DATA DASHBOARD: FAQ

The CSI data dashboard focuses on state-level data and does not include estimates of records for federal arrests or convictions. Our methodology focuses on state conviction and non-conviction records and does not attempt to model federal records or the overlap between state and federal records.

The numbers in the CSI data dashboard may differ from other estimates due to our unique methodology, which incorporates a variety of metrics, including conviction rates, recidivism rates, deportation rates, mortality/survival rates, and interstate mobility rates, building on the research conducted by Shannon et al. (2017) and the Brennan Center for Justice (2020). Other estimates use different methodologies and data sources.

Our dashboard presents the estimated number of people with records and the impacts of several potential configurations of Clean Slate legislation, while our state fact sheets present impact estimates based on the specific legislation that’s been submitted to the legislature.

The CSI data model is not directly based on criminal history repositories, state police, courts, or corrections data. Instead, it uses methodologies from published research and data from the sources detailed in our Methodology Document and Overview.

We frequently compare our estimates to a variety of state administrative data sources. You can read more about how our estimates compare to other estimates and data sources in Part 2: CSI’s Estimates Compared to Other Data Sources of our Methodology.

Our dashboard is not connected in any way to personally identifiable information. Our methodology focuses on aggregate data to estimate the number of unique individuals impacted by records in all 50 states and Washington, D.C., ensuring privacy and confidentiality.

The CSI data model is distinct in its focus on estimating the number of unique individuals impacted by records across all 50 states and D.C., using a comprehensive set of metrics that include arrest, conviction, recidivism, deportation, mortality/survival, and interstate mobility rates. This approach allows for a broad understanding of the scope and impact of criminal records across demographic groups and is different from other databases that rely on state administrative data sources or include personally identifiable information.

CSI calculates the number of people in each state and D.C. by 1) estimating the number of people with felony convictions using prison release data, state-level recidivism data, and adjusting for mortality, interstate mobility, and deportation rates, 2) estimating the number of people with misdemeanor convictions using arrest data from the Uniform Crime Reporting Program, conviction rates, and recidivism data, and then adjusting for the same factors as felony convictions, 3) estimating the number of non-convictions by calculating the total number of people charged but not convicted based on state-level conviction rates, 4) adjusting these estimates by demographic data to account for race, ethnicity, and sex differences, and 5) summing the estimated populations with each type of conviction and non-conviction record to calculate the total number of people with a record in each state.

Our website shows an overview of our methodology, data sources, and detailed methodology documentation.

CSI uses a variety of data sources to construct our data model and dashboard, including:

National Corrections Reporting Program (NCRP), Survey of Prison Inmates, and National Prisoner Statistics Program for prison release data

Annual Survey of Probation for probation data

Uniform Crime Reporting Program (UCR), National Center for State Courts, and Measures for Justice for arrest, conviction and reconviction data

CDC age-adjusted death rates for mortality adjustments

U.S. Census Bureau data for interstate mobility estimates

Transactional Records Access Clearinghouse (TRAC) for deportation data

CSI focuses on the number of people impacted to accurately represent how criminal records affect individuals and communities. This approach highlights the human aspect of criminal legal system data and enables an understanding of the potential for Clean Slate laws to offer relief to those impacted.

CSI’s methodology builds upon previous research by incorporating methods established by Shannon and colleagues (2017) and the Brennan Center for Justice (2020) for estimating the populations impacted by criminal records. We refine these methods with additional data sources and adjustments for recidivism, mortality, and other factors affecting population estimates' accuracy. We also expand upon the scope of earlier research by breaking down data by race, ethnicity, sex, and record type and adjusting for dynamic factors like interstate mobility and deportations.

The CSI data dashboard is designed to be as accurate as possible by using the most reliable data from official federal and state databases and accounting for known challenges such as incomplete or erroneous data, variance in categorizations, and dynamic factors affecting the population.

While our methodology is designed to produce accurate estimates, we acknowledge there are limitations in estimating the population with records due to the lack of availability and quality of state-level data and potential variations across states. Our methodology document describes our data model as consistent with other data sources, often producing more conservative estimates.

The CSI data dashboard can be valuable for policymaking and research endeavors. It is designed to provide estimates of the populations impacted by criminal records in each state, broken down by types of records and demographics. The dashboard can enable users to examine the racial equity implications of Clean Slate legislation and its components and can assist policymakers and advocates in identifying and addressing potential disparities. Our dashboard can inform more equitable outcomes in implementing Clean Slate policies and can serve as a resource for further research into the impact of criminal records on individuals and communities.

The frequency of updates to the CSI data dashboard is based on the availability of new data and research findings. We strive to ensure that our dashboard reflects the most current and accurate information to support stakeholders in their efforts to advance Clean Slate policies.

The CSI data dashboard estimates the total number of people with a conviction or non-conviction record. However, there are a few types of records that are not included in our estimates due to limitations of the official data sources available. For example, arrests where the highest charge was a traffic violation (not including DUI) are not included in our model because these types of arrests are not reported to the federal Uniform Crime Report Program. This may result in particularly conservative estimates of the population with a record in states like Georgia where many traffic offenses are classified as misdemeanors. Additionally, our dashboard does not include estimates of how many people have out-of-state or federal records in addition to a record within their state. Similarly, we do not include estimates of the population that owes court-mandated fines, fees, or restitution associated with a prior conviction.

Some estimates of Native Americans and Latinos with records are likely underreported due to limitations in how official federal and state data sources report information on race and ethnicity. These limitations can result in the underrepresentation of these groups in criminal legal system data, affecting our estimates, which rely on these aggregate data sources. We flag these instances as “Likely Underreported” to acknowledge the potential discrepancies and advise caution in interpreting these figures.

We have provided a video and a written user guide on our dashboard webpage that each give a comprehensive walk through of the dashboard.

Should you have further questions or require personalized support, our team is here to help. Please reach out to us via email at data@cleanslateinitiative.org.