The 2006 National Health Interview Survey NHIS Paradata File: Overview And Applications

1
The 2006 National Health Interview Survey (NHIS)
Paradata File Overview And Applications

• Beth L. Taylor
• 2008 NCHS Data Users Conference
• August 13th, 2008

2
2006 Paradata File

• Paradata is information about the data
• collection process
• Data collected during and immediately after the
  NHIS interview
• Part of the NHIS annual data release

3
National Health Interview Survey

• NHIS Annual survey of the civilian,
  non-institutionalized
• population of the U.S. sponsored by the
  National Center for
• Health Statistics
• Approximately 35,000 families interviewed
  annually
• In-person interview with telephone follow-up
  allowed
• 4 ongoing core modules (Household, Family, Sample
  Child, Sample
• Adult) plus annual supplements
• Census Bureau (contractor) carries out field work
  for
• Sponsor NCHS

4
Sources of Paradata

• Contact History Instrument (CHI)
• - Introduced in 2004
• - Produced by the Census Bureau
• - Used on other Census surveys
• - Launches each time interviewer accesses the
  CAPI instrument
• - Captures data on each visit attempt
• - In-scope responding and nonresponding
  households
• - Out-of-scope households

5
Contact History Instrument

• Captures whether attempt via phone or personal
• visit
• Contact path
• - Description of contact attempt
• - Reluctance
• - Strategies used to complete interviews
• Noncontact path
• - Description of attempt
• - Strategies

6
Contact History Instrument

• Examples of Strategies
• - Advance Letter given
• - Scheduled appointment
• - Left note/appointment card
• - Staked-out household
• - Checked with neighbors

7
Contact History Instrument

• CHI data
• - Family-level file
• - Summary of contacts, noncontacts, strategies
  used during interview period
• - Attempt-level file
• - Date and time of contact attempt
• - Description of contact, noncontact,
  reluctance, strategies used for that visit
  attempt

8
Sources of NHIS Paradata

• Front/Back sections of survey instrument
• - Present on NHIS since late 1990s
• - Tailored to NHIS
• - Language of interview
• - Cooperativeness of respondent
• - Mode of interview (personal visit or phone)
• - Reasons for partial/break-off interviews
• - Type of noninterview case

9
Sources of NHIS Paradata

• Time file
• - Interview times (length of interview
  start-finish)
• - Module/Section times (time of 4 core modules
  and within-module sections)
• Audit trails
• - Record of keystrokes
• - Field times (length of time per question)
• - Dates
• - Interviewer notes

10
Paradata File Creation

• In early 2007 NHIS Paradata Committee
• established at NCHS to
• - Select variables to include on the initial
• Paradata File release
• - Weighed issues of confidentiality, usability
• - Example Created numerous recodes for CHI file
  variables

11
Paradata File Release

• First NHIS Paradata File released in
• January, 2008 on the Internet (2006 NHIS data)
• 125 Variables
• - Family-level file (one record equals one case)
• Dual uses
• - Can be analyzed alone or linked to the 2006
  health data files

12
Table 1 2006 Paradata File Frequency
Distribution of Cases by Outcome
13
2006 Paradata File

• Internet Release
• Paradata File Description Document
• Variable Summary
• Variable Layout
• Dataset
• Sample SAS Input Program
• Variable Frequencies

14
2006 Paradata File Release

• File Description Document
• - Sample design
• - Weighting and Variance Estimation
• - Conceptual grouping of variables
• - Measures of time, contactability, cooperation
• - Contact strategies
• - Partials/Break-offs
• - Mode measures
• - Case-level information

15
2006 Paradata File Release

• Variable summary
• - Variable name and brief description of each
  variable on the file
• Variable layout
• - More detailed description
• - Variable universe
• - Source (Contact History File, etc.)
• - Question/Response codes from survey instrument
  (if applicable)
• - Notes, recode information

16
2006 Paradata File Release

• ASCII dataset (PARADATA.EXE)
• Sample SAS input program
• - Formats
• Variable frequencies

17
Uses of the 2006 Paradata File

• Stand alone data file (Examples)
• - The average number of contact attempts it
  takes to complete the interview
• - Point in the interview period and time of day
  when the Sample Adult module was started (Early,
  Middle, Late) / (Morning, Afternoon, Evening)
• - Which strategies employed by the interviewer
  led to successful completion of the interview

18
Uses of the 2006 Paradata File

• Linked to 2006 health microdata (merge with NHIS
• Sample Adult File, Family File, etc.)
• - Item of note Will not have a 11 match with
  health files because Paradata File has
  information on nonresponding out of scope cases
• - Paradata File has a slightly larger number of
  
• interviewed cases than health files
• - Paradata is field data before cleaning

19
Uses of the 2006 Paradata File

• Examples of possible research using linked
• files
• - Characterizing the characteristics of
  hard-to-contact families
• - Modeling the impact of the interview mode on
  health outcomes

20
Weighting

• When using the Paradata File as a stand-alone
• file, use the WTIA_PD variable if making
• population inferences
• - This weight reflects the probability of
  household selection
• When using the Paradata File to support analyses
• with the health data files, use the weight
  variable from the
• health data file
• - Example Use WTFA_SA when merging with the
  Sample Adult File

21
Variance Estimation

• NHIS data are obtained through a complex sample
• design involving stratification, clustering, and
• multistage sampling
• - Recommended that users utilize computer
  software that provides the capability of variance
  estimation and hypothesis testing for complex
  sample designs (SUDAAN)
• - PSU and STRATUM variables included on Paradata
  File

22
Current NCHS Research Using NHIS Paradata

• Assessing data quality from the field
• - Started Quality Assurance Workgroup jointly
  with Census Bureau to monitor data quality
• Upcoming presentations

23
Future Steps

• 2007 Paradata File released in June, 2008
• - Nearly identical to 2006 Paradata File release
• - Telephone items revised for 2007
• Expansion for 2008 NHIS Paradata Release
• - To include data from visit-attempt file,
  possibly audit trail data (use of function keys,
  language of interview)
• Feedback from data users

24
NHIS Paradata File

• For more information about the NHIS Paradata
  File, please visit
• http//www.cdc.gov/nchs/about/major/
• nhis/2006paradata.htm