15

tended to occur in the evenings. In fact, slightly more than half (52%) of the CMS serious
assaults in lounges etc. involved juveniles assaulting juveniles, more than a quarter
(28.3%) involved mutually instigated fights and the remainder (19.7%) involved
juveniles assaulting staff. The three highest months for CMS serious assaults were May
(8), June (8) and July (10). More CMS serious assaults in lounges etc. occurred on
Wednesdays than any other day, however, a relatively high proportion of serious assaults
also occurred on Tuesdays and Thursdays. More than two thirds (67.7%) of the CMS
serious assaults that occurred in lounges etc. happened during the swing shift; two-thirds
(62.9%) occurred between 4 and 8 pm.

The committee expressed interest in having more qualitative information on serious
assaults. As a result, R/D staff undertook a review of 20 randomly selected incident
reports involving serious assaults that occurred during 2003 at CMS or BCS. The
committee expressed interest in the serious assaults that occurred in either the education
and recreation areas, so the cases we selected occurred in those areas. Results of our
review are organized by facility.

The typical CMS assailant was a young, minority gang member with a history of violent
behavior. In at least half of the CMS cases, the serious assault was the result of gang
activity. The CMS serious assaults examined overwhelmingly occurred between youth
who were from different housing units, thus suggesting that Education/Recreation may
have provided an opportunity for motivated juveniles to assault others. Nearly half of the
juveniles committing the serious assaults had family histories of domestic violence. Many

16

(70%) of the suspects were younger than their victims. This finding is inconsistent with
the notion that assaults commonly involve older juveniles preying upon younger ones.
Specific findings for the 10 CMS serious assaults we examined include:
1) Juvenile Demographics
a) All of the suspects were minorities.
b) Almost all (90%) of the suspects had a history of violent behavior.
c) Many (70%) of the suspects were gang members.
d) More than half (60%) of the suspects were younger than their victims.
e) Victims were evenly divided between Whites (50%) and Mexican-Americans
(50%)
2) Situational Characteristics
a) Almost all of the cases (90%) involved only two juveniles.
b) Almost all (90%) of the serious assault cases involved juveniles from different
housing units.
c) Half (50%) of the serious assaults involved gang activities.
d) One third (30%) of the cases occurred when no YCO was present; one third
(30%) occurred when a YCO I was present and 20% occurred when a YCO II was
present.
The typical BCS suspect was a Hispanic or White juvenile with a history of violent
behavior. She was just as likely to assault a staff member as another youth. No pattern
was evident regarding the age of the BCS suspects, other than juveniles between the ages
of 15 and 15 ½ were more likely to assault staff than each other. The juveniles were
equally likely to belong to the same housing unit, so education or recreation did not
provide BCS females with their only opportunity to assault one another. BCS assaults
typically occurred when no staff members were present or when only teachers were
present. Specific findings for the 10 BCS serious assaults we examined include:
1) Juvenile Demographics
a) Almost all (92%) of the suspects had a history of violent behavior.
b) Suspects were slightly more likely to be minorities (54%) than Whites (46%).
c) Staff (50%) and juveniles (50%) were equally likely to be victims of the assaults.
d) One third (38%) of the suspects had a gang affiliation.
e) One third (38%) of the suspects had a history of domestic violence.

17

2) Situational Characteristics
a) Most (70%) of the serious assaults occurred when neither a YCO I nor a YCO II
were present
b) Most (60%) of the juvenile assaults on staff occurred as a result of staff efforts to
redirect the juveniles
c) Most (60%) of the cases involved only two juveniles, 20% involved three or more
juveniles and it was impossible to ascertain the number of juveniles involved in
the remaining 20% of the cases.
The committee had two additional questions on serious assailants. First, they wanted to
know the probability of a juvenile becoming a chronic, serious assailant if s/he had
committed an assault during their first month with ADJC. Second, they wanted to know
the probability of a juvenile becoming a chronic, serious assailant if s/he had any history
of confrontation with a police or correctional officer. Using logistic regression analysis
we found that the commission of an assault within the first month of commitment did
represent a significant indicator of a juvenile’s future likelihood of becoming a chronic,
serious assailant. For example, the likelihood of a 16 year old, medium risk, White
juvenile becoming a chronic serious assailant tripled, from 3% to 9% when it was found
they committed an assault during their first month. On the other hand, we found that a
juvenile’s history of confrontation with a police or correctional officer was not a
significant predictor for becoming a chronic, serious assailant.

There was a moderately strong (r=.40, p≤ .05, N=31) relationship between staff vacancy
rates and serious assault rates. Our data for this issue was from 31 ADJC housing units
during the second half of 2003. To control for variations in housing unit population we
used the housing unit’s serious assault rate (per 100 juveniles) as the dependent variable.

18

We found a slightly stronger relationship (r=.50, p≤ .05, N=16) between staff vacancy
rates and assault rates at the AMS.

We were unable to find empirical support for the notion that when visitation day occurs
affects when assaults occur. Indeed, the notion was raised that juveniles would curtail
their assaults and other misbehaviors immediately prior to visitation, so they wouldn’t
risk losing visitation privileges. We examined this notion by identifying when visitation
normally occurred at each ADJC housing unit and then examining the corresponding data
on serious assaults at each housing unit. Fully 94% (30 out of 32) of the ADJC housing
units had visitation or Saturdays and Sundays; the only exceptions were the Alpha Unit
with visitation on Tuesdays, and the Phoenix Unit with visitation on Fridays. We found
that the days immediately prior to visitation at the Alpha Unit were high serious assault
days (Saturday = 20%, Sunday = 22% and Monday = 20%) and that visitation day at the
Phoenix Unit was the single highest serious assault day of the week (18.3%). Both of
these findings are contrary to the notion that the scheduling of visitation day affects when
juveniles engage in serious assaults. Despite lack of support for the notion as presented
by the Alpha and Phoenix housing units, the overall trend for assaults to occur more than
would be expected on Mondays and Tuesdays with most visitations occurring on
Saturdays and Sundays, still provides some general backing for this notion.

19

Figure 9: Serious Assaults in 2003
by Week With School Breaks Indicated
35
Summer Break

Intersession
Break
30

Serious Assaults

25

20

15

10

5

0
1

3

5

7

9

11

13

15

17

19

21

23

25

27

29

31

33

35

37

39

41

43

45

47

49

51

Weeks

There may be a relationship between the onset of school and when assaults occur. An
issue raised by the Assaults Committee involved the notion that an inordinate number of
assaults seemed to occur immediately after school started. We examined this issue by
identifying the school intersession and summer breaks and then comparing the weekly
pattern in serious assaults to the school break schedule. Figure 9 displays some support
for this notion, in that the weeks immediately after the first (week 1), second (week 12)
and third (week 42) intersession breaks had increases in assaults. The period after the
extended summer break did not support this notion, however. This notion that the onset
of school correlates with assaults may deserve closer scrutiny in order to better
understand the nature of this relationship. This notion could be examined for prior years
to see if this is a stable relationship, and/or the data could be disaggregated by different

20

ADJC facility to reveal if perhaps this effect tends to happen at certain secure school
and/or housing units and not others.

We could not find support for the notion that exercise inhibits assaults. We examined this
notion by using a Performance Based Standards (PBS) data element on the percent of
juveniles who report receiving at least one hour of large muscle exercise each day on
weekdays and two hours each day on weekends. We examined this for the last three PBS
reporting periods7 by ADJC facility and we did not find a statistically significant
relationship. A limitation to this research may be that using this PBS measure may be far
too crude of a measure to be sensitive to the actual relationship (if there is one). Perhaps a
more effective approach to address this question might entail the analysis of PBS data for
individual juveniles related to their exercise and assaultive habits.

Determinants of a Chronic Serious Assailant
The following is an attempt to determine the characteristics of a chronic serious assailant.
A chronic serious assailant is a juvenile who has committed four or more injury assaults.
Chronic serious assailants accounted for over 70% of all injury assaults during the
calendar year of 2003. The strategy adopted was to find significant factors or
characteristics of a chronic serious assailant using a logistic regression model, and to
calculate the probability of a juvenile becoming a chronic serious assailant. If the
probability exceed 50%, it may be safe to conclude that the juvenile is a potentially
chronic serious assailant.

7

October 2002, April 2003 and October 2003.

21

The sample consisted of 1652 juveniles who were in ADJC secure care during
the calendar year 2003, of which 117 were chronic serious assailants. Among a
host of potential factors, the following were found to be statistically significant:
low, medium, high risk levels, race of the youth categorized as White, African
American, Hispanic, all the rest, age of the youth, and emotional stability. Risk
scores of juveniles who were assessed on the old criterion were recalculated
based on the new formula. The estimates of the logit model are presented below.
VARIABLE
ESTIMATED
NAME
COEFFICIENT
Med. Risk level
0.81906
High Risk level
0.51005
African American
0.99135
Hispanic
0.50609
All others
0.60014
AGE
-0.28893
Emotional Instability
1.0302
CONSTANT
0.36407
LIKELIHOOD RATIO TEST = 53.2257

T-RATIO
2.0072
1.1696
3.1329
2.0735
1.5953
-3.4611
5.0750
0.2670
WITH
7

D.F.

The odds of being a chronic serious assailant increases by 2.3 if a juvenile has a medium
risk level as compared to low risk level, and by 1.7 if a juvenile has high risk level as to
compared to low risk level. The result that a juvenile has a higher chance of being a
chronic serious assailant if their risk level is medium as compared to high risk level is
problematic. One possible explanation of this result is that risk levels, though statistically
significant, are not sensitive enough to capture differences among chronic and nonchronic assailants. In other words, risk level may not be a good predictor of assaults.
Further investigation is needed to resolve this result. An African American youth is 2.7
times more prone to being a chronic serious assailant than a White, a Hispanic 1.7 times
more than White and, other races 1.8 times more than a White. An increase in age
decreases the chance of being a chronic serious assailant by 1.3 times more than a youth

22

who is a year younger. An emotionally unstable juvenile is 2.8 times more likely to be a
chronic serious assailant than one who is emotionally stable.

The table below calculates the probabilities for a 16 year old emotionally unstable
juvenile falling in different race and risk level categories. For example, the probability of
a 16 year old emotionally disturbed White youth with medium risk level being a chronic
serious assailant is 8%.

Table 3
Chronic Assault Probabilities
RISK LEVEL
RACE

LOW MEDIUM HIGH

WHITE

.04

.08

.06

AFRICAN AMERICAN

.10

.19

.15

HISPANIC

.06

.13

.10

ALL OTHERS

.07

.14

.11

The objective of this query was to determine the characteristics of a chronic serious
assailant which together account for 50% or more chance of being a chronic serious
assailant. As table 3 indicates, none of the probabilities calculated there are even
remotely close to the bench-mark. A possible explanation is that none of the factors
considered account for a substantive influence on the outcome variable: presence or

23

absence of chronic serious assault. A distinction needs to be made between statistically
significant factors and substantively significant factors. While all the factors considered
in the study are statistically significant, none of them individually, nor all of them
collectively are substantively significant. Potentially significant factors of substantive
value are conspicuous by their absence in the study. Studies on recidivism conducted by
the Research and Development Section have been more or less successful in isolating
substantively and statistically significant factors that account for recidivism8 suggesting
that the data elements contained in Youthbase while being good predictors of recidivism
are insufficient predictors of assaults.

3.
Survey of Superintendents, Housing Unit Managers and Security
Captains
ADJC Procedure 1190.02, Incident Debriefing, requires debriefing sessions be held after
staff or juvenile assaults:

8

I.

…in order to constructively review the circumstances and response to an
incident, analyze problem areas (if any) facilitate team interaction,
identify immediate corrective action (as necessary) and review other
options that were possibly available in resolving the incident. Debriefing
sessions shall be informal and informational in nature.

II4a.

By the fifth day of the month, the Superintendent shall forward to the
Assistant Director of Secure Schools and the Security Administrator a
report detailing all action taken as a result of the previous month’s
debriefings. This report shall identify any trends or areas requiring
immediate attention and list, as appropriate, any corrective action taken,
any commendations, any recommended revisions or additions to existing
policies and procedures.

See Outcome Evaluation: Fifth Annual Report, January 2, 2002, p.43.

24

All 4 Superintendents, 30 ADJC Housing Unit Managers and 4 Security Captains were
sent a questionnaire in January of 2004 regarding the debriefings held at their facilities
during the previous 90 days. We received responses from 14 Housing Unit Managers and
2 Captains.

Only 11 of the Housing Unit Managers reported they held incident debriefings within the
previous 90 days. As shown in Table 4, the debriefings that were held reportedly
complied with almost all of the debriefing requirements.
Table 4
Reported Compliance with 1190.02 Debriefing Requirements
Requirement
Percent
81.8%
Debriefings are held within 5 days of an assault
100%
Review of circumstances of the incident are discussed
81.8%
Effect of the incident on staff, juveniles, etc. were discussed
100%
Cause of potential cause of the incident was reviewed
72.7%
Security Captain was chair of the debriefing
100%
Potential or known deficiencies in operational procedures discussed
Neither of the two Security Captains who returned our questionnaire reported that they
had conducted incident debriefings after a juvenile or staff assault within the past 90
days. One Captain noted “unit managers are responsible to do debriefings and complete
the Youthbase form …”

In talking to ADJC staff we learned that some consider a follow-up to be the same as a
debriefing, and this confusion seemed to be more of an issue for juvenile assaults than
staff assaults. In fact, we found that 81.1% of the time a juvenile assault occurred, no
debriefing was done, however, follow-ups were done 84% of the time no debriefing was

25

done. In addition, we found that 52.7% of the time a staff assault occurred, no debriefing
was done, however, follow-ups were done 81.5% of the time no debriefing was done.

To help us understand how well the ADJC procedural requirement on assaults compared
to similar organizations we sent requests for assistance to members of both the Council of
Juvenile Correctional Administrators (CJCA) and National Institute of Corrections (NIC)
listservs. We received responses from 8 different organizations9. They provided us with
copies of the incident reporting systems and suggestions on what to look for in a good
system. In addition, the Arizona Department of Corrections (ADC) provided us with a
copy of their restricted access Department Order 706, Incident Management System.

An analysis of the different incident/assault reporting and debriefing systems revealed
that the design of our system compared favorably to others. The lack of a full response to
the internal questionnaire that we sent to secure school managers precludes us from
reaching definitive statements regarding how well our system is currently operating. An
audit of randomly selected assaults which occurred during a recent time period could
provide more definitive management information on how well ADJC staff are complying
with assault debriefing requirements.

4.

Suggestions

9

Missouri Department of Social Services, Division of Youth Services; Virginia Department of Juvenile
Justice; Montana Department of Corrections, Peoria, Illinois; South Carolina Department of Juvenile
Justice; American Correctional Association, Prince William Adult Detention Center, Virginia; Lane County
Sheriff’s Office, Oregon.

26

On March 15, 2004, R/D staff met and briefed members of the committee on the findings
obtained to date. The committee provided suggestions for additional data collection
efforts and they also reviewed 22 recommendations submitted by ADJC staff and gleaned
from the relevant literature. The following three suggestions for addressing this important
dilemma were supported by the committee.
1.
Develop a better behavior management program for assaultive and nonassaultive youth. This should include rewards for appropriate behavior and
consequences for inappropriate behavior.
Mr. Hillyard announced that soon his Division would be unveiling a new behavior
management system, and the committee supported that new system. Mr. Veloz suggested
that the new system consider establishing positive exercise-based incentives e.g., play
basketball with others and negative exercise disincentives e.g., running laps, to promote
better behavior management. One of the respondents to our national request for
assistance who was from an adult correctional agency confirmed the importance of
establishing an effective behavior management program with both incentives and
disincentives. Implementation of the new ADJC behavior management system might
benefit from a careful monitoring of assault data to check its affect upon this problem.

2.
Fully implement the existing ADJC procedural requirement to conduct
assault debriefings.
A recommendation was made by ADJC staff that ADJC more fully comply with existing
procedural requirements to hold debriefings after assaults occur. To avoid unnecessary
work and to focus effort in important issues, it might be wise to modify the procedure to
require debriefings only after serious assaults. Mr. Joseph Vattilana from the Delaware
Department of Services for Children, Youth and Families provided us with a suggestion

27

on how the ADJC might provide additional structure to the debriefings. Mr. Vattilana has
proposed a formal Root Cause Analysis process which might prove beneficial to ADJC
managers to employ after serious assaults. Based in part upon the Total Quality
Management movement (Scholtes, 1988), the Root Cause Analysis material provided by
Mr. Vattilana might provide better guidance to secure school staff as they work through
the issues relative to serious assault than what is currently prescribed by ADJC
procedure.

3.
Pursue Dr. Wortley’s recommendation to reduce assaults by providing more
structure and routines for times and places where assaults occur.
Wortley notes for adult prisons that:
Most prison programmes do not have the explicit objective of facilitating control
but are usually justified in terms of their presumed rehabilitative or therapeutic
effects. However, at a situation level prison programmes may assist in the control
task because of their time-structuring properties and their imposition of routine
on prisoners. (Wortley, 2002, p.95)
Our research found that more assaults with injuries occurred than expected in the
following locations: lounges, dayrooms, hallways, or grounds. Wortley’s research
suggests that ADJC carefully consider the nature of juvenile activities while at these
locations and consider if there might be some ways to provide additional structure and
routine for the juveniles which should then, in turn, lead to a reduction in assaults.

5.

Conclusions

It is clear from the research conducted on this project that ADJC is facing a serious
dilemma with assaults, however, our research has uncovered some hopeful clues on how
to begin to address this dilemma. Our research found patterns to the assaults relative to

28

both where and when they were most likely to occur. We found that AMS had fewer
assaults than would be expected and CMS and ENC had more assaults than would be
expected. We found statistical support for the contentions that larger housing units and
housing units with higher staff vacancy rates had more assaults. We also identified four
factors that were statistically significant predictors of becoming a chronic serious
assailant. We found that more serious assaults occurred at BCS and CMS lounges,
hallways or grounds than would be expected. The serious assaults that occurred in
lounges, etc. commonly occurred in the early evenings. We found that the commission of
an assault within the first month of commitment represented a significant indicator of a
juvenile’s future likelihood of becoming a chronic, serious assailant.

Our research found that we have a procedure that proscribes debriefings be done after
assaults occur. We compared our procedure to similar procedures provided by other
agencies and we found that our procedural requirements compare favorably.
Unfortunately, a survey we conducted for this project indicated that the procedure may
not be universally followed. Only one-third of the ADJC housing unit managers reported
that they have done debriefings over the last 90 days. We were unable to obtain
information from Superintendents regarding their required assault reports to the Assistant
Director.
Three suggestions have been proffered by this research:
•

Develop a better behavior management program for assaultive and non-assaultive
juveniles.

•

Fully implement the existing procedural requirement to conduct assault
debriefings.

29

•

Consider adopting Dr. Wortley’s recommendations to provide more structure and
routines for times and places where assaults occur.

If approved and implemented, the three suggestions should help ADJC reduce the
number of assaults within its secure care facilities. All three suggestions require focused
management attention within the institutions to succeed. As one author stated with regard
to adult correctional facilities:
If most prisons have failed, it is because they have been ill-managed, undermanaged, or not managed at all…where prison managers effect a strong
administrative regime…serious disorder are less frequent, meaningful treatment
programs more plentiful and recidivism rates less startling. (Dilulio, 1987, 7)

30

APPENDIX

31

FLOW CHART OF ASSAULT REPORTING REQUIREMENT

Assault
Occurs

Initiate
SIR

SIR Approved
by Assistant
Director,
Operations

All involved
staff fill out IR
ASAP or
before end of
shift

Goes to
YPS for
review

IR given to
Dispatch to
assign
Incident
control #
Graveyard enters
IR's into
Youthbase by the
end of shift

Filed in
Juvenile
Master File

32

Juvenile is sent
to Separation
or Health Unit

Staff have
90 minutes
to turn in IR
to YPS

Copy to
Separation
File

FOLLOW-UP REQUIREMENTS
IR's shall be
reviewed within
7 days of the
completion of
the IR by:

Assault Occurs

IR's shall be
reviewed in the
Daily Incident
Report Recap
each morning by:

Superintendent
* review trends
*documention & notification
are done for SIR and
Debriefings
*investigative follow-up

Health Unit Supervisor:

YPS and caseworker:

Security Supevisor:

*ensure that injury incidents
match IR
*ensure mechanical restraints
are documented in IR
*ensure appropriate follow-up

*to coordinate the necessary
needed responses
*respond to IR affecting
juvenile program goals
*to ensure follow-up is
documented in the IR

*reveiw IR for Control #
*Use of Force
*ensure narrative describing
incidents matches IR

Asst Supt
*review IR w/in administrative
area to identify trends
*May issue directions regarding
follow-up to be documented in
IR

Debriefing
should be
conducted for
serious IR's
within 5 days

Issues to be discussed at the
debriefing include but are not
limited to:
* a review of the circumstances of
the incident
* the effect of the incident on staff,
juveniles, etc
* the cause or potential causes of
the incident;
* identification of known or potential
deficiencies in operational
procedures and/or practices
* Need for immediate corrective
action and steps taken to
accomplish this, as necessary

33

Security Captain shall act as
Chairperson and complete a
debriefing form

Superintendent of each facility
shall report by the 5th of each
month a detailing off all action
taken on previous month
debriefings to the AD,
Operations

References:
Clarke, Ronald, 1980, Absconding from Residential Institutions for Young Offenders,
Out of School: Modern Perspective in Truancy and School Refusal, Chichester: John
Wiley and Sons.
Dilulio, John, 1987, Governing Prisons: A Comparative Study of Correctional
Management, New York: The Free Press.
Gaes, Gerald and William McGuire, 1985, Prison Violence: The Contribution of
Crowding Versus Other Determinants of Prison Assault Rates, Journal of Research in
Crime and Delinquency, v22., no. 1.
MacDonald, John, 1999, Violence and Drug Use in Juvenile Institutions, Journal of
Criminal Justice, v27., no. 1.
Maitland, Angela and Richard Sluder, 1998, Victimization and Youthful Prison Inmates:
An Empirical Analysis, The Prison Journal, v78., no. 1.
Scholtes, Peter, 1988, The Team Handbook: How to Use Teams to Improve Quality,
Joiner.
Wooldredge, John, 1994, Inmate Crime and Victimization in a Southwestern Correctional
Facility, Journal of Criminal Justice, v22., no. 4.
Wortley, Richard, 2002, Situational Prison Control: Crime Prevention in Correctional
Institutions, Cambridge, U.K.: Cambridge University Press.

34