Protobi can handle complex structures in survey data. This tutorial provides a quick intro for common cases.

The Eurobarometer Robots survey dataset is also a great case study for importing data and advanced configuration options, as it codes collections of related items in the datafile in a way typical of some survey platforms.

Here, question Q3 is a check-all-that-apply question "QA3 Have you ever used, or are you currently using such robots at home or at work (e.g. a robotic vacuum cleaner at home or an industrial robot at work)?"

In the survey, respondents were able to select four response items:

  • "Yes, at home"
  • "Yes, at work"
  • "Yes, others"
  • "No"

But in the dataset, all four fields have identical variable labels, corresponding to the collection. Here's what it looked when we first imported the data file (after running Auto Group):

Extract common text

To standardize this, click the square icon for the parent group qa3 and select "Extract common text...". This will pull shared text from the child elements into the parent group.

Protobi will recognize they all start with the same text, so the default choice is fine. This will extract the text from the individual fields and make that the title for the parent group:

But now the fields have no title text at all. The only way to know what each option is is in the value formats.

Extract titles from formats

So again, click the square icon for qa3, select "Extract titles from formats..." and enter the value `1` (if you toggle formats off/on you'll see that the selections are coded as 1, with different value formats). It now looks as follows:

Compact display

Now we can compact the display nicely. Again, click the square icon for qa3 and select "Compact to..." and enter `1` which is the unformatted value for selected items.

Each element is now displayed as a concise line item showing the percent of people who selected it in the survey: