Practical Data Management and Statistical Computing (BioEp691F)

Contacts

Outline

Assignments

Resources

---

Outline: Lec21 Lec22 Lec23 lec24 Lec25
Lectures: Lec21 Lec22 Lec23 Lec24 Lec25

---

Lecture 21

A coding manual or codebook completely describes the data and how it was processed. It is the last step in data management in a study, and serves as a reference for analyses. The codebook includes:

• Brief summary of data collection protocol, pre-processing protocol, and questionnaire with variable names.
• Data entry protocol for the study, including special files for entry and verification.
• Special variable creation and coding decisions.
• Contents and names of ASCII and SAS data sets.
• Optional - Proc FREQ (Univariate) for all variables.
• Optional - Printed list,
• ASCII Data set and list
• Names and locations of all data sets 

Pre-processing

Review questionnaires, enter ID number if missing if adjacent forms match sex and birthdate, verify that circle codes are blackened when response is given.

Data Entry Protocol

• Order questionnaires by ID number from smallest to largest.
• Enter data beginning with ID, study date, date of birth, gender, age, and then other questions.
• Enter a 2 digit ID, with leading "0".
• Use a single character for sex (m or f).
• Enter leading "0" for leading blanks with dates, with no "/". All should be 6 columns.
• Enter a 2 digit age, including a leading "0" if necessary.
• Leave missing values blank.
• Verify all data entry (with EpiInfo).

(see LEC22P1.SAS for example)

Special Variable Creation/Coding Decisions

• Check age calculation, and replace with age based on study dates and birthdate if different. Create list of ages where there is a miss-match, and manually examine them.
• Check for consistency of NA responses for smoking.(see LEC22P2.SAS and output).

A Study WEB site is a valuable way to document data. Such a site is

• permanent
• organized
• easy to find

The WEB site need not be elaborate. A simple site will consist of the following:

• Site home page
  • Title
  • Brief (3 line) description
  • Links to:
    • Instrument (pdf)
    • Codebook Page
      • Variable definitions
      • Special considerations
    • Data Page
      • Diagrams of data set creation
      • Proc contents of data set
      • Link to:
        • SAS Programs
        • Entry programs
    • More Details Page
  • Diagram of the site

Example