The categories used in the Placement History Map were all created using the fields available in the public records version of the Florida Safe Families Network database. The original database fields and the calculated fields for this project are described below.
- afcars_id: a random number assigned to identify an individual child in the records. There were 294,899 afcars_ids in the database — one for each child.
- removal_date, discharge_date: the date and time that a child was removed or discharged from care. The database covers 342,688 removals and 316,237 discharges. The earliest removal date was June 2, 1981. The database period ends at April 16, 2019.
- placement_begin_date, placement_end_date: the date and time that a child was placed with a specific provider. There were 1,142,045 placement entries in the database.
- placement_end_reason, discharge_reason: the reasons a placement ended or a child was discharged from care.
- provider_id, provider_name, provider_zip: details about each individual provider. Many foster parents had the same name, so the provider_id was used to differentiate them. The zip codes appear to be for the providers’ mailing addresses and are not guaranteed to be where a given child was actually placed, especially for institutional placements.
- placement_setting, service_category, service_type: descriptive labels assigned to a provider for that given placement. The category and type fields contained a wide range of entries, but the placement_setting field broke all placements into these 8 groups:
- Foster Family Home (Non-Relative)
- Pre-Adoption Home
- Foster Family Home (Relative)
- Group Home
- Supervised Independent Living
- Trial Home Visit
- [some entries were blank]
- 29 fields were also included that designated various types of maltreatment that caused the child to come into care during that removal. Examples of these fields include: emotional abuse/neglect, domestic violence, abandonment, inadequate supervision, and inadequate housing.
Using those fields, the following fields were then calculated for use in the project.
- placement_length, placement_number, placement_number_no_admin: the placement length was calculated as the number of days between a placement beginning and end. Where the placement had not ended yet, the date that the database was provided was used and a field was created to mark placements that were ongoing as of the end of the data. The placement numbers were calculated by incrementing for each new placement a child was in. A second “no admin” count was made skipping over runs, visitations, and other administrative entries.
- longitude, latitude, move_miles, placement_county, placement_state: using the provider_zip, the coordinates of the zip code, county and state were looked up from an online database. An estimate of the number of miles a child was moved could be calculated based on the zip code coordinates. This is only an estimate and does not account for where in a given zip code a child was placed or whether a child was moved within a provider’s network without a new entry in the database.
- roommate_count: roommates were calculated by identifying overlapping placement dates for children with a given provider. Some providers have multiple locations or units, so it is possible that some of the children identified as roommates were actually not placed in the same home or facility. The term “roommates” is a shorthand and does not mean they actually lived in the same room.
- provider_average_rate, provider_stddev_rate, child_total_avg, child_total_high: placement payments were obtained through a separate public records request with DCF. Many providers in the placement database matched exactly with the payment data. However, many did not. For those that did not match, estimated day rates were calculated based on the average payment rate for the placement_category for the provider’s lead_agency. If there were no other payments for that category by that lead agency, then the statewide average for all providers of that category was used.