Report on Terminology and Data Analysis

advertisement
Information Sharing and Integration Committee
Report on Terminology and Data Analysis
OVERVIEW
The Information Sharing and Integration Committee of the CJCC was charged with
developing a report that could be viewed by the CJCC on a regular basis in order to
guide policy discussions and identify trends. To meet this charge, members from
the Sheriff, Trial Court, District Court, Community Corrections and Support Services
met several times to determine how criminal justice data could be reported
expeditiously and accurately to the CJCC. This workgroup met several times and
developed this report, which details the proposed methodology and definitions of
key terms. The following summarizes key elements of the report and discussions.






Through partnership between Community Corrections and the Sheriff, an
accurate data set has been established. It is a point in time analysis from
March, 2005 (592 total inmates on over 1100 charges) and is a combination
of jail and court data – the first data set of its kind for Washtenaw County.
Many inmates have multiple charges which are in multiple stages in the
criminal justice process. Categorizing such individuals has engendered much
discussion and this document has recommendations about how the CJCC may
wish to do so.
Lack of consistency around definition of terms has been a barrier. This
document attempts to begin to develop consensus around the terminology.
This document has been developed to be the beginning of a conversation
about definitions. Through deliberations of the CJCC, enhancements and
changes to definitions are expected.
Ongoing data collection and analysis of this nature does not fall neatly into
the jurisdiction any discrete CJCC agency. As such, a long range plan will
require staff resources from several CJCC members.
As policy discussions ensue, methods in which data is gathered may be
identified. These issues may require training resources within areas managed
by CJCC members.
INTRODUCTION
Jail utilization data, like all data, has little intrinsic value until it is interpreted into
meaningful information. For many years, Washtenaw County has had the
resources to piece together fragments of utilization data, but the process was
tedious, labor intensive, cost prohibitive, and difficult to accomplish because the task
required access to, and familiarity with, the different data systems of the courts and
the jail. Each data system had its strengths, weaknesses and navigational
idiosyncrasies. Recent collaborative efforts between the jail and the county’s judicial
and executive branch have made the prospect of converting jail utilization data into
information more attainable. The installation of the search engine utility connected
to the AS400 jail data system at Community Corrections has significantly reduced
the time formerly required to assemble inmate data in a roster format. As a result,
a jail admission cohort study from March 2005 is underway as a pilot for a planned
ongoing stratified sample of jail bookings.
The pilot study involves a query of approximately 30 data elements from the AS 400
jail record system, cross referenced with approximately 6 criminal record fields from
the eNACT court computer. The data elements include demographic information
(name, age, sex, race, employment status and city of residence), booking and
release dates and times; legal status and charge level at admission and release;
offense description, bond amount and bond type; arresting agency and charge
information; dates and dispositions for court appearances that occur while in
custody; information regarding failure to appear or probation violations and other
failures to comply with sentence prior to jail admission; and, release reasons. The
data has been exported to an excel spreadsheet, transferred to an access
spreadsheet for cross tabulation, then returned to excel and will be manipulated
primarily through pivot table reports. Basic spreadsheet formulae allow the
calculation of lengths of stay within hundredths of a day across a number of
different variables. It is expected that these projects will yield important new
information to guide policy making for the Criminal Justice Collaborative Council. A
glossary of these data elements is included in the appendices.
The first obstacle in the process of converting data to information is to define the
main data elements in the study project. Nationwide, there is an astonishing lack of
uniformity of definition, even among justice professionals, for terms they use
everyday. For example, not every state has the same definition of a felony. The
commonly used textbook definition, “an offense punishable for more than a year”,
does not always apply in Michigan, where certain crimes punishable by up to two
years in prison are called “High Court Misdemeanors”. If a prison diversion program
wanted to use jail utilization information to identify unsentenced inmates for
inclusion in programming, then classifying an inmate as a (High Court)
misdemeanant or felon would not be a trivial issue.
The process of defining data elements is extremely complex, but not
insurmountable. For one, there is a fair amount of consistency in terms of which
variables most jail utilization studies attempt to measure, and the sampling methods
they employ. The Department of Justice, Bureau of Justice Statistics (BJS) is a
typical example. The BJS jail studies measure the lengths of stay of inmates
according to their legal status (sentenced or unsentenced) and charge level
(felony or misdemeanor) at booking and release, with special attention given to the
recidivism rates of violent offenders. In view of the fact that many offenders
enter jail with multiple charges at various stages of the judicial process, the BJS
designates a “single most serious offense” to allow analysis of aggregate data
representing the main reason that each offender was in custody.
All five of the measurements crucial to most jail utilization analysis (legal status,
charge level, violent offender, recidivism and most serious offense) can be
interpreted differently, based on the defensible assumptions or legitimate intended
uses of those collecting and interpreting the data.
DEFINING “SINGLE MOST SERIOUS OFFENSE
Issue
The definition of “single most serious offense” is illustrative of the complexity of the
definition issue. Common practice dictates that charge level is the first determinant
of crime severity, and felonies are assumed to be more severe offenses than
misdemeanors. In reality, this does not necessarily mean that all felons are more
severe offenders than all misdemeanants, because charge level does not take into
account the very important risk factor of an inmate’s prior criminal history. Even so,
if the jail utilization information was being used in the context of public safety, such
as which inmates should be released early due to overcrowding, then it could be
reasonably argued that a misdemeanant domestic violence offender may be
objectively more serious an offender than a drug abuser convicted of felony drug
possession.
An alternative measure to distinguish “most serious offender” (rather than offense)
might be to use jail classification as a guide. Inmate classification level is based on
both the legal status of the inmate and the inmate’s history of violent felony
offenses and institutional misconduct. Under the jail classification system,
unsentenced inmates are often classified higher than sentenced inmates, and while
an assaultive misdemeanor like domestic violence will not raise an inmate’s
classification level, a pending misdemeanor warrant for retail fraud will. In the
context of public safety, using jail classification to determine early release decisions
would prevent the release of, for example, a property damage misdemeanant with a
history of armed robbery, but relying on jail classification might fail to identify, or
worse, exacerbate, process issues related to the pre-trial detention of objectively
less serious offenders. In the latter case, a better definition facilitating recognition
of potential system processing improvements at the pre-trial stage could potentially
alleviate the need for early release of the very inmates that might be identified by
using classification.
There are also political and practical implications in establishing a hierarchy of
classification of offenses for statistical analysis. Victim groups or other advocacy
interests may contend that the subordination of a particular class of crimes to
another class (e.g., crimes against women, versus drug or alcohol offenses or crimes
involving firearms) is an attempt to underscore the scope and severity of a particular
social problem at the expense of another. Nonetheless, attempting to stay within
existing frameworks for collecting and reporting data (such as adopting much of the
format of the Department of Corrections’ statewide Jail Population Information
Systems, or JPIS) reports allows comparative analysis between Washtenaw and
other counties across the state.
Proposed Solution
Developing a decision tree based primarily on charge level, but incorporating
elements of state felony sentencing law, such as differentiating multiple felony
offenses on the basis of assaultiveness and crime class (severity), appears to be a
viable solution that addresses many of the concerns listed above. Misdemeanor
offenses can be differentiated for tracking purposes based specifically on violence
(e.g., stalking, domestic violence or assault), on persistently occurring offenses
(Operated While Impaired), or by statutory maximum. Exhibit 1 is a draft of a
decision tree to determine most serious offense, and it includes the outcome of the
hierarchical placement of the top 40 crimes in the study project. (As this is a draft,
it is important to note that additional sorts may be needed at level three and level
seven.)
DEFINING “VIOLENT OFFENSE”
Issue
Closely related to the issue of “most serious offense” is the definition of “violent
offense” or violent offender. There is no consensus among justice practitioners as
to what constitutes a “violent offense”. Some jurisdictions appear to interpret
“violent offense” synonymously with “dangerous offense”, thereby including a range
of negligent behaviors such as drunk driving or reckless discharge of a firearm.
There are at least four potential available references for defining “violent offense”:
The Michigan Sentencing Guidelines, The Department of Corrections Assaultive
Crime List, The FBI Uniform Crime Reports and the Michigan Court Rules.
With the Truth in Sentencing reforms of 1999, the Michigan Legislature identified
felonies considered assaultive as part of adopting a new set of sentencing guidelines
meant to send fewer, but more dangerous, offenders to prison for longer periods of
time. There are approximately 972 felony criminal offenses listed in the Michigan
Sentencing Guidelines Crime List. Many of the crimes listed are archaic (i.e.,
Seduction), redundant (i.e., 257.602a & 750.749A are both Fleeing & Eluding-First
Degree) or variations on a single offense depending on the degree of harm suffered,
amount of controlled substance involved or value of property damaged or lost (i.e.,
Placing an Explosive Causing: Death, Property Damage, Serious Impairment or
Injury). The Sentencing Guidelines lists 208 offenses considered as Crimes Against
a Person, and divides these offense into “high Severity” or “Low Severity” according
to Crime Class. A small number of low-severity Crimes Against a Person have
relatively low prison commitment rates, and one in particular,
Resisting/Obstructing/Assaulting Police, seldom results in jail or prison confinement.
The advantage of using sentencing guidelines to define assaultiveness is that they
are comprehensive, covering every felony offense, and disposition information
(including jail) is tracked statewide by the Department of Corrections and available
for comparison. The disadvantage of using sentencing guidelines to define
assaultive offenses is that they make no reference to misdemeanors and include
some felonies that local sentencing trends suggest are not considered so dangerous
to warrant even temporary removal from the community.
The Department of Corrections compiles an “Assaultive Crime List” to assist parole
officers in parolee disciplinary policy. A parolee charged with any of the offenses on
the list is subject to more severe action, up to revocation of parole. Though less
expansive than the sentencing guidelines, the problem with the list is that it is
updated and changed (often every quarter), and it sometimes includes relatively
minor offenses, such as misdemeanor trespassing. It would be impractical to
retrofit definitions every time the list changed, or to know what the previous
definitions may have been.
The FBI Uniform Crime Reports classifies violent offenses according to six
categories: Murder/non-negligent manslaughter, Rape, Other sexual assault,
Robbery, Aggravated Assault, and Other violent (i.e. Negligent manslaughter or
kidnapping). The advantage of using the FBI Crime Reports is that the focus on
violent crime appears narrower and, on the whole, more severe than the sentencing
guidelines.
The Michigan Court Rules as they relate to the setting of bail enumerate just a
handful of violent offense groups for which a judge may deny bond based on the
grave nature of the offense. The disadvantage of using the criteria spelled out in
the court rules is that it may be too restrictive to describe a variety of behaviors that
the public may consider “violent”.
Proposed Solution
It appears the benefits of using the “Crimes Against a Person” crime list in the
sentencing guidelines outweigh the shortcomings, particularly as it relates to
comparing felony jail sentencing patterns statewide. The subcommittee should
address the need for further defining the list of assaultive misdemeanors beyond
stalking, assault and domestic violence.
DEFINING “LEGAL STATUS” AND “CHARGE LEVEL”
Issue
Less problematic is the task of determining an inmate’s legal status (unsentenced or
sentenced) and charge level (misdemeanor or felony), but there are potential
problems of definition nonetheless.
For example, JPIS reports classify all
probationers lodged on probation violation warrants under the legal status
“sentenced”, despite the fact that most felony probation absconders are remanded
to jail with a high bond (or no bond) pending a hearing and adjudication. It seems
incongruent for a “sentenced” inmate to have a bond. Similarly, the legal status
category “unsentenced” or “pre-sentence” contains a number of subgroups that are
likely treated substantially different from one another in terms of bond consideration
or length of stay. An inmate arrested and brought before the bail magistrate on a
new charge is likely to be granted bond and released much faster than an inmate
arrested on a bench warrant for failing to appear at a pre-trial hearing. Though it is
appropriate to classify both of these inmates as “unsentenced”, it may be wise to
differentiate among the subgroups as a separate data element.
The inmate whose legal status and/or charge level changes while in custody
presents other problems of definition, especially when multiple charges are involved.
An unsentenced inmate who failed to post bond on a non-violent felony in Circuit
Court and failed to bond on a violent misdemeanor in District Court suddenly has the
non-violent felony pled to a non-violent misdemeanor. What portion of the days
that inmate spent in custody should be considered unsentenced as a felon?
Consider the following situation: a misdemeanant probation violator is arrested in
connection with a new felony offense. The probationer is brought to district court
and sentenced to a lengthy jail term on the misdemeanor. The former probationer
refuses to post a modest bond on the felony because it will not affect the sentence
on the violation, and posting bond will relinquish any claim s/he has to jail credit.
The probationer eventually pleads to the new felony and the felony judge “brackets”
the probationer’s jail sentence within the admission and release dates of the
misdemeanor. In this case, it is likely that the misdemeanor sentence impacted the
felony sentence, while clearly the lower charge was keeping the inmate in custody.
Proposed Solution
Once again, the complexity of the issue suggests that a matrix or decision tree is a
viable alternative to subjectively evaluating each jail record. The decision tree in
Exhibit 2 starts with the presumption that any case already sentenced is the
controlling record. This is consistent with the unsentenced inmate’s presumption of
innocence; however, the decision tree also acknowledges instances where a pending
high severity felony is the main reason the inmate is in jail. The decision tree
parallels the JPIS categories of Unsentenced-No Change in Status (meaning
admitted and released unsentenced), Sentenced-No Change in Status, and
Sentenced after Admission. This allows computation of days unsentenced and
sentenced for an inmate based on legal status at release. Following the JPIS format
allows comparison to local JPIS data for validation and comparative analysis with
statewide data.
It is important to note that it may be theoretically possible to run a jail record
through the “legal status” decision tree and “most serious offense” decision tree
with different outcomes. In such a case, the “legal status” tree should hold
precedence.
DEFINING “RECIDIVISM”
Issue
Perhaps the most elusive of all the definitions justice practitioners use regularly is
the definition of “recidivism”. In a 15 state study of felony offender recidivism, the
BJS defined recidivism as “criminal acts that resulted in the rearrest, reconviction or
return to prison with or without a new sentence during a three year period following
the prisoner’s release”. Simply put, if the prisoner returned to prison as a result of
a new arrest or technical violation, it was counted as recidivism. This implies that
an arrest to jail on a misdemeanor did not necessarily mean a parole revocation
was pursued. The range of definitions of recidivism vary from quite restrictive (only
counting new felony convictions) to very broad (counting any new arrest or jail
booking, including probation violation). Irrespective of whatever definition is used,
it is arguable that the way recidivism is measured is inherently flawed in that
recidivism does not account for changes in the scope or frequency of detected
criminal activity by an offender. In other words, an offender who commits one
public order offense after 20 months is not differentiated from an offender who
commits 3 felonies in the first 4 months after release. Both offenders have a 100%
recidivism rate, with no measure of “survival time” between offenses.
Proposed Solution
Inasmuch as the purpose of the pilot admission cohort study, and the subsequent
stratified admission sample are designed to measure jail utilization, it makes
sense that to track inmate bookings as a measure of “recidivism”, then further
subdivide those records by type of arrest or time between arrests. However, given
the importance and broad degrees of interpretation of this term, it is recommended
that the Inmate Reintegration Committee study this issue to make a
recommendation. Doing so is consistent with their charge and can help ensure a
thorough discussion and resolution.
GLOSSARY OF DATA ELEMENTS
Arraignment on the …is the first step in the felony process after an offender is
bound over
Information
from district court. The AOI is a formal reading of the charges
to the defendant that details the elements of the crime and the
date and location where the offense allegedly occurred.
Arresting Agency
…is the police agency that took the inmate into custody. In
some cases, the arrest took place in a court to which the
Arresting Agency was assigned duty.
Average Daily
…is computed by calculating (category by category) the number
of
days individuals spent in jail during each reported month, and
then summing those offender days for all reported months in
the report period and dividing those sums by the total number
of calendar days in the reported months.
Population (ADP)
Average Length
Of Stay (ALOS)
…is the sum of the release date minus admission date divided
by the
total number of offenders in a specified group (i.e.,
Unsentenced, white females, drunk drivers, etc.). The average
length of stay can be calculated across a wide array of
variables.
Boarded In
… refers to inmates that another county is responsible to
incarcerate, but who are lodged in the county jail.
Boarded Out
…refers to inmates that the county is responsible to incarcerate,
but who are lodged in another county jail.
Bond Amount
…is an obligation to appear in court made binding by money
forfeit. Bond is generally set high enough to discourage an
individual from fleeing prosecution, but the constitution protects
individuals from a bond that is excessively high.
Bond Type
…may be a promise (personal recognizance), ten percent of the
bond amount, or cash/surety (the whole bond amount or
equivalent value in collateral).
Booking Date
…is the Date the inmate was formerly admitted to the jail.
Booking Number
…is the sequential number assigned to each inmate at the time
of booking. Differentiated from the jacket number, which is the
file assigned to each inmate that is re-openened every time an
inmate is booked. An inmate may have several booking
numbers over time, but only one jacket number.
Booking Time
…is the time of day an inmate is booked into jail, expressed as
military time.
Charge Level
…refers to whether the offense is a felony, misdemeanor, noncriminal traffic offense or civil offense/parole hold or unspecified
hold.
Classification Level …is an inmate management tool designed to identify special
needs inmates and segregate objectively higher risk inmates
from lower risk inmates based on custody and security risk
factors. Classification is based on criminal history, interviews,
legal status and charge level. Classification can determine
which inmates are appropriate to work in the jail as trustees or
participate in jail-based programs. Classification involves a
numerical designation from 1 to 8, with 1 being the highest risk.
Court Case Number …refers to the docket number assigned to a court case. It is
distinguished from a police case number; however, sometimes
the AS 400 system will substitute the police case number for the
court case number, particularly if an inmate is released prior to
prosecutor authorization of a charge.
Court ORI
…is the court of origin for a case, expressed as the court
number (i.e., CC-6 for Circuit Court courtroom #6).
Credit Days
…is the number of days subtracted from an inmate’s sentence
for remaining free of misconduct in jail. Known as “sheriff’s
good time”, Credit days are authorized by statute to represent
16% of an inmate’s sentence.
Crime Class
…for the purpose of this project refers to any of the nine
categories of crime severity contained in the Michigan
Sentencing Guidelines as designated by the letters A-H or M2,
with A-D representing “High Severity” crimes, along with M2 for
Second Degree Murder. Crime Class should not to be confused
with the JPIS data term “Crime Class”, which is the designation
of the “Charge Level” (misdemeanants and felons).
Disposition
…generally refers to the ultimate outcome of a criminal charge;
however, in the context of this project, disposition information,
including disposition date, refers to the outcome of a specific
court date. Thus, the disposition for a preliminary exam date
might be “BO” for bound over, followed by “GP” at the felony
arraignment date and “SJ” for sentenced to jail at the sentence
date. The AS 400 records all of the above actions as separate
dispositions on the same case.
Disposition Date
…see the above definition for Disposition.
Employment Status …is the inmate’s self-reported employment status at the time of
booking. Employment status may include full-time, part-time,
none, unemployed or student.
Extradition
…is the process whereby one state provides a defendant due
process before surrendering an inmate to another state having
jurisdiction to try that inmate’s pending criminal matter.
Failure to Appear
…is designated yes or no contingent upon whether the last
court action prior to the inmate being arrested was a missed
court date and a bench warrant issued.
Hold
…is authorization to keep a person in custody pending an event
(usually a court date or administrative hearing). The AS400
system often does not distinguish between whether an outcounty hold has been adjudicated or not.
Housed
…is the JPIS term representing all the offenders a county is
responsible to incarcerate who are lodged in the county jail.
Incarceration
…is the detailed designation for an inmate’s legal status at
admission. Incarceration differentiates, for example, between
unsentenced clients who are booked in on a writ, who come to
jail prior to arraignment, or those booked into jail after
arraignment. The 9 incarceration codes currently used are:
Court Appearance (Writ), Extradition, Hold for Out-County
Agency, Immigration Detainer, Parole Hold, Sentenced, Prearraignment, Pre-sentence and Witness.
JPIS
…stands for Jail Population Information Systems, a system of
non-delineated data maintained by the Department of
Corrections. Almost 90% of counties report JPIS data, including
Washtenaw County. The JPIS data from Washtenaw County
does not automatically update every time an offender’s legal
status changes.
Download