March 2014  Vol. 32, No. 1


The JRSA Forum is supported by the U.S. Department of Justice, Bureau of Justice Statistics. JRSA is a national nonprofit organization. For membership or other information, call (202) 842-9330, e-mail cjinfo@jrsa.org, or visit our Web site: http://www.jrsa.org.

Karen F. Maline, Editor
Nancy Michel, Managing Editor

JRSA OFFICERS AND STAFF:

Stephen Haas, President
Janeena J. Wing, Vice President
Lisa Shoaf, Secretary/ Treasurer
Danette Buskovick, Delegate
Mark Myrent Delegate
Roger Przybylski, Appointed Delegate
Phillip Stevenson, Past President

Joan C. Weiss, Executive Director

Shawn Flower, Research Associate
Karen F. Maline, Director of Member Services
Nancy Michel, Director of Publications
Stan Orchowsky, Research Director
Jason Trask, Program Associate
Lisa Walbolt Wagner, Research Associate
Carrie Williamson, Research Associate


























































JRSA'S IBRRC Posts Reports from Vermont and West Virginia SACs

In 2013, JRSA's Incident-Based Reporting Resource Center (IBRRC) awarded funds to two Statistical Analysis Centers, Vermont and West Virginia, to conduct analyses of data from their state's incident-based reporting (IBR) system, and report on their findings. The purpose of the IBRRC project is to provide federal, state and local analysts with information and resources on the analysis and use of crime data generated from the National Incident-Based Reporting System (NIBRS), and from NIBRS-compatible state and local systems.

The Vermont study, An Analysis of Intimate Partner Violence Case Processing and Sentencing Using NIBRS Data, Adjudication Data, and Corrections Data, was written by Robin Weber, Research Director of the Vermont Center for Justice Research (VCJR), which serves as the Vermont SAC. The study is the first analysis in Vermont conducted by merging three disparate datasets to perform a more cohesive analysis of domestic violence incidents, processing, and sentencing in Vermont. Data elements from NIBRS were merged with the VCJR Adjudication Database to perform incident-level analysis of court dispositions. The data from the new combined dataset were then merged with information from the Department of Corrections' Most Serious Status Extract to determine the type of sentence served. For the first time, Vermont was able to analyze the impact of the circumstances of the offense as reported in NIBRS on court and sentencing decisions.

The Vermont project built upon a prior JRSA-funded project using NIBRS to analyze domestic violence incidents reported from 2003 through the second quarter of 2011. That study found that the type of agency and the location of the agency were statistically significant in predicting whether a defendant was arrested or was cited to appear on the charge. This project analyzed the effect that the decision to arrest or issue a citation has on court processes. Of the more than 4,000 domestic violence arrests made during the study period, 69% resulted in a criminal charge. The project also used variables in the Adjudication Database to analyze the citation or arrest decision and found that the police officer's decision about arrest influences the bail and charging decisions.

Prior research in Vermont indicated that women were statistically more likely to be incarcerated for domestic violence offenses than men. Policy makers and stakeholders posited that it could be because there are no female batterer programs outside of a facility. This project enhances that prior research by using NIBRS data elements combined with Department of Corrections' data and the Adjudication Database to examine the role of gender in sentencing and shows that men were more likely to be incarcerated than women.

The West Virginia study, Testing the Validity of Demonstrated Imputation Methods on Longitudinal NIBRS Data, was conducted by Christina R. LaValle, Stephen M. Haas, and James J. Nolan of the West Virginia Office of Research and Strategic Planning (the West Virginia SAC).

In many states, recorded criminal offenses collected by law enforcement agencies using either the Uniform Crime Report (UCR) or NIBRS format are the only available source of crime data used to generate reports and crime statistics at the state and local levels. Despite the voluntary and inherent limitations of crime data collection systems, data are most often reported "as is" and assumed correct. Previous studies on state IBR data and national UCR data have demonstrated various issues with completeness (i.e., partial and nonreporting agencies) and accuracy (i.e., classifying zeros and irregular reporting). Therefore, reporting data "as is" is not the most accurate depiction of IBR data. This research sought to test simple, practical tools for assessing data quality and measuring the performance of partial and nonreporting imputation methods applied to longitudinal data to improve the accuracy of state NIBRS data, especially when used for state and county trend analyses over time.

The data quality and imputation methods applied were developed by the West Virginia SAC in previous research. Data quality methods involved testing guidelines used for classifying crime counts of zero as true zeros or missing data using criteria based on the total number of property crimes reported in one year and the number of consecutive months in which all crime counts (violent, property, and non-index) were zero. In addition, two outlier detection methods were used that involved calculating two statistics. One method measured how many times larger or smaller a monthly crime count was compared to its median, and the second method calculated a statistic to measure a "gap" and "range" based on maximum, minimum, and median monthly crime counts. The West Virginia imputation methods estimated missing data for partial reporting agencies using a quarterly average and used modified population groups to estimate data for nonreporting agencies.

Using a longitudinal dataset spanning from 2007 to 2011, a missing value pattern and full reporting dataset were established for a simulation study to investigate the performance of the West Virginia imputation methods. The imputation methods for partial reporting and nonreporting agencies were applied simultaneously to the simulated data, and accuracy statistics, mean absolute error, root mean square error, and bias were calculated and compared to FBI imputation methods. The imputation methods developed by the West Virginia SAC outperformed the current FBI imputation methods when applied to state IBR data. The accuracy statistics suggest that the West Virginia imputation methods are more accurate and precise in estimating data for partial reporting and nonreporting agencies when compared to the FBI methods.

Overall, the West Virginia imputation methods seemed to reliably estimate missing data for producing stable crime trends as a means to count crime not reported, and they outperformed methods currently used by the FBI.