Codebook

A codebook explains the fields and coded values in a dataset. It connects short labels, numbers, or abbreviations with their intended meanings and often describes missing values and relevant collection rules. For public-safety data, it helps readers interpret the agency's codes without guessing from the characters alone.

A code needs its context

The Inter-university Consortium for Political and Social Research describes a codebook as documentation of a data collection's contents and structure. Its guidance includes variable names, labels, values, missing-data codes, and explanatory notes. A variable is simply a field describing a characteristic of each record. ICPSR: What is a codebook?.

An agency's call-type code and its disposition code may look similar while serving different purposes. The first can describe the kind of request; the second can describe how it was handled. The codebook identifies which field a definition belongs to.

Example

Hypothetical: A dataset has a field named result_code. Its codebook defines 1 as “report taken,” 2 as “referred,” and 9 as “not recorded.”

The numbers are labels, not quantities. Adding them or calculating their average would not produce a meaningful average result. Treating 9 as a severe outcome would also be wrong: the documentation says it represents missing information. These invented codes illustrate the need to read the actual source codebook.

Common confusion

A codebook and a data dictionary often overlap. A dictionary may emphasize field structure and allowed types, while a codebook may provide detailed value meanings and survey or collection context. Publishers use the names differently, so the contents matter more than the title.

A codebook also differs from a general glossary. A glossary explains a concept across contexts. A codebook defines how a particular dataset represents that concept. A local code can therefore require a narrower explanation than the general term.

Use the version that matches the data

Check the agency, system, field name, and effective period. Codes can be added, retired, or reassigned. Applying today's list to an older file can silently change the interpretation of historical records.

Read notes about blanks, unknown values, and inapplicable fields before calculating totals or percentages. If a code is absent from the documentation, keep it unresolved rather than assigning the closest-sounding meaning. A clear analysis can report that unknown code's frequency while preserving the distinction between verified categories and uncertain ones.

Related terms

• Data dictionary
• Metadata
• Not applicable
• Reclassification

Sources

• ICPSR, University of Michigan: What is a codebook? — Explains field-level documentation, coded values, missing-data labels, and contextual notes. The result codes here are hypothetical.