A codebook is a document that describes the structure and content of a research data file. The term codex is also often used.

A codebook conveniently consolidates all data from a research study, facilitating the preparation of data entry, processing, interpretation of processing results, and subsequent analysis. This document is a useful tool for understanding between researchers and between researchers and statisticians. Thanks to the codebook, it is possible to use the same data by other researchers for an unlimited period of time, such as, for example, in the analysis of secondary data or meta-analysis.

The codebook should include:

  • document description,
  • research description,
  • data file description and
  • variable description.

The document description should include information about the author and institution, address, etc.

In addition to the research title, the research description should include:

  • brief description of the project and research,
  • time and geographical determinant of the research,
  • population description,
  • sample description,
  • unit of analysis and unit of observation and
  • instruments used in the research.

The most extensive part of the code is the description of variables.

Questionnaires and other instruments are most often unsuitable for describing variables because they are question-oriented, while in the code the basic unit is the variable. Questionnaires are constructed so that they can be understood by the respondent, and the code by researchers and statisticians.

Instruments often do not contain all the information used in the research and are not easy to understand for data entry, so they are not used as a codebook.

The following table shows an example of a description of one variable.

Variable Question Column/Position Text Modality
1    1  1  Ord. number 1 - 189
2 1  2  Gender

1 - male
2 - female
3 2  3  Success at the end of the previous year of study

2 - satisfactory
3 - good
4 - very good
5 - excellent
4    4    
5    5    

The first column lists the ordinal numbers of all variables used in the study. The rules state that all data entered about a single respondent is treated as a variable.

The second column contains the data needed to identify the variable, such as the question number in the questionnaire.

The third column contains data about the column or position that this variable occupies in the code sheet.

The text of the question or a detailed description of the variable is in the fourth column.

The fifth column shows the content of the variable, i.e. all the values that the variable can have in that study. The rules state that the respondents' answers are numerically coded, which is why the values on nominal and ordinal variables are explicitly marked. No special marking is required for continuous variables.

Depending on the research, the codebook can be expanded. For example, if multiple instruments were used in the research, a column for the instrument label is introduced. The column with the variable position can be removed.