Using the NCVS Interactive Code Builder

The Interactive Code Builder provides an easy means of selecting specific variables, subsetting cases, and structuring an analysis file at the chosen unit of analysis. Unlike most datasets, the NCVS full (hierarchical) file requires that these tasks be performed for each specific data analysis. The statistical programming for these tasks can be done manually; however, it is a time consuming process, and the potential for error is high. The Interactive Code Builder was developed to expedite the creation of specifically constructed files for statistical analysis. The goals are to free researchers from the burden manipulating data files and allow them to spend more time on analysis.

Essentially the NCVS Code Builder is a series of choices that the user makes. The product is an SPSS syntax file. The user copies the syntax file from the Web into an SPSS syntax window and runs it on a desktop computer. An SPSS system file structured according to the user's parameters is save on the user's computer. The file is ready for statistical analysis.

Users need to complete two steps before the Code Builder can be fully utilized. (The Interactive Code Builder can of course be used before this, but no data files will be created.)

Step 1 - Get the Data

NCVS Data are available at no charge from the National Archive of Criminal Justice Data. The dataset contains several different ASCII data files; for each data year there is a full (or hierarchical) file, an incident-level file, an incident-level rape subset, and a concatenated incident-level file. This code builder utilizes the full (hierarchical) file.

Step 2 - Use the Code Builder

After the NCVS data have been downloaded and saved locally, it is a simple process to create custom analysis files. The NCVS Code Builder prompts the user with a series of questions and instructions.

First the user selects all of the variables of interest. The variables are listed by NCVS record type.

Next, based on the variables selected, the user picks the desired level of analysis (ie, at the household, person or incident level).

Depending upon the variables and unit of analysis chosen, the user chooses how many related records to read.

NCVS can record up to 13 persons per household and 18 incidents per person. The number you may choose is limited based on how many records actually appear in the data of the year you have chosen. Users should realize that the number of variables in the final file can increase dramatically. For example, suppose that a household-level file is being created, and for that particular year there are a maximum of 10 people in a household with a maximum of 7 incidents per person. Further, suppose that only 3 variables were originally selected:

V2021 - TYPE OF LIVING QUARTERS (ALLOCATED)
V3014 - AGE (ALLOCATED)
V4011 - SCREEN Q. NO. WHICH PROMPTED REPORT

Selecting these 3 variables with the maximum number of related records results in 81 variables in the final analysis file:

V2021 X 1 time = 1 variable
V3014 X 10 times = 10 variables
V4011 X (10 X 7) times = 70 variables

Choosing to create a file with the maximum number of related records will result in a large file that has mostly missing data.

Next, the user selects whether to subset cases. Subsetting is available by NCVS crime type.

The above choices are submitted, and in a couple of seconds an SPSS syntax file is displayed.

This syntax file is the product of the Interactive Code Builder. It is a file that can be used to create an SPSS system file built to the user supplied specifications. This file should be copied into an SPSS syntax window and run. The path and data file name may need to be edited. An SPSS system file will be saved onto the user's PC.

Use the Interactive Code Builder