Filters can be used to define only specific parts from your data set to be recoded. This can be defined by:
- Categorical variables
- Open ended variables
- Numeric variables
The data can be filtered before recoding, so we only recode specific parts of the data. Within Filter you can select the right group and your existing answers will appear. See image below: here we could recode for example only the data of one or multiple cities.
Open ended variables
The data can be filtered before recoding on any open ended variable. You can also define if the text should match the whole string or just part of the string.
Example: The data for respondent id 1000 is firstname.lastname@example.org in the Email variable (an open-ended text variable). If the filter is defined as “dapresy.com” and “Match part of the string” is applied this respondent will fulfil the filter criteria, if the option “Match whole string” is applied it will not fulfil the filter criteria.
Here we see an example of using an open-ended variable as filter in a recoding rule.
Open numeric variables
Open numeric variables can be used as filters when setting up recoding rules. This allows easy correction of data as a data set does not need to be corrected outside Dapresy Pro in cases of having to update a small subset of the data.
When using Open numeric variables as the filter you can select to add a range or single values. If single values are used you can enter multiple values separated by a semi-colon (;). Both options are shown in the example image below. Here we see an example of using an Open Numeric variable as filter in a recoding rule, as shown the option Single value has been selected.
Here we see an example of using an Open Numeric variable as filter in a recoding rule, as shown the option Range has been selected.
Besides the regular survey variables, RespondentID and ResponseDate can be used to filter on, to do really specific data corrections on a respondent level or time period level. Filtering on RespondentID works in the same way as an open numeric. When using ResponseDate as the filter in a recoding rule you can select between five different options as shown in the image below, all of these are described further down. Here we see the five different options that are available when using ResponseDate as filter when in a recoding rule.
Range is used to recode respondents that belong to a certain time period, a start and stop date is defined. Here we see the setup of a range, a start and stop date is defined. The format for the date in the field should be YYYY-MM-DD.
ResponseDates before certain date
ResponseDate before a certain date is used to recode respondents until a certain date. If the date for example is set to 2017-12-01 all respondents with an older date will be recoded.
ResponseDates after certain date
ResponseDate after a certain date is used to recode respondents after a certain date. If the date for example is set to 2017-12-01 all respondents with a newer date will be recoded.
ResponseDates older than X days
This is a floating option which recodes respondents that have an interview date older than x number of days from the data activation date.
ResponseDates newer than X days
This is also a floating option which recodes respondents that have an interview date newer than x number of days from the data activation date.
Note: it is not possible to recode the question type "Date".