DAISIE logoThe basics of survey processing > Variables French site English site

The management of variables means that the information is presented exactly as it should be

Variables

The data is organised in “variables”. Descriptive statistics and data analysis depend on variable kind. A variable can be:

  • Continuous
  • Discrete
  • Ordered
  • Textual

DAISIE offers great flexibility in survey processing because it is able to identify what kind of variable it is and take account of this in the processing. In all calculations and analyses, DAISIE offers the possibility to change what kind of variable it is using what we call “recoding” functions. Hence for example, the continuous variable ‘Age’ can be turned into an ordered variable by putting the ages into groups. All variables can be turned from one kind into another.

The particularities of survey variables

When a variable includes some No answers, a decision has be taken by the user as to how to deal with them. DAISIE automatically isolates the no answers of a variable. It is therefore easy for the user to decide what to do with them: add them on as an extra modality, assimilate them with another response, assign them the mean of all responses, etc... or simply take them out of the statistical calculation bases. Good management of no answers is essential for meaningful statistical results.
Survey variables also have some other particularities. Multi-response variables and batteries of scores can be evoked among other things. An automated approach to these characteristics means not only comfort for the user, but also greater security as to the quality of the statistical results. For example, with a multi-response variable, what base should be chosen: Respondents, Interviewees or Total responses? You have to at least be able to choose your calculation base.

Working variables

The variables observed are rarely raw variables, i.e. variables that come straight from the questionnaire. They are more likely to be what we call “working variables”, obtained by recoding one or several variables.

Examples of working variables:

  • In a question about products, the creation of families of products by grouping the products together, each family being followed by the products contained in it
  • In a satisfaction scale, adding in the modalities “Total satisfied” and “Total not satisfied”, but without altering the “ordered” nature of the result
  • In a question about allocated budgets, translating local currencies into Euros and / or putting ‘outlandish’ values in no answer
  • In a question about preferences, organising the variable in terms of order of preference
  • And anything else you can imagine...
 In DAISIE, the possibilities for creating working variables are limitless.

Example of a working variable

A working variable represents the information in its ideal form. In this example, we have chosen to group answers to the question “What European countries have you visited?” by the main geographical areas:

  • Western Europe
  • Southern Europe
  • Eastern Europe

The Western European countries are grouped into geographical sub-zones:

  • Benelux
  • Scandinavia
  • Other countries

Simple writing enables all subtotals to be inserted and the modalities to be grouped by geographical zone.

NB.: DAISIE automatically identifies no answers, respondents, interviewees and total responses. The latter includes the totals and subtotals which are now responses, but DAISIE also knows how to identify the real number of multi-responses.

Working variable