Variables (questions respondents do not see) are used to collect data from closed- or open-end questions respondents do answer. This includes question types such as quotas, click balance quotas (CBQs), coded single select, coded multiple select, and coded number. These variables are generally asked for in the survey document to keep track of quota groups or other data collections.
For information on visible/respondent-facing questions and variables, see Questions and variables. Overviews of the most commonly used question types can be found in the Question types overview and Table types overview articles.
Quotas
Quotas (type: quotas
) control how many respondents in a particular group are allowed to take a survey. Respondents must meet certain requirements to qualify for a quota group. Some quotas limit the number of respondents that can answer specific questions, such as screening questions, while other quotas limit the number of respondents allowed to complete the full survey.
Click balance quotas
A click balance quota or cbq (type: click balance quota
) is a quota based on all incoming traffic, not just completes, and is used for representative sampling. The term "click balance" is used because the incoming traffic is often referred to as "clicks" by panel providers when respondents click on the survey invitation, and "balance" because the clicks are proportionally balanced to the research targets (e.g., Census data). Tracking clicks for some quotas is useful because clients may require a certain number of responses to the screening section, even if not all of those respondents complete the survey.
Coded variables
Coded variables collect data to be used elsewhere in the survey for either logic or reporting purposes and can be created via the question types coded single select, coded multiple select, and coded number, respectively. These question types are hidden/non-respondent facing variables derived from responses to other questions, and can work alongside both quotas and click balance quotas.
Both type: coded single select
and type: coded multiple select
operate as if selectby: condition
is the default. This can be overwritten as needed.
Coded number variables (type: coded number
) calculate input based on the logic expressions or math specified in a cvalue
tag.
Other types of variables
Survey variables
The invisible: y
tag hides a survey question from respondent view, thus making it a survey variable. The invisible field generates no page content. This type of question is generally not written into the questionnaire but is needed for programming purposes to make the survey function as intended. Therefore, while it is hidden from the respondent, it is a survey variable, and not a hidden question.
System variables
Often there is a need to define variables within the survey source file that:
- Are not part of the questionnaire. If not part of the questionnaire, then it's unlikely the survey designer will want them in the final data output.
- Do not provide any useful data for reporting survey results. Perhaps the variable is used for controlling a skip pattern, calculating a value for use elsewhere in the survey, or even determining which text or content to show on screen.
These variables are called system variables and often need to be hidden from the data output and reporting. To denote a variable as a system variable, use the tag systemvar
. When systemvar: y
is used, the variable is placed within the System fields chapter in the Field Selector. System variables are excluded by default from any data exports and are not included in the data map layout for fixed field exports. You can apply the systemvar
tag to a question, class, or table, or include it under the set defaults
widget.
Comments
0 comments
Please sign in to leave a comment.