A table describing each FARS variable name, value, and corresponding value label.
Format
A data frame with 132,454 rows and 8 variables:
- source
The source of the data (either FARS or GES/CRSS)
- years
Years of the data element definition.
- file
The data file that contains the given variable.
- name_ncsa
The original name of the data element.
- name_rfars
The modified data element name used in rfars
- label
The label of the data element itself (not its constituent values).
- value
The original value of the data element.
- value_label
The de-coded value label.
Details
This codebook serves as a useful reference for researchers using FARS data. The 'source' variable is intended to help combine with the gescrss_codebook. Data elements are relatively stable but are occasionally discontinued, created anew, or modified. The 'year' variable helps indicate the availability of data elements, and differentiates between different definitions over time. Users should always check for discontinuities when tabulating cases.
The 'file' variable indicates the file in which the given data element originally appeared. Here, files refers to the SAS files downloaded from NHTSA. Most data elements stayed in their original file. Those that did not were moved to the multi_ files. For example, 'weather' originates from the 'accident' file, but appears in the multi_acc data object created by rfars.
The 'name_ncsa' variable describes the data element's name as assigned by NCSA (the organization within NHTSA that manages the database). To maximize compatibility between years and ease of use for programming, 'name_rfars' provides a cleaned naming convention (via janitor::clean_names()). Both names are provided here to help users find the corresponding entry in the Analytical User’s Manual but only the latter are used in the data produced by get_fars().
Each data element has a 'label', a more human-readable version of the element names. For example, the label for 'road_fnc' is 'Roadway Function Class'. These are not definitions but may provide enough information to help users conduct their analysis. Consult the Analytical User’s Manual for definitions and further details.
Each data element has multiple 'value'-'value_label' pairs: 'value' represents the original, non-human-readable value (usually a number), and 'value_label' represents the corresponding text value. For example, for 'road_fnc', 1 (the 'value') corresponds to 'Rural-Principal Arterial-Interstate' (the 'value_label'), 2 corresponds to 'Rural-Principal Arterial-Other', etc.