Data Quality Check 2

2.  Ensure that the required homicide (09A, 09B, 09C) offense segment data fields are complete.

Click here to download SPSS code
Download sample data

For incidents that involve the homicide offenses of murder and nonnegligent manslaughter, negligent homicide, or justifiable homicide, are the required offense segment data fields completed and are the values valid? The following code creates the variables necessary to perform the data quality check and creates frequency tables of the number of homicide offense codes, the number completed/attempted, and the type of weapons used. When using this code, be sure to insert the path and file name of the data file to be used. If you need any assistance working with the syntax provided, please contact us.

For the above-mentioned offenses, the following data fields are required: UCR offense code, offense attempted or completed, and type of weapon/force involved. The allowable values for these data fields are listed in the FBI's Data Collection Guidelines and are incorporated into this example. This data check can be performed on data sets at the local or state level. The following example uses a single agency data set. The results show which offense segment records need to be researched and corrected before submission at the state or national level. The narrative below explains the SPSS code used to prepare the file and to produce the lists and tables needed for quality control purposes. To illustrate how an agency can self-audit, one month's data from one of the NIBRS-certified localities reporting to the FBI in 1999 was modified.


Defining Homicide Offenses and Selecting Records for Analysis

The block of code below retrieves and reviews the offense segment records to identify those with a homicide offense. The FILTER selects only those offense segment records with any one of the three homicide offenses (murder and nonnegligent manslaughter, or 09A; negligent homicide, or 09B; or justifiable homicide, or 09C).

GET FILE = 'Directory:\Path\Offense Segment.sav'.
COMPUTE filter_$ =(off_code = '09A' or off_code = '09B' or off_code = '09C').
VARIABLE LABEL filter_$ "off_code = '09A' or off_code = '09B' or off_code = '09C' (FILTER)".
VALUE LABELS filter_$ 0 'Not Selected' 1 'Selected'.
FORMAT filter_$ (f1.0).
FILTER BY filter_$.

Defining Value Labels for Required Variables

Next, the valid value labels for offense code, attempted/completed, and weapon type are assigned.

VALUE LABELS
  off_code
    '09A' 'Nonnegligent Homicide'
    '09B' 'Negligent Homicide'
    '09C' 'Justifiable Homicide'/

  att_comp
    'A' 'Attempted'
    'C' 'Completed'/

  weapon1
    11 'Firearm'
    12 'Handgun'
    13 'Rifle'
    14 'Shotgun'
    15 'Other Firearm'
    20 'Knife/Cutting Instrument'
    30 'Blunt Object'
    35 'Motor Vehicle'
    40 'Personal Weapons'
    50 'Poison'
    60 'Explosives'
    65 'Fire/Incendiary Device'
    70 'Drugs/Narcotics/Sleeping Pills
    85 'Asphyxiation'
    90 'Other'
    95 'Unknown'
    99 'None'.

Generating Lists and Tables

For the selected records, a FREQUENCIES table is requested for the UCR offense code, offense attempted or completed, and type of weapon/force involved. The frequency distribution for UCR offense codes 09A, 09B, and 09C will provide the total number of records to be evaluated. For these three homicide offenses, the data element offenses attempted/completed must be coded as completed (C). A frequency distribution on ATT_COMP will show the joint distribution for the selected homicide offense records and offense attempted/completed. A visual scan of the table will show whether all relevant records were coded with the only valid code (C).

Up to three types of weapons/force involved can be recorded for each offense. For this example, only the first occurrence of this variable -- WEAPON1 -- is checked for the presence of data. If there is a valid value for the data element WEAPON1, then the requirement is met.

Finally, the TEMPORARY / SELECT IF command searches for homicide offense records for which data are missing from the two conditionally required fields - ATT_COMP and WEAPON1. LIST requests the display of the values for originating agency identifier, incident number, UCR offense code, offense attempted or completed, and type of weapon/force involved (first occurrence as shown in variable WEAPON1) so that any records missing a value in ATT_COMP and/or WEAPON1 are identified so that they can be researched and corrected.

FREQUENCIES
   VARIABLES=off_code att_comp weapon1.
   TEMPORARY.
SELECT IF ((off_code = '09A' or off_code = '09B' or off_code = '09C') AND (att_comp = ' ' or weapon1 = ' ')).
LIST ori inc_num off_code att_comp weapon1.

Reviewing the Lists and Tables

The first frequency table produced is the offense code table, which provides the total number of relevant records (680) for the data quality check.  There are 680 records for which an offense code of murder and nonnegligent homicide (09A) or negligent homicide (09B) or justifiable homicide (09C) is present.



The second table, the offense attempted/completed table, shows that 680 records have a valid value for this conditionally required data field. Therefore, all 680 records were coded correctly. Any records with missing data in either of these two fields would have been identified by the originating agency identifier (ORI) and incident number, so that these incidents can be researched and corrected.

The final table, the type of weapon/force involved, shows that 680 records have a value for this conditionally required data field. All 680 records were coded correctly in that at least one code was recorded. A visual scan of the codes in the first column confirms that all values are valid. Any records with missing data in these fields would have been identified by the originating agency identifier (ORI) and incident number, so that these incidents can be researched and corrected.