This codebook describes the criminal CSV file that fjc-idb creates from the
FJC Integrated Database (IDB) files for criminal defendants in the U.S.
district courts. It follows the FJC's Criminal Integrated Database (IDB): 1996
to Present codebook (August 2016), adapted to the column names, labels, and
added columns that fjc-idb produces. The original codebook is available from
the FJC IDB page.
One row per defendant. Unlike the civil and appellate files, each row is
a defendant, not a case. A case with several defendants has several rows,
identified by defendant_number or defendant_key.
Years. The IDB has one file for each fiscal year, which runs from
October 1 through September 30. Each file is a snapshot with one record for
every defendant filed or terminated during the year, or still pending at its
end. A record holds both the filing and, once the defendant is terminated,
the closing data. The record is the most recent one the court sent, so
earlier information (such as the charges in a superseded indictment) can be
overwritten.
Pending defendants. A pending defendant is repeated in every file until
terminated, so fjc-idb keeps pending defendants only from the latest file
in the range. In those rows, pending is 1 and data_file_year is 2099.
A defendant terminated in one year and terminated again in a later year
(for example, after a retrial) has a row for each termination.
Code changes. The FJC updated the data back to 1996 to use the codes in
place since fiscal year 2012. Some fields were created in 2012 and are
missing or calculated for earlier years, as noted below.
Column order. The CSV keeps the order of the original file, with each
added column placed after the columns it comes from. The tables below group
the columns by subject instead.
Each original IDB field is renamed to a descriptive name. Pass rename_columns=False (or --rename-columns false) to keep the IDB names, such as FOFFCD1.
Missing data
The IDB code -8 and the date 1900-01-01 mean missing data. They are written as NA in date columns and left blank in all other columns.
Labels
Columns marked with a label table have their codes replaced by labels. Codes with no label are kept as they are.
Indicators
Added indicator columns are 1 for yes and 0 for no.
Added columns
Columns marked added are created by fjc-idb and are not in the IDB.
Numbered columns
The IDB records up to five offenses for each defendant at filing and at termination, ranked by severity. Columns ending in _1 through _5 describe offense 1 (the most severe) through offense 5.
Type of the originating case, if the case was transferred.
case_type_magistrate
TYPEMAG
text
Type of the magistrate judge case, if the case was merged from one.
defendant_key
DEFLGKY
text
Unique key for the defendant: district, office, docket number, case type, defendant number, and reopen sequence.
case_key
CASLGKY
text
Unique key for the case: district, office, docket number, case type, and reopen sequence.
magistrate_docket
MAGDOCK
text
Docket number of a magistrate judge case later merged into the criminal case.
magistrate_defendant_number
MAGDEF
integer
Defendant number in that magistrate judge case.
county
COUNTY
text
Five-digit FIPS code of the county where the offense was committed (two digits for the state and three for the county). 88888 means outside the home state and 99999 outside the United States.
Date the defendant became a fugitive. Missing before fiscal year 2012.
fugitive_end_date
FGENDDATE
date
Date a fugitive defendant was taken into custody. Missing before fiscal year 2012.
source
SOURCE
text
Where the record came from: CMECF (sent by the court through CM/ECF), DSNAPSHOT or FUGITIVE (snapshots from the AO's legacy system), or MASTER (the legacy live database).
version
VER
integer
0 for the first version of the record and increasing by one with each update. Missing before fiscal year 2012.
load_date
LOADDATE
date
Date the record was loaded into the AO's database.
last_update_date
UPDATE
date
Date of the last action on the record. Discontinued in March 2012.
Added. One column for each group of offense codes, which is 1 if any of the defendant's five filing offenses is in the group. Blank if all five are missing. See Offense groups.
offense indicators at termination
D2TOFFCD1–5
integer
Added. The same columns built from the five termination offenses, each with _at_termination added to its name. Blank for defendants who have not been terminated.
The groups use the D2 codes because they are the current classification and
cover every year in the data.
Date the final sentence was entered on the docket.
termination_date
TERMDATE
date
Date the case was closed: the disposition date if all counts were dismissed or ended in acquittal, or the sentencing date if the defendant was found guilty on any count. The AO uses this date to count terminations.
filing_to_termination
FILEDATE, TERMDATE
integer
Added. Days from filing_date to termination_date. Blank if either date is missing.
days_to_proceeding
INT1
integer
Days from the earlier of the filing date and first appearance date to the proceeding date.
days_proceeding_to_disposition
INT2
integer
Days from the proceeding date to the disposition date.
days_disposition_to_sentencing
INT3
integer
Days from the disposition date to the sentencing date.
termination_courthouse
TERMOFF
text
Office where the case was terminated, as "City, ST". See Courthouses.
termination_judge
TJUDGE
text
Code of the judge who terminated the case. Blank in the public files.
For each offense from 1 to 5, ranked by the severity of the disposition and
penalty. The AO reports disposition_1 as the defendant's final disposition.
These columns are 1 if the defendant counts toward the AO's published
statistics for the fiscal year and 0 if not. They were created in fiscal year
2012 and calculated for fiscal years 2005 through 2011.
Column
IDB field
Counts defendants
count_filed_including_transfers
CTFILTRN
filed, including transfers between districts
count_filed
CTFIL
filed, excluding transfers between districts
count_filed_without_reopens
CTFILWOR
filed as original proceedings
count_filed_reopens
CTFILR
filed by reopening, remand, appeal, or retrial
count_terminated_including_transfers
CTTRTRN
terminated, including transfers between districts
count_terminated
CTTR
terminated, excluding transfers between districts
count_terminated_without_reopens
CTTRWOR
terminated in original proceedings
count_terminated_reopens
CTTRR
terminated after reopening, remand, appeal, or retrial
count_pending
CTPN
pending at the end of the year, including long-term fugitives
count_pending_without_fugitives
CTPNWOF
pending at the end of the year, excluding long-term fugitives
The IDB stores a two-character district code. fjc-idb adds the two-digit
circuit in front of it (circuit 9 and district 73 become 0973), which is
how the appellate codebook lists districts.
Office codes repeat across districts, so each office is identified by its
combined district code followed by the office code (09732 is Los Angeles).
The names come from Appendix D of the appellate codebook, which uses the same
office codes, and are followed by the state's two-letter abbreviation. By
default, twelve names misspelled in that codebook are corrected; pass
fix_city_spellings=False (or --fix-city-spellings false) to keep the codebook
spellings. The appellate codebook lists every courthouse code.
A severity code has three characters. The first gives the maximum prison
sentence, the second the type of crime, and the third the maximum fine. For
example, 235 means four to five years of prison, a crime against a person,
and a fine of $5,001 to $10,000. Codes range from A00 (lowest) to 939
(highest).
Code 1 also covers dismissals of counts alleging only criminal forfeiture and
cases remanded or with an appeal affirmed or denied. Dismissals without
prejudice use code 15.