Disparity Documentation Manual - Justice Research and Statistics

Disparity Documentation Manual - Justice Research and Statistics
Disparity Index Computation Database User’s Manual
Prepared for:
Justice Research and Statistics Association
Prepared by:
The Illinois Criminal Justice Information Authority
Robert Bauer, Senior Research Analyst
Ernst Melchior, Senior Sys tems Analyst
Phillip Stevenson, Senior Research Analyst
March 2003
This project was supported by grant #98-JN-FX-0112 (S-2), awarded to the Illinois Criminal
Justice Information Authority by the Office of Juvenile Justice and Delinquency Prevention,
through the Justice Research and Statistics Association. Points of view or opinions contained
within this document are those of the authors and do not necessarily represent the official
position or policies of the Justice Research and Statistics Association or the Illinois Criminal
Justice Information Authority.
Introduction
The Disparity Index Computation Database allows users to assess the level of racial/ethnic
disparity at various stages of the criminal justice system process (i.e., disproportionate
minority representation). This database can be used to determine whether a disparity exists
and, in most cases, it can help determine where in the process disparity is being introduced.
The database is a single-user Microsoft Access system designed to handle one interpretation
of one system process. The database can handle multiple categorizations of cases within that
process, and the interpretation can include up to 250 system process steps.
Although the database can be used to assess disparity in any system, the examples used in
this manual demonstrate the functionality of the database as it pertains to the assessment of
racial disparity in the juvenile justice system.
The Basics
This manual is intended to be accessible to everyone, including those who have not tried to
assess disproportionate minority representation (DMR) before and those who have never
before worked with MS Access. For those who have had some exposure to DMR or
Microsoft Access, there are sections of this manual that will appear to be overly simplistic.
For others, the manual may be complex at first. Given that assessing DMR is a complex
problem, we decided to confront the complexity, but to try to keep that complexity in the
manual and not in the database. Once you get through this manual, you will find the database
easy to use.
Two things may help you have an easier time as you progress through this manual.
1)
2)
Make a copy of the database as a “tutorial” database. 1 Go to the forms as you
read about them in this documentation, so you can more easily see what is being
discussed.
Photocopy the flow chart on page 2. It will be easier to follow the examples given
in this documentation as we use Illinois’ juvenile justice system process as the
example.
A Flow Chart of Illinois’ Juvenile Justice System Process
To help the user learn how to use the database to assess DMR in a justice system process,
Illinois’ juvenile justice system process is used as an example. The flow chart in Figure 1
shows some of the paths a minor’s case can take in a juvenile justice system.
1
The database file can be copied in the same way as any Microsoft file. We also recommend that you make a
copy of the original for backup purposes. Further, if you wish to deal with multiple systems or multiple
interpretations of the same system, you can make a copy of the database file and use the copy to create the new
interpretation.
Disparity Index Computation Database User’s Manual
1
Figure 1: Example Flow of a Juvenile Justice System2
Juvenile Population
Arrests
Override up
Screened for
detention
Override
down
No override
Not screened
for detention
(i.e., not
detained)
Case Referred to
State’s Attorney
Detained
Secure
detention
Station Adjustments
Transfer to adult court
Not detained
Probation
adjustment
Non-secure
detention
Petition filed
Found
delinquent
Charges
dropped
Found non-delinquent
Detention hearing
IDOC
Secure
detention
Non-secure
detention
Not detained
Probation
Temporary
detention
Continued
under
supervision
2
As juvenile justice system professionals will notice, the flow chart used in this document does not capture all
the possible paths of the juvenile justice system process. This flow chart was created with the specific purpose
of aiding the user in understanding the functions of the database.
Disparity Index Computation Database User’s Manual
2
The Decisions the User Must Make
The user decides what events in the system to track (i.e., grid points) and what variables will
be used (e.g., race, ethnicity, sex, etc.).
Grid Points
These are the decision points and outcomes at which the user wants to assess levels of
disparity. For example, an arrest (a decision point) can result in a station adjustment (an
outcome) or a referral to court (both an outcome and a decision point). In addition to the
description of the grid point, the user needs to number the grid points. Although it is not
required, it makes sense to number the grid points in a manner that is consistent with the flow
of the system.
Variables
The database was created with three variables already programmed for the user: race,
ethnicity, and sex. Later in the manual, the procedure to be followed for creating new
variables, including interactions between multiple variables, will be discussed.
Reports
The database is pre-programmed to give the user an option between two reports, a “Main
Report” and a “Point-to-Point Report.” These reports show the representation and disparity
indices at every point in the system (Main Report) or between any two points selected by the
user (Point-to-Point Report). Both reports show both indices of disparity in the system; one
based on the population at large (Representation Index) and the other based on the population
at the previous point in the process (Disparity Index).
Main Report
This report will automatically provide the user with the Representation Index (RI) and the
Disparity Index (DI) for every decision point and outcome in the system, by each variable
selected by the user (i.e., race, ethnicity, sex, etc.).
Point-to-Point Report
This report limits its analysis to any two points in your system. For example, the user may
want to know if disparity is evident for the outcomes of the cases that reached the
“delinquency petition filed” stage. This is useful if disparity is identified at the “delinquency
petition filed” stage and the user wants to know if further disparities are introduced in the
system after delinquency petitions are filed. In this example, the origin point is “delinquency
petition filed” point and the destination point is any point that comes after the “delinquency
petition filed” point. This report allows users to narrow the focus of a report to the points of
the system in which they are most interested.
Disparity Index Computation Database User’s Manual
3
Using the Database
Before using the database, users must have Microsoft Access installed on their computer or
network. Fortunately, using this database does not require the user to know very much about
Microsoft Access.
To open the database, double click on the icon (if loaded to your desktop) and you will see
the Main Form. Many different forms will open for you as you work with the database.
Figure 2: Main Form of the Database
The database is designed to make the process of entering data and assessing disparity in a
system quick and easy. However, getting to the point of data entry and assessing disparity
requires considerable thought as to the flow of the system and subsequent availability of data
for each decision point in the system.
It is strongly recommended that the user create a flow chart to help understand how cases
move through the system. Of specific importance are the points in the system where
decisions are made (i.e., decision points) to keep the juvenile in the system, versus diverting
the juvenile out of the system, and where less punitive and more punitive decisions are made.
These are the decision points where disparity is of most concern. Put another way, it is
important for juvenile justice system personnel to understand where they are treating youth
differently, and why that is the case. This database allows the user to answer the first of these
two important questions.
Disparity Index Computation Database User’s Manual
4
Once you have an image of the process (i.e., a flow chart), you can begin to build the
database that will allow you to enter the data for each of the decision points identified in the
flow chart. The database has been programmed to look at all your grid points and variables
and creates all the records you need for creating a new data set.
If the user discovers errors in the flow chart or specification of the decision points and issues
the “BUILD IT!” instruction, the user will have built the data set improperly and the program
will produce incorrect results. The system provides you a way to get out of this trap quickly.
When the user realizes this, the user can hit the “DELETE DATA” button, pick the data set
to delete, and delete it. Then, the user can fix whatever was wrong, hit “BUILD IT” again,
and the problem will be gone. This loss of time will be about two minutes. However, if you
spent an hour’s worth of time entering data AFTER the undesirable “BUILD IT” command,
this data entry will also need to be repeated. Therefore you should try to get your system
entered correctly before hitting the “BUILD IT” button.
Understanding the process and getting it defined in a computer- friendly way is the hard part.
Once this is done, entering the data into the computer is not difficult.
Recreating your juvenile justice system
In the example provided earlier in this manual, a flow chart of the juvenile justice system
process was created to guide the identification of grid points (i.e. decision points and possible
outcomes). The database is preset to include general population and arrests as points at which
data must be entered in order to assess disparity in a criminal justice system. From the flow
chart, users can identify the points at which they will enter data. More importantly, the flow
chart can be used to guide the creation of grid point connections; essentially the paths that a
case can follow through the system. Any link between two decision points or outcomes is a
grid point connection.
Identifying the variables by which disparity will be identified
Already included in the database are the variables: race, ethnicity, and sex. These variables
will allow the user to assess disparity by these characteristics. Of course, the user chooses the
variables of interest, and the flexibility of this database allows the user to identify other
variables (e.g., participation in an intervention program) by which disparity can be assessed.
Entering the Grid Points
After you have identified the decision points in the process, the outcomes of the process, and
the connections between them, you are ready to begin constructing your database. Click the
“Grid Points” button on the main form to begin. The form shown in Figure 3 will appear.
Disparity Index Computation Database User’s Manual
5
Figure 3: Grid Point Entry Form
The two fields are the number of the grid point and the description of it. Using your flow
chart as a guide, the user should enter in the grid point number and grid point description for
all the decision points and outcomes in this table. Although the order in which the grid points
are entered isn’t important (later you will enter the connections between grid points),
following a logical order is recommended.
Grid point 1 is always population – you aren’t supposed to change that. Grid point 2 is
“Arrests” – you should not change the number of grid point 2, but if the first event to occur to
your population is something other than Arrests, you can change the title as you choose.
Entering the Grid Connections
Grid connections are the links between two grid points in the system. In our example flow
chart, the possible outcomes stemming from the filing of a delinquency petition are the youth
being found delinquent or the youth being found non-delinquent. By entering the connections
(represented as arrows on our flow chart) you allow the computer to determine how the grid
points are connected. The connection between Grid Point 1 and Grid Point 2 has already been
entered for you.
After clicking the “Grid Connections” button at the bottom of the left column of the main
screen, you will see the screen shown in Figure 4.
Disparity Index Computation Database User’s Manual
6
Figure 4: Example Grid Connections Screen
For every arrow on your flow chart, you will type the number of the origin point in the
“origin” field, and the number of the destination point in the in the “destination” field. The
gray shaded boxes to the right of the screen will tell you the names of the fields you have
selected. It is important to remember that if you have set up the flow chart as we recommend
with lower numbers flowing into higher numbers, every destination point must have a
number higher than its origin point.
Entering the variables
Although the database is programmed with race, ethnicity, and sex already as pre-established
variables, the user has the ability to enter other variables of interest. For example, if the user
is interested in disparity by age or level of income, the user can create these variables and the
database will calculate indices based on these variables. For example, in Cook County, IL,
there are important differences in the populations of Chicago and the surrounding suburbs. In
our assessment of racial disparity in Cook County’s juvenile justice system, we not only
looked at racial disparity, but also racial disparity by sex and location.
To enter the names of the variables you create, click the “Variables” button on the main page.
The screen shown in Figure 5 will appear.
Disparity Index Computation Database User’s Manual
7
Figure 5: Variable Name Screen
As can be seen, there are spaces for you to enter up to six variable names. To use a variable
in the analysis, the user needs to type in its name and click on the “In Use” check box to
indicate that it is a variable that will be used in the analysis.
This form will not allow the user to delete or add any records. Users will always have
exactly six variable s available, although all of them do not have to be used. If the user would
like to do analysis of disparity by more than the three new variables for which there is room
(remember that race, sex, and ethnicity are three variables that are preprogrammed), any of
the preprogrammed variables can be deleted and new values added. For example, if the user
wanted to conduct disparity analysis on race, ethnicity, age, income, felony/misdemeanor,
and location, the sex variable can be deleted to make room for one of the other variables.
Another option would be to copy the master database and conduct separate analyses in
separate databases. Either of these options will allow the user to succeed in the analysis.
Entering values for the variables
Once the variables have been entered, for any new variable(s) the user must identify the
values to be used (e.g., for the Sex variable we have male and female. For your new
variables, you need to tell it what the choices are.). To get to the form in which values are to
be entered, the user should click on the button titled “Values” on the main form of the
database. The user can then enter the numbers corresponding to the variables that have been
created.
Disparity Index Computation Database User’s Manual
8
Figure 6: Example Form for entering variable values
Of the four columns on the form, the user is required to enter only two. In the first column
the user inputs the number of the variable of interest. For each variable value that exists the
user must enter the variable number. For example, if “Participant in Training Program” is a
variable the user created and it has two values (e.g., Participant, Non-Participant), the number
corresponding to the sex variable must be entered twice. The database will use the second
column to display the name of the variable the user has entered. The third column is an
internal computer code. In the fourth column the user enters the text names of the variables.
You can also modify the pre-programmed variables on this form. For example, if you have
some data in which race is not supplied, you might need to add a “Not Known” value to the
race variable.
Combining variables
The database also allows users to identify combinations of variables to be used in the
analysis. For example, in addition to assessing disparity by race and sex, the user may want
to know the disparity indices for interactions between these terms. In our report for Cook
County, we not only assessed disparity by race, sex, and location, we also developed
disparity indices for the combination of these variables (e.g., female African-American, male
African-American, female White, male White, etc.).
By clicking on the button titled “Interactions” on the main form, the user can review the
variables selected for analysis and create new variables that are combinations of existing
Disparity Index Computation Database User’s Manual
9
variables. In the example screen in Figure 7, across the top of the form are the variables that
have been identified as being used for analysis. To create a new variable that is a
combination of two or more pre-existing variables, users need to fill in the new variable
name in the first column and check the boxes in the row pertaining to the new variable that is
under the pre-existing variables. In Figure 7, a race/sex variable has been created.
Figure 7: Example Interactions Screen
Building the Data set
After creating your system, naming the decision points and outcomes, and defining the
variables of interest, the database will create a form into which the user can enter the data.
The first step is to name the data set. Often different data sets will represent different years,
of data. Thus, if you have data from 2000 and 2001, you can get reports for both years
without needing to re-enter you flow chart and variable choices. You will just create one data
set for 2000, and another for 2001. When 2002 data are available to you, you will be able to
create another data set for that. These are the steps you will follow. Click the button on the
main form labeled, “Add Data Set.” Figure 8 shows the form that will appear after clicking
“Add Data Set.”
Disparity Index Computation Database User’s Manual
10
Figure 8: “Add Data Set” Form
Note that the gray shaded “Status” area indicates that the data set has not yet been created.
Once you have named the data set, the user can close this form. After the data set has been
named, the next step is to instruct the database to build the data set.
After clicking the button titled “Build Data Set,” the screen in Figure 9 will appear.
Figure 9: “Build Data Set” Screen
Disparity Index Computation Database User’s Manual
11
The user should use the drop down box to select the data set to be built. The next step is to
click on the “Build It!” button. The data set will be built and the user will be returned to the
main screen.
To enter data into a data set, users should click on the “Enter Data” button and a screen
similar to that in Figure 10 will appear. On this screen all the user needs to do is select the
data set into which the user wants to add data. If the data set has not been built yet, the name
of the data set will not appear. Once users have selected the data set with which to work, they
should click “Edit this Data Set” to make the data entry form appear (see Figure 11 for an
example).
Figure 10: “Enter Data” Screen
Disparity Index Computation Database User’s Manual
12
Figure 11: Data Entry Form Screen
At this point, it is simply a matter of entering the numbers for each variable, at each decision
point or outcome. As can be seen in Figure 11, the program has created a data entry form for
all the variables and their categories, decision points, and outcomes. On this screen, the only
field that can be edited is the “Persons” field. If at this point users notice that they built the
data set incorrectly, they will have to return to the screen(s) where the error occurred, fix the
error, and build a new data set.
Creating Reports
As discussed earlier, users are able to generate two different reports from their data, a “Main
Report” and a “Starting Point Report.” The “Main Report” will display the Representation
Index and Disparity Index for all decision points and outcomes in the user’s system. The
“Starting Point Report” displays the Representation Index and Disparity Index for a
destination point (i.e., the point at which disparity is to be assessed) based on the user’s
selection of an origin point (i.e., the starting point at which the Representation Index and
Disparity Index will be calculated).
By clicking on the “Reports” button, a report specification screen will appear (see Figure 12).
This screen asks users to identify the data set from which the reports are to be generated.
After identifying the data set, users have a choice between the “Main Report” and the “
Starting Point Report.” After clicking on the button titled “Main Report,” the database will
calculate the Representation Index and Disparity Index for all the points in the system and
display them on screen. If the user wants a “Starting Point Report,” where a grid point is used
as the starting population, the user will select the starting node. Then the user can click the
Disparity Index Computation Database User’s Manual
13
button for “All Following Nodes” or can select which following nodes to analyze by clicking
on the “Specified Following Nodes” button.
For the “Main Report” and the “All Following Nodes” report, the user can select “Display a
Query” instead of as “Display a Report.” This is done in the radio control box at the top of
the form. If “Display a Query” is chosen, the results will be fo rmatted for downloading to a
spreadsheet or other analysis package. “Display a Report” is better for viewing on the screen
or printing.
Figure 12: Report Specification Screen
Example Reports
The “Main Report” generated by the database shows the Representation Index and the
Disparity Index for each point in the system by each variable selected for analysis. Figure 13
is an example of a “Main Report.” The “Main Report” and the “Point-to-Point Report” are
formatted similarly. These reports are straightforward, giving the user the variables used in
the report, the population, the number individuals at the previous point in the system, the
number of individuals at each point in the system, and the Representation Index and
Disparity Index for each variable.
Disparity Index Computation Database User’s Manual
14
Figure 13: Example of Main Report
Disparity Index DataBase Main Report
Test Data
Population
Arrests
Persons at
Previous
Point(s)
Persons at
This Grid
Point
Disparity By
Population
(DI)
Disparity
From
Previous
Point(s)
(RI)
Race
Majority
400,000
400,000
3,333
0.73
0.73
Minority
100,000
100,000
2,357
2.07
2.07
Summary for Arrests (2 detail records)
500,000
500,000
5,690
Screened
Race
Majority
400,000
3,333
3,193
0.73
0.99
Minority
100,000
2,357
2,307
2.1
1.01
5,690
5,500
Summary for Screened (2 detail records)
500,000
Not Screened
Race
Majority
400,000
3,333
140
0.92
1.26
Minority
100,000
2,357
50
1.32
0.64
Summary for Not Screened (2 detail records)
500,000
5,690
190
Detained
Race
Majority
400,000
3,193
2,927
0.71
0.98
Minority
100,000
2,307
2,205
2.15
1.02
Summary for Detained (2 detail records)
500,000
5,500
5,132
Not Detained
Race
Majority
400,000
3,193
266
0.9
1.25
Minority
100,000
2,307
102
1.39
0.66
Summary for Not Detained (2 detail records)
500,000
5,500
368
Page 1 of 8
Disparity Index Computation Database User’s Manual
15
Interpreting Representation and Disparity Indices
Generally speaking, the interpretation of the Representation Index and Disparity Index are
very straightforward. Any Representation Index or Disparity Index over 1.0 indicates overrepresentation. Any Representation Index or Disparity Index under 1.0 indicates underrepresentation. A Representation Index or Disparity Index of 1.0 indicates no disparity. The
important point to remember is that the Representation Index is based on the general
population. For example, the formula used to calculate the Representation Index based on
minority group status is:
Representation Index (RI) =Percent of Minority Group at Juvenile Justice Decision Point
Percent of Minority Group in Juvenile Population
The Disparity Index is based on the population “at risk” for further participation in the
system. In other words, the Disparity Index is adjusted for the composition of the population
at the point immediately preceding the point being assessed. For example, the formula used
to calculate the Disparity Index based on minority group status is:
Disparity Index (DI)=Percent of Minority Group at Juvenile Justice Decision Point
Percent of Minority Group at Previous Juvenile Justice Decision Point
In this way, the Disparity Index can tell you the degree to which disparity in the system is
added to, or negated by each point in the system.
Still, the representation and disparity indices are subject to a great deal of interpretation. The
database can tell you what the indices are, but cannot tell why they are what they are. For
example, the database cannot tell the user whether the arrests of a disproportionate proportion
of people from a group are appropriate. If criminal offenders disproportionately come from
Group X, one would expect that they would also disproportionately be arrested. In addition,
the database cannot determine whether the disparity in the distribution of station adjustments
is a good thing or not. If Group X offenders are more likely to be arrested for less serious
crimes than Group Y offenders, one may hope that Group X offenders are diverted out of the
system at the arrest stage at a higher rate than Group Y offenders. In other words, although
this database can identify disparity in the justice system, it does not control for frequency of
crime, seriousness of crime, past offending history, etc. Remember, this database is to be
used to assess disparity, not explain disparity. Knowing if disparity exists is just the first step
to addressing Disproportionate Minority Representation.
Disparity Index Computation Database User’s Manual
16
120 S. Riverside Plaza, Suite 1016
Chicago, Illinois 60606-3997
Telephone: 312-793-8550
TDD: 312-793-4170
Fax: 312-793-8422
http://www.icjia.state.il.us
Illinois Criminal Justice Information Authority
Rod R. Blagojevich, Governor
Candice M. Kane, Executive Director
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement