Documentation Tools in the Survey Lifecycle

1
Documentation Tools in the Survey Lifecycle
2
Outline

• What is NSFG Webdoc?
• Instrument documentation ! Survey documentation
• Data Cleaning/Processing in the Data Lifecycle
• Unexpected dividends Documentation Tool as
  Management Tool

3
National Survey of Family Growth, Cycle 6

• Conducted by the U.S. National Center for Health
  Statistics of the Department of Health and Human
  Services in 2002

• Cases
• 7643 female cases (age 15-44)
• 4928 male cases
• 13593 pregnancy cases

• 5300 variables
• Intended and unintended births
• Sexual intercourse
• Contraceptive use
• Infertility
• HIV testing

4
NSFG WebDoc

• Previously, for NSFG Cycle 5
• Codebook 6000 pages
• Questionnaires 600 pages
• User Guide 450 pages
• Features of WebDoc for NSFG Cycle 6
• Web browsable codebook
• PDFs for printable codebooks
• Search functionality
• (based on Oracle Text Indexing)

5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
(No Transcript)
9
WebDoc Components

• Document Assembly
• Component
• DDI from Blaise
• Word documents
• Excel spreadsheets
• Text documents
• SAS data definition statements
• SAS output

• Web Display
• Component
• Cocoon
• XSLT

80
20
10
Instrument Documentation
Survey Documentation
?

• Blaise instrument
• documentation
• Female questionnaire
• Male questionnaire
• 800 questions

• Final public release
• 3 data files (female, male , pregnancy)
• 5300 variables
• Raw
• Computed
• Recodes
• Imputation flags
• Weights
• intermediate

• Inhouse data files
• full data files
• Interviewer observations
• Audio CASI files

11
Documentation in the Lifecycle
Data Archiving
Study Design
Data Collection
Data Processing
Data Distribution
Data Discovery
Data Analysis
Instrument Doc
Survey Doc
12
Consistent Codes for DK, RF, NA
Confidentiality Review
Consistency Checks
Perturbation
Multiple regression Logical Imputation
Recodes (10 levels)
How do you manage this process?
A Documentation Tool can help!!
13
Unexpected Dividend Documentation Tool as
Management/Diagnostic Aid

• WebDocs was started near the end of the Data
  Collection Phase of the NSFG project.
• A working version of WebDocs was finished early
  in the Data Processing/Cleaning Phase.
• Small unplanned modifications to the working
  prototype was used to produce diagnostics/metrics
  of the data cleaning process.

14
Documentation in the Lifecycle
Data Archiving
Study Design
Data Collection
Data Processing
Data Distribution
Data Discovery
Data Analysis
Survey Doc
15
WebDoc as diagnostic tool

• Features
• Identifies problems
• Missing Documentation Pieces question text,
  description,
• Editing errors missing labels, misspelled
  variable names,
• Data errors missing values, out-of-range values
  (incorrect computation or recode logic),
  inconsistent missing codes, SYS-MIS
• Wiki used to track comments on each variable
  (almost zero programming necessary)

16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
Documentation Tool as a Diagnostic/Management
Tool

• Do not treat documentation as an afterthought.
• Have the documentation tool available early at
  the beginning of the phase in order to use it as
  a diagnostic/management tool to manage the
  process
• Metadata in a database can be repurposed.
• Another example SRCs BlaiseDoc
• Can be used to produce HTML Questionnaire with
  programmed skip patterns
• HTML Questionnaire can be used by non-programmers
  to quickly check the work of instrument
  programming.

21
(No Transcript)
22
Documentation in the Lifecycle
Data Archiving
Study Design
Data Collection
Data Processing
Data Distribution
Data Discovery
Data Analysis
Survey Doc
Instrument Doc
Instrument Doc
23
Summary

• Make Documentation Tools available early.
• Use a database so documentation metadata can be
  easily repurposed.
• Documentation tools can be used by
  non-programmers to inspect/manage the product
  quality and the process.
• NSFG WebDoc
• http//www.icpsr.umich.edu/NSFG