 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
 |
|
|
|
NCHS Home | NHIS Home | Questionnaires, Datasets, and Related Documentation | 1997-2008 Data and Related Documentation | 1996-Prior Data and Related Documentation | NHIS on Disability | Coming Events | Methods | Reports | Data Linked to NHIS Data | Related Sites | CDC/NCHS Privacy Policy Notice | Accessibility | Search NCHS | Contact us |
6.2. Tobacco Related Codebooks
This section provides general and detailed information for data analysts who need to understand the changes to the tobacco related codebooks over the years. See Section 6.2.2 for a list of codebook files. See Section 6.2.3 for the web links.
6.2.1 Overview of the NHIS Tobacco Codebook Files
In order to analyze data on tobacco use, the analyst needs to consult the NHIS codebooks for the desired years to determined which variables to use in the analysis. The NHIS codebook format has undergone major changes over the years, so the essential information is provided here to help the analyst.
NHIS codebooks are abbreviated documents that present condensed information from the questionnaire and the dataset to assist the analyst in selecting variables and writing computer programs to analyze data using standard statistical packages such as SAS, SPSS, SUDAAN, or STATA. Synonyms for codebook used in the NHIS documentation include: file layout, tape layout, and data dictionary.
Note that there is not always a one-to-one relationship between a question in the questionnaire and a variable in the public use dataset. For some questions, there are several variables, either because more than one answer can be selected, or because recodes were computed. Alternatively, a given recode may be a composite of several variables. For some questions, the answers are not released in the NHIS public use dataset to protect confidentiality, but recoded information can be released. There can be multiple files for each NHIS data year and there is a corresponding codebook for each data file.
Generally, a NHIS codebook contains the following components:
For the
whole dataset:
The
name and data year for the dataset.
The
sample size.
For each
variable:
A
descriptive label for each variable. This label is usually an abbreviation
of the question from the questionnaire. It could also be a brief
description of a recode or computed variable (variable label).
The
location of each variable (inclusive column numbers)
A
list of answer codes for categorical variables (value).
A
label for each categorical code (value label).
A
range of answers for continuous variables (such as height or weight)
The
unweighted frequencies for each value of each variable in that dataset so
that the analyst can check the accuracy of his/her output.
In
recent years, the codebook may also include variable names that follow the
SAS naming conventions (8 characters or less; starts with a letter; contains
only numbers, letters and a limited number of special characters; does not
include any embedded spaces.)
Variable
names in NHIS Codebooks:
Starting
in 1997, when the NHIS switched from paper forms to computer assisted
personal interviewing (CAPI), each NHIS variable and recode was assigned a
SAS compatible variable name (8 characters or less, starting with an
alphabetic character).
Some
supplements from 1996 and before years were retroactively assigned variable
names when they were released for public use (such as the 1992 Cancer
Control and Epidemiology files). The HIS core files from 1996 and before
continued to list the column numbers, not the variable names in identifying
the location of each field.
Major Changes to the NHIS Codebook Format:
1965-1996: The NHIS codebooks from these years included: sample size, frequencies, column numbers, variable labels, answer codes, and value labels. These codebooks did not usually include variable names.
1997-2003: With the advent of the 1997 questionnaire redesign, variable names were systematically created for each variable, and included for every variable in each codebook.
2004 to present: Beginning with the 2004 data year NHIS codebook information can be found in the following files:
Variable
summary:
This file contains the question number, the variable name, the variable
label, the column location and the column length (width).
Variable
layout:
This file contains the question number, the variable name, the value codes
and the value labels.
Variable
frequencies:
This file contains the question number, the variable name, unweighted
frequencies and percentages for the categorical variables in each file. The
responses for certain continuous variables, such as age, are given in broad
categories.
Return
to Introduction
Return
to Detailed Outline
This page last reviewed October 15, 2008
|
