Evaluation of the Transfer of Responsibility for Child Protective Services to Law Enforcement in Manatee, Pasco, and Pinellas Counties, Florida, 1995-2001 (ICPSR 3933)

Time Period(s) View help for Time Period(s)

Date of Collection View help for Date of Collection

Data Collection Notes View help for Data Collection Notes

Hide

Study Purpose View help for Study Purpose

In 1998 Florida law required that the responsibility for child maltreatment investigations be transferred to the Sheriff's Offices in Manatee, Pasco, and Pinellas Counties. The investigations had previously been conducted by the Department of Children and Families, a statewide child welfare agency. In Manatee County the Sheriff's Office began accepting all maltreatment cases in January 1998, in Pinellas County in November 1999, and in Pasco County in April 2000. The aim of this study was to evaluate the outcomes produced by this change. The research questions that drove the evaluation were: (1) Were the children who had been reported as abused or neglected in the experimental counties safer as a result of the transfer of responsibility for investigations to the sheriff's office? (2) Were the perpetrators of criminal abuse or neglect more likely to face criminal sanctions in a county where the sheriff's office conducts maltreatment investigations? (3) Were there impacts or consequences in other parts of the child welfare system as a result of the transfer to the sheriff's office? and (4) Was the new system created in the sheriff's office actually implemented as intended, and were there implementation factors that affected the outcomes?

Study Design View help for Study Design

This study used a nonequivalent control group design. Performance in Manatee, Pasco, and Pinellas Counties was compared to performance in selected counties before and after the change in service took place. Pinellas County was compared to Hillsborough County, and Manatee and Pasco Counties were compared to Lee County. Data were obtained from the Florida Abuse Hotline Information System (FAHIS), an online data system that contained records of all incidents of abuse and neglect reported in the state of Florida. The hotline, located in Tallahassee, was a statewide service run by the Department of Children and Families. Intake counselors responded to phone reports of child maltreatment, decided which of the reports would be accepted for investigation, and transmitted report data to county field offices or sheriff's offices where investigations were initiated. FAHIS records were updated as investigations proceeded from initial visit through final court disposition. The study included all reports of maltreatment that were received between January 1995 and December 2000. Data on investigations were included through August 2001. This study used data from four datasets obtained from the FAHIS relational database: (1) a report dataset (one record per report with records of four report types: Initial Abuse Report, Duplicate Report, Additional Investigation Report, and Supplemental Information Report), (2) a dependent dataset (one record per child [victim or other] per incident), (3) a caretaker dataset (one record per adult [caretaker or other] per incident), and (4) a reporter dataset (one record per reporter). Due to the fact that Lee County was the comparison county for both Manatee and Pasco Counties and Manatee and Pasco Counties underwent the change in service at different dates, separate Lee County data were created for the Manatee and Pasco comparisons. Parts 1, 3, and 5 contain the Lee County data used to compare Lee County to Manatee County (and the data for the other four counties in the study). Parts 2, 4, and 6 contain the Lee County data used to compare Lee County to Pasco County. In Parts 1, 3, and 5, Lee County cases reported between January 1, 1995, and December 31, 1997, are identified as pre-experimental, and those reported between January 1, 1998, and December 31, 2000, are identified as post-experimental. In Parts 2, 4, and 6, Lee County cases reported between January 1, 1995, and March 31, 2000, are identified as pre-experimental, and those reported between April 1, 2000, and December 31, 2000, are identified as post-experimental. Part 1, Report Data -- Five Counties, was created by merging those records marked as Initial Abuse Report from the FAHIS report dataset and the reporter dataset and selected variables from the dependent dataset. This file contains specific information on reports and investigations, maltreatment allegations and findings, summary information on victims and family members, and dispositions. Part 2, Report Data -- Lee County, contains the same data on Lee County as Part 1, except that different cases were identified as pre- and post-experimental to allow comparison to Pasco County. Part 3, Family Data -- Five Counties, was created using the data in Part 1 and the FAHIS dependent dataset. This file includes information on family-specific report history, summary information on victims, abuse allegations, findings, and dispositions. Part 4, Family Data -- Lee County, contains the same data on Lee County as Part 3 except that different cases were identified as pre- and post-experimental to allow comparison to Pasco County. Part 5, Victim Data -- Five Counties, was created from the dependent dataset by including only reports that identified a dependent as a victim. Reports were dropped that involved only a dependent who was involved in an incident but not victimized by it. Part 6, Victim Data -- Lee County, contains the same data on Lee County as Part 5 except that different cases were identified as pre- and post-experimental to allow comparison to Pasco County.

Sample View help for Sample

Comparison counties were selected for two reasons: (1) they were geographically close to the experimental counties (to minimize travel costs), and (2) they were similar enough in terms of selected demographic variables. Parts 3 and 4: If a report included multiple families, it was not included in these files. Approximately 8 percent of the reports included multiple families and were dropped. Families that moved from one county to another during either the pre- or post-experimental period were also excluded. Two percent of pre-stage families and less than 0.1 percent of post-stage families were eliminated for this reason. Families were not eliminated that only changed residence counties across stages. Parts 3 and 5: Reports that were entered during the transitional period in Manatee (July 1997 through December 1997) were excluded from these files for both Manatee and Lee, Manatee's comparison county. During this period both the Manatee Sheriff's Office and DCF handled investigations and it was impossible to attribute outcomes to a specific agency. Parts 5 and 6: Victims were excluded who changed their residence during either experimental stage. Approximately 2 percent of pre-stage victims and 1.2 percent of post-stage victims were excluded for this reason.

Universe View help for Universe

Parts 1 and 2: Reports of child maltreatment received by the Florida Abuse Hotline between January 1995 and December 2000. Parts 3 and 4: Families involved in reports of child maltreatment received by the Florida Abuse Hotline between January 1995 and December 2000. Parts 5 and 6: Children alleged to be victims of maltreatment in reports received by the Florida Abuse Hotline between January 1995 and December 2000.

Unit(s) of Observation View help for Unit(s) of Observation

Data Source View help for Data Source

Data Type(s) View help for Data Type(s)

Description of Variables View help for Description of Variables

Part 1, Report Data -- Five Counties, variables include investigation close date, case priority status, number of dependents in report, reporter type, whether the report was made before or after the sheriff's takeover, number of family members in the report, number of victims in report, number of substantiated victims, and the ZIP code for the first child in the report. Part 1 also includes the following information for the first six children in the report: interim placement, disposition code, disposition date, race, gender, age, whether maltreatment was found, and the type of maltreatment. Part 2, Report Data -- Lee County, contains the same variables as Part 1. In Parts 3 through 6 pre-experimental and post-experimental data were differentiated by creating two sets of variables: one set identified as "pre" and the other as "post." As a result, Parts 3 through 6 contain two sets of variables that provide the same information but apply to different time periods. Part 3, Family Data -- Five Counties, includes these variables for the pre- and post-experimental periods: residence county, number of reports on family, ethnicity of the first child in the first report on the family, total substantiated reports on the family, number of victims in each report, family ZIP code in first dependent record, and a family maltreatment severity score. Part 3 also includes the following information on the first six pre-experimental and the first six post-experimental reports on each family: report received dates, number of substantiated abuse victims, caretaker and dependent counts, disposition index, whether maltreatment was found, type of maltreatment, and investigation close date. Part 4, Family Data -- Lee County, contains the same variables as Part 3 except that it does not have family maltreatment severity scores. Part 5, Victim Data -- Five Counties, includes these pre- and post-experimental variables: residence county, gender, ethnicity, number of reports on dependent, and number of reports substantiated. Part 5 also includes the following information for the first 24 pre-experimental reports and the first 12 post-experimental reports: report received date, age, reporter type, interim placement, disposition index, disposition date, whether maltreatment was found, type of maltreatment, and investigation close date. Part 6, Victim Data -- Lee County, contains the same variables as Part 5 except that Part 6 includes data on the first 22 pre-experimental and the first seven post-experimental reports.

Response Rates View help for Response Rates

Not applicable.

Presence of Common Scales View help for Presence of Common Scales

Part 3 includes a family maltreatment severity score.

Hide

Original Release Date View help for Original Release Date

2004-07-01

Version History View help for Version History

Hide