Section 7 Codebook

For proper use of the ELSOC COES database, we recommend that researchers work with the Codebook (presented in this section) and the Variable List (downloadable in excel format at the following link, or on the COES Website).

The Codebook and the Variable List shows the longitudinal code of each variable; the associated variable label; the phrasing of each preamble, question and item; the different response categories and associated codes; and observations on the variables that have undergone modifications throughout the measurements of the study. In addition, the Codebook includes frequency tables for each question according to wave and sample (continuous and text variables are omitted).

We designed the codebook to summarize all relevant information about the variables in the database into a standard format for ease of use. Generally, the variables included in the database take the following form:

Longitudinal Code. Variable label

Question phrasing

Response Codes of response categories

Notes

Frequency table

We present the questions by questionnaire module to facilitate the understanding.

The Longitudinal Code is associated with each questionnaire item in the database and codebook. Through these codes, we can identify different items. In the wide form longitudinal database, the code includes *_w01, _w02, _w03, _w04* or *_w05* at the end to denote whether the variable corresponds to waves 2016, 2017, 2018, 2019 or 2021, respectively.

We included variable labels in the Codebook and the databases. The ELSOC team designed them intending to briefly describe the phenomenon or dimension to be measured7. The question’s phrasing follows the labels, including preambles, response codes and categories. In constructing the database, we entered response codes as numerical values and response categories as labels.

Finally, we included observations associated with possible changes over time or aspects to be considered when using the variables.

The string variables (text) do not present codes since they are literal verbal responses of the interviewers. There are no response categories in items where we request a numerical answer because we record the value indicated by the respondent.

  1. We eliminated accents and other symbols not included in all statistical software (e.g. accents and ñ)↩︎