Find and identify straightline respondents
June 10, 2018
Straightliners. You know they must be somewhere in your sample … respondents who give the same answer to every question in a section.
If you could see the answers for one respondent for one section, it’d be easy to spot. But how do you quickly identify all straightlines? It’s pretty easy to find them in Protobi using this one trick…
Look for a standard deviation of zero
A straightliner is someone whose answers don’t vary. Standard deviation measures variation. So a straightliner is a respondent who has a standard deviation of zero.
To identify straightliner for a section, we just need to calculate the standard deviation of responses for that section for each respondent. And select those respondents with zero deviation.
Here’s an example from the Pew Research Gender and Generations survey. The graph below shows the distribution of all respondents.
We could click to find respondents who answered any one response in Q5a, Q5b, Q5, etc. But otherwise it’s not obvious who gave the same answer to each question.
Calculate standard deviation
Select “Transform…” from the context menu and choose “Standard deviation” from the dialog. Also deselect “Hide children”.
This will calculate the standard deviation for that section for each respondent and show the result as a distribution.
Here we’ve also set “Round by” to “log” to see the distribution better:
Select respondents with zero deviation
Here we can see there are 124 respondents who gave the same answer to each question in this section:
Here there are 25 respondents who said “Very strong conflicts” to every question in this section. And 8 who respondend “There are not conflicts “ to every question.
Just because they answered the same answers in one section doesn’t mean they didn’t mean it. They could be honestly giving the same answer.
But straightline answers are an important consideration in context. For instance, a respondent who completes a 45 minute questionnaire in 15 minutes and has several sections with flatline answers in more than one section would legitimately be suspect.
- Find a section of your survey where you’re concerned about straightliner.
- Set the Transform for that section to ‘Standard deviation’
- Click to select respondents with zero deviation
Then see if there are other aspects of their answers that would trigger ‘red’ or ‘yellow’ flags…
Want to discuss your survey? Contact our team at firstname.lastname@example.org