Start-List Coding with a Structured Codebook and Check-Coding

Description

Build a provisional code list from the framework and questions before fieldwork, define each code operationally, and keep it reliable through repeated check-coding as it evolves.

For Miles and Huberman 'coding is analysis'. They distinguish descriptive, interpretive and pattern codes and prefer a start list made before fieldwork from the conceptual framework, research questions and key variables: a dozen to 60 codes grouped under conceptual families, each with an operational definition and a mnemonic name rather than a number. Two coders code the same pages and compute agreement as agreements divided by agreements plus disagreements, expecting about 70% at first and aiming for about 90%. Each coder also re-codes early pages days later to check self-consistency. The codebook is then revised as codes decay, swell or emerge, while keeping a governing structure. Notes from each contact are coded before the next visit, so analysis steers later fieldwork. A flat alphabetical 'ragbag' list, or more than two coding days per contact day, signals trouble.

Attributes

Loss type
Vestigial
Effort
Weeks

Steps

  1. Derive a start list of a dozen to 60 codes under a few conceptual families, keyed to research questions, on one sheet.
  2. Write an operational definition for each code and give it a mnemonic name, not a number.
  3. Apply the list to the first field notes or transcripts and examine fit.
  4. Check-code: two people code the same 5-10 pages; compute agreements / (agreements + disagreements), expecting about 70% at first and aiming for 90%.
  5. Have each coder re-code early pages a few days later to check self-consistency.
  6. Revise the codebook: drop decaying codes, split swollen ones, add emergent ones, and keep the family structure.
  7. Code each contact's material before the next field contact, and repeat check-coding about two-thirds of the way through.

Sources

  • Miles, M. & Huberman, A. M., Qualitative Data Analysis: An Expanded Sourcebook, 2nd ed. (1994)

Also known as

  • Provisional start list of codes
  • Check-coding
  • Codebook with operational definitions

Categories

  • User Research
  • Qualitative Analysis & Synthesis