Measuring Justice Dashboard: North Carolina’s Criminal Justice Data Visualization Platform

The Measuring Justice Dashboard helps stakeholders understand state and local systems, see where they are doing well, and spot areas that may need attention.

Choose your metric from the list at the left and navigate through our data visualizations, tailoring content as you go. You will be able to:

• Compare your county to other jurisdictions
• Look at state & county-specific results
• Examine different time periods
• Drill down on specific case types
• Explore metrics by race

Click here for a 3-min dashboard tutorial.

Our Dashboard is growing; check back for new metrics!

Data

Unless otherwise noted, criminal justice data come from our database of North Carolina Automated Criminal/Infraction System (ACIS) records. For information about other data sources, please visit the Methodology section.

Have suggestions for the dashboard? Connect with us here.

1. Data Sources

1. Court system data: North Carolina Administrative Office of the Courts (NC AOC) Automated Criminal/Infraction System (ACIS).

2. Population data: U.S. Census via the National Cancer Institute, which has processed census data for populations in racial categories that can be mapped to ACIS data (see “Demographic Groups” below).

3. County Federal Information Processing Standard (FIPS) and Zip Codes: the U.S. Department of Housing and Urban Development (HUD).

2. Connecting Individuals Across Cases

ACIS data does not include a field or identifier that connects individuals to all cases they may have in the system. We thus developed a matching process to do so. The individual identifier is needed for various analyses, such as tracking whether a person with a pending incident acquires a new criminal charge during the pretrial period.

3. Missing Data

If no data are reported for a particular metric (e.g., no warrantless arrests for a particular month in a very small county), graphs will have a break and maps and charts will have a missing element. Additionally, to avoid the risk of allowing individuals to be identified, data visualizations will not be displayed if there were three or fewer data points. When a record is missing data on a field that is needed for a particular analysis (such an individual's date of birth for an analysis that calculates age), we excluded the record from the analysis.

The Second Chance Act, S.L. 2020-35, became effective December 1, 2021. Per data reported from the NC AOC, in the three months after the statute's effective date 143,320 cases were expunged (an average of 47,773 per month). Operation of this law impacts Dashboard data for December 2021.

4. Race Tabs, Demographic Groups & Dropped Cases Due to Inconsistent Racial Data

ACIS includes the following demographic categories: Asian, Black, Hispanic, American Indian, Other, Unknown, White, and Non-person. When processing ACIS data for racial metrics, we dropped cases where the person was identified as Unknown or Non-person (e.g., a business entity).

To obtain demographic information for populations at the state and county level used in our racial metrics, we extracted demographic data from the Surveillance, Epidemiology, and End Results Program at the National Cancer Institute (NCI). That program collects and structures U.S. Census Bureau data into age and racial categories. Importantly, we could match the NCI racial categories for Non-Hispanic White, Non-Hispanic Black, Non-Hispanic Native American, and Hispanic to ACIS race categories. Because the ACIS Asian category did not directly map to the related NCI category, which includes Asian and Pacific Islanders, we categorized ACIS Asian and NCI Asian and Pacific Islander as Other. The Other category also includes people who identified as part of two or more racial groups.

We pull NCI population data for individuals aged 16-74. Limiting the data by age in this way helped ensure the NCI served as a good comparison because that age range accounted for 99.4% of all charged persons in the Dashboard criminal justice data set in use at the time this methodology was adopted (January 1, 2014 - December 31, 2021).

NCI data starts in 2014 and currently ends in 2020. The Dashboard includes ACIS data from 2014 through 2021. Whenever a dashboard analysis displays county or state populations, we show population figures for the same year or the most recent year available (e.g., a 2019 analysis will show 2019 population whereas a 2021 analysis will show 2020 data). If more than one year is selected, we will show the most recent population figure (e.g., if 2019 through 2021 is selected, we will show 2020 data). As new NCI data are released, we will update the Dashboard.

For approximately 2% of all individuals in our dataset (about 91,000 people), ACIS data includes inconsistent demographic information either within a single case or across multiple cases for a single individual. We excluded these individuals from all analyses. We will monitor this issue as our dataset grows and modify our methodology (with a note here) as appropriate.

5. Charging Data Analysis

Included Charges

We include all charges at case initiation in served cases. This applies to cases initiated in district court and to cases initiated directly in superior court. For cases that begin in district court and move to superior court, we include only charges upon case initiation in district court. In later analyses, we will examine how charges change as a case moves from district court to superior court. Because this analysis presents charges by service date, it includes only charges in cases where there is a service date.

We dropped charges that are not substantive offenses under North Carolina law, including:

• ARMED HABITUAL FELON
• BILL OF PARTICULARS
• CIVIL REVOCATION DR LIC (10)
• CIVIL REVOCATION DR LIC (30)
• CONTEMPT BY PROBATIONER
• CRIMINAL CONTEMPT
• EXTRADITION/FUGITIVE OTH STATE
• FEL HIT/RUN FAIL STOP INJ RV 1
• FEL HIT/RUN FAIL STOP INJ RV 2
• FEL PROB VIOL OUT OF COUNTY
• FELONY PROBATION VIOLATION
• FUGITIVE
• GOVERNOR'S WARRANT
• H/I FELONY PROBATION VIOLATION
• HABEAS CORPUS
• HABITUAL FELON
• MISD PROB VIOL OUT OF COUNTY
• MISDEMEANOR PROBATION VIOL
• MOTION TO WITHHOLD WAGES
• MOTIONS
• PROBATION REVOCATION APPEAL
• PROVISIONAL CIVIL REV DL (30)
• SEIZURE/FORFEITURE OF VEHICLE
• SHOW CAUSE
• VIOLATION OF A COURT ORDER
• VIOLENT HABITUAL FELON

ACIS data includes some offenses charged as “Free text” offenses. Where we could not identify these offenses as felonies or misdemeanors, we coded them as “unknown” and dropped them from our analysis. This group of dropped offenses include the following:

• ASSAULT - FREE TEXT
• BRIBERY - FREE TEXT
• COMMERCIAL SEX - FREE TEXT
• DANGEROUS DRUGS - FREE TEXT
• FAMILY - FREE TEXT
• FLIGHT/ESCAPE - FREE TEXT
• FORGERY - FREE TEXT
• FRAUD - FREE TEXT
• HEALTH LAW - FREE TEXT
• INVADE PRIVACY - FREE TEXT
• LARCENY - FREE TEXT
• LIQUOR - FREE TEXT
• OBSCENITY - FREE TEXT
• OBSTRUCT POLICE - FREE TEXT
• OTHER - FREE TEXT
• PROPERTY - FREE TEXT
• PUBLIC INTOXICATION - FREE TEX
• PUBLIC ORDER - FREE TEXT
• STOLEN VEHICLE - FREE TEXT
• TAX REVENUE - FREE TEXT
• WEAPON OFFENSE - FREE TEXT

Additionally, ACIS data include some offenses identified as a felony or misdemeanor but without additional detail from which we could categorize the offense as violent or non-violent. We also coded these offenses as “unknown” and dropped them from our analysis. This group of dropped offenses include the following:

• ACCESSORY AFTER THE FACT (F)
• ACCESSORY AFTER THE FACT (M)
• FELONY AID AND ABET
• FELONY CONSPIRACY
• MISDEMEANOR AID AND ABET
• MISDEMEANOR CONSPIRACY

Dropping offenses leads to some understating of total criminal charges, individuals and cases, though some of the excluded charges may have been for infractions and thus properly excluded from the analysis, which examines only criminal offenses.

Violent and Non-violent Offense

Some offenses obviously are violent, such as first-degree murder, or nonviolent, such as food stamp fraud. Notes on the violent or nonviolent categorizations we applied to other offenses include:

• All forms of child abuse under G.S. 14-318.4 categorized as violent (note that statute includes abuse by prostitution; abuse by negligence)
• All homicides categorized as violent (e.g., involuntary manslaughter; 2nd-deg murder by drug distribution)
• All sexual touchings categorized as violent (e.g., rape, sex offense, crime against nature)
• Breaking and entering: (1) With intent to terrorize categorized as violent (2) Other breaking and entering and variants categorized as nonviolent
• Burglary categorized as violent
• Cruelty to animals categorized as nonviolent
• Disabled and elder adult abuse and neglect offenses categorized as violent
• Disclosure of private images categorized as nonviolent
• Disorderly conduct categorized as nonviolent
• Drug crimes categorized as nonviolent, including those where drugs are sold to a minor or where a minor is used to facilitate a drug crime
• Explosive offenses categorized as violent (e.g., malicious use of explosive to damage property)
• Exposing children to fire categorized as violent (like child abuse)
• Failure to report disappearance of child under G.S. 14-318.5 categorized as nonviolent; Failure to report crime against juvenile under G.S. 14-318.6 categorized as nonviolent
• False report and hoax offenses categorized as nonviolent
• Felony arson & burning offenses categorized as violent; misdemeanor burning offenses categorized as nonviolent (fail to extinguish fire; set fire to woods, lands, fields)
• Firearm and other weapon possession offenses categorized as nonviolent (poss. stolen firearm; felon in possession; purchasing firearm in violation of domestic violence order; possession of weapon by prisoner; possession of weapon of mass destruction; weapons on educational property; carrying concealed)
• Fleeing to elude arrest and elude arrest/passenger flee categorized as nonviolent (on the theory that where there is injury or death, the violent aspect will be captured under separate offenses charged for that conduct).
• Gang crimes categorized as violent
• General crimes categorized the same as the intended offense (e.g., attempted murder categorized like murder)
• Human trafficking and servitude offenses categorized as violent (involuntary servitude; sexual servitude)
• Impaired driving-related charges not directly involving impaired driving categorized as nonviolent or other (DWLR after impaired revocation; FTA 2 yrs after implied consent offense)
• Kidnapping and restraint offenses categorized as violent (e.g., misdemeanor false imprisonment; abduction of children; felonious restraint; felony custody order violation; taking hostage by prisoner; confinement of minor or disabled person for purposes of prostitution)
• Larceny from the person categorized as nonviolent
• Malicious conduct by prisoner categorized as violent except version involving exposing genitalia, which is classified as nonviolent
• Misdemeanor ordinance crimes categorized as nonviolent
• Open container and alcohol consumption while driving offenses categorized as DWI
• Patient abuse and neglect categorized as violent, including violations under 122C-66(a).
• Post-homicide offenses categorized as nonviolent (e.g., conceal/fail to report death; destroy remains)
• Riot offenses categorized as violent
• Sex offender registration offenses categorized as nonviolent
• Shooting offenses without immediate human victim categorized as violent (e.g., discharging firearm w/in an enclosure to incite fear; discharging firearm into occupied premises). But note that per rule above, if offense is defined by ordinance (discharge firearm in city), it is categorized as nonviolent.
• Solicitation of child by computer offenses categorized as violent
• Soliciting and patronizing prostitution offenses categorized as nonviolent except that any prostitution offense involving a child/minor treated as violent, to be consistent with child abuse
• Threat/harassment/intimidation/stalking/hazing offenses categorized as violent (communicating threats; harassing phone calls; intimidating witnesses; cyberstalking)
• Traffic offenses resulting in death or injury: (1) Felony traffic offenses resulting in death or injury categorized as violent (fail to move over serious injury/death; felony serious injury by vehicle; felony death by vehicle; passing stopped bus striking a person; hit and run); (2) non-felony traffic offenses resulting in death or injury categorized as nonviolent

Drug offense

The term drug offense includes all NC AOC offense codes associated with offenses proscribed by General Statutes Chapter 90 (Medicine and Allied Occupations), except for the following offense codes, which are excluded:

• 6039, PRACTICE PHARMACY W/O LICENSE, G.S. 90-85.21B
• 6043, PRACTICE MEDICINE W/O LICENSE, G.S. 90-18(A)
• 6049, PRACT MEDICINE W/ NON-NC LIC, G.S. 90-18(A)

Also excluded from the term drug offense are:

• Drug-related offenses proscribed in other Chapters of the General Statutes
• NC AOC offense code 3599, DANGEROUS DRUGS – FREE TEXT
• Drug-related ordinance offenses

DWI

DWI offense means any type of impaired driving and other alcohol/driving offenses (e.g., open container in passenger area).

Traffic Offense

Traffic offense means any misdemeanor proscribed by North Carolina General Statutes Chapter 20 (Motor Vehicles).

Other, Non-Violent Misdemeanors

This group contains misdemeanor offenses not classified anywhere else – e.g., ordinance and non-drug offenses.

Number of Charges

The goal of the Incident Tab is to show the number of incidents with single or multiple charges and the demographics of populations for each offense count level (specifically, 1-6 charges per incident and 7+ charges per incident). Thus, an incident includes all charges against an individual with the same service date regardless of case number.

Age

For the age analysis, we calculate a charged person's age at the time the charging instrument for the first offense in the incident was served.

6. Citation v. Arrest Analysis

We limited our analysis to highest charge misdemeanor incidents that were initiated by citation or magistrate's order. A magistrate's order is the process issued after a warrantless arrest. Because we were focused on arrests made in connection with the initiation of charges by officers, arrests made pursuant to warrants or Orders for Arrest are not included. An incident includes all cases against a person on the same date and in the same county.

For the tabs titled Map, Time, Count, and Race, we categorize incidents so that they take the highest level of offense severity in that incident. So, if one incident included both a traffic and non-traffic offense, the incident would be counted as non-traffic rather than as traffic. The Rank tab includes all offenses that resulted in an arrest.

7. Summons v. Warrant Analysis

We limited our analysis to highest charge misdemeanor incidents that were initiated by summons or warrant for arrest. An incident includes all cases against a person on the same date and in the same county.

For the tabs titled Map, Time, Count, and Race, we categorize incidents so that they take the highest level of offense severity in that incident. So, if one incident included both a violent and a non-violent offense, the incident would be counted as violent rather than non-violent. The Rank tab includes all offenses that resulted in an arrest.

Unserved, active warrants are not included in ACIS data that we receive from the court system. Thus, summons rates may be lower and counts for warrants may be higher than shown on the Dashboard.

8. Court Non-Appearance Analysis

Our analysis includes all incidents initiated on or after January 1, 2014, and all later initiated charges in those incidents regardless of what court the matter was in (district or superior) when the charges were served.

An incident includes all charges served on a person on the same day in the same county, even if those charges were initiated in different cases.

We classify an incident as having a non-appearance if any of its offenses has an ACIS entry for a called and failed or a motor vehicle failure to appear. We classify an offense as having a non-appearance if it has an entry in either of those fields.

Because court records record only the last non-appearance, we can report only whether an incident had a non-appearance. Thus, non-appearance data are shown at the incident level, by incident initiation year. For example, 2019 reflects non-appearance rates and counts for incidents initiated in 2019. Because older incidents afford charged individuals longer periods of time when they can accrue a non-appearance and may involve more required appearances, we would expect to see higher non-appearance rates for older incidents. For example, an incident that was initiated on January 1, 2019, and disposed on December 31, 2021, was pending for 3 years. By comparison, the maximum pending time for an incident initiated on January 1, 2021, is shorter.

For the age analysis, we calculate a charged person's age at the time the first offense in the incident was served.

For the incident duration analysis, we use the earliest offense service date as the date of incident initiation. For closed incidents, we use the last case disposition date as the date the incident is disposed. For pending incidents, we use the last date of AOC data in the Dashboard for that data point. We classify an incident as having been disposed of in less than one month if it was disposed of before a full calendar month passed; as having been disposed of in one month if it was disposed of between one and two calendar months; and so on. For incidents initiated in district court that proceed to superior court, we calculate incident duration as the total time the incident was pending in both courts (e.g., for incidents that resolve in superior court, from first offense service date in district court to superior court incident disposition date).

We compute a charged person's residency (in-county, out-of-county, in-state, out-of-state) using the zip code of their address in ACIS. We categorize zip codes that cross county and state lines as in-county and in-state.

For this analysis, we dropped approximately 0.7% of incidents where reported case disposition date was before service date and where the individual's zip code was missing. For this reason, the number of incidents reported in the Court Non-Appearance metric does not match that reported in other metrics.

Stakeholders participating in engaged projects with the Lab may see some variation between non-appearance data reported here and in reports we produce in those projects. In local project reports, we typically are comparing changes in non-appearance rates from a pre- to post-implementation period. To ensure that we are making apples-to-apples comparisons in those reports, we control the time periods at issue e.g., looking at non-appearances for a fixed time period (e.g., 18 or 24 months) in both the pre- and post-implementation periods. Because the Dashboard is not executing such a comparison, the time periods used in the Dashboard non-appearance analyses are limited only by an incident's disposition date. Additionally, and as noted under the heading “Race Tabs, Demographic Groups & Dropped Cases Due to Inconsistent Racial Data” and elsewhere in this section, some records are dropped from the Dashboard data set because of missing or inconsistent data.

9. New Criminal Activity Analysis

We executed new criminal activity analyses at the incident level. An incident includes all cases against a person on the same date and in the same county.

On the Map and Race tabs, data are reported at the incident level, by service year. For example, when the year 2019 is selected for these tabs, the Dashboard displays information for all incidents served in 2019, both pending and disposed. On the Count, Rank, and Duration tabs, data are reported on new pretrial charges and the time periods for those tabs reflect pretrial offense dates. Thus, selecting 2019 for these tabs brings up all new pretrial charges with offense dates in 2019, regardless of when the original incident occurred.

To compute new pretrial criminal activity, we begin by examining whether the original incident is followed by any new criminal charges while the original incident was pending. An incident is pending from its service date until the last disposition date of any included case. For example, if an incident served on January 1, 2022, was composed of two cases, one disposed of on March 1, 2022, and a second disposed of on June 1, 2022, the incident would be pending between January 1, 2022, to June 1, 2022. If an incident includes any cases that have not been disposed, we use the last date of ACIS data in the Dashboard as a proxy for disposition date.

When computing new criminal activity in the pretrial period, we examine the offense date rather than the service date of the new criminal charges. For example, if an incident was pending between January 1, 2022, and June 1, 2022, and the person was served with a new criminal offense on March 15, 2022, with an offense date of February 15, 2022, the new charge is considered new criminal activity and the criminal activity date is the offense date, February 15, 2022. If, however, the offense date of the new offense is June 15, 2022, it would not be counted because the offense occurred after the original incident was disposed.

In the Rank tab, the column “Percent Total” represents the percentage of all new pretrial criminal charges attributable to the relevant offense.

In the Duration tab, we show the time it takes for an individual to acquire a new pretrial criminal charge. Time to charge is calculated by computing the difference between the service date of the original incident and the offense date of the new pretrial criminal charge.

Stakeholders participating in engaged projects with the Lab may see some variation between new pretrial criminal activity data reported here and in reports we produce in those projects. Several factors explain these variations. First, the unit of analysis in those project reports typically is at the defendant level. In the Dashboard, however, the unit of analysis is at the incident level, and this may lead to higher pretrial criminal activity rates. For example, suppose that Joe is arrested in January for crime X. While that incident is pending, Joe acquires two more arrests, one in February and one in March. When pretrial criminal activity is calculated at the defendant level, Joe would appear once in the data—as a single defendant. When pretrial criminal activity is calculated at the incident level, Joe would appear multiple times, for each of his separate incidents.

A second possible reason for differences between the Dashboard data and estimates in engaged projects is that in local project reports, we typically are comparing changes in pretrial criminal activity rates from a pre- to post-implementation period. To ensure that we are making apples-to-apples comparisons in those reports, we control the time periods at issue e.g., checking for pretrial criminal activity for a fixed time period (e.g., 18 or 24 months) in both the pre- and post-implementation periods. Because the Dashboard is not executing such a comparison, the time periods used in the Dashboard are limited only by an incident's disposition date. As a result, Dashboard pretrial criminal activity rates may be higher than rates presented in local project reports.

A third possible reason for differences between the Dashboard data and estimates in engaged projects is that, as noted above (see “Race Tabs, Demographic Groups & Dropped Cases Due to Inconsistent Racial Data”), some records are dropped from the Dashboard data set due to inconsistent data.

10. Definitions

Citation. A citation is a directive, issued by a law enforcement officer or other person authorized by statute, that a person appear in court and answer a misdemeanor or infraction charge or charges.

Magistrate's Order. A magistrate's order is used only when a person has been arrested without a warrant. When a person is arrested without a warrant, he or she must be brought, without unnecessary delay, to a magistrate for an initial appearance. If the magistrate finds that there is probable cause to charge the person with a crime, the magistrate must issue a magistrate's order.

Order for Arrest. An order for arrest is an order issued by a justice, judge, clerk, or magistrate that a law-enforcement officer take a named person into custody.

Summons. A criminal summons consists of a statement of the crime or infraction charged and an order directing the accused to appear in court and answer the charges; it does not order a law enforcement officer to take the person into custody. A criminal summons is based on a showing of probable cause supported by oath or affirmation.

Warrant for Arrest. A warrant for arrest consists of a statement of the crime charged and an order directing that the accused be arrested and held to answer the charges. Thus, unlike the citation and criminal summons, which merely direct an individual to appear in court to answer charges, the warrant for arrest directs law enforcement officers to arrest the accused. The warrant must be based on a showing of probable cause that the person committed a crime. That showing may be made on oath or affirmation.

11. Dashboard Updates

We continue to improve and refine the Dashboard and may occasionally update the data to address issues discovered during quality control. For more information about updates, please reach out using the weblink on the Dashboard's Home page.

Contributors and Acknowledgments

Andre Assumpcao, Post-Doctoral Fellow, UNC School of Government Criminal Justice Innovation Lab, for concept, design, analysis, and development of the Dashboard.

Jessica Smith, W.R. Kenan Jr. Distinguished Professor & Director, UNC School of Government Criminal Justice Innovation Lab, for concept and development of the Dashboard and Dashboard text.

Christopher Tyner, former Legal Research Associate, UNC School of Government Criminal Justice Innovation Lab, for offense coding.

Ethan Rex, Project Manager, UNC School of Government Criminal Justice Innovation Lab, for data support.

Alex Cowell, Research Director, UNC School of Government Criminal Justice Innovation Lab, for help with modifications, improvements and quality control.

Jonathan Crabtree, Assistant Director for Research Data Information Systems, Odum Institute, The University of North Carolina at Chapel Hill, for data science and management support.

Matthew Dunlap, Systems Programmer/Analyst, Odum Institute, The University of North Carolina at Chapel Hill, for data programming support.

Rebecca Tippett, former Director, Carolina Demography, The University of North Carolina at Chapel Hill, for advice regarding demographic categories and analysis.

Michelle Russell, Principal Associate, Public Safety Performance Project, The Pew Charitable Trusts, for her foundational research and data visualizations of North Carolina non-appearance rates in connection with the North Carolina Court Appearance Project.

Jamie Vaske, Assistant Department Head, Director of Administration of Justice Program, Criminology and Criminal Justice Program, Western Carolina University, for her support and advice on creating the pretrial failure metric.