Quality check responses and flag outliers

Updated at August 10th, 2022

Protobi has a built-in quality check or "QC" tool that allows you to identify outliers and flag respondents with low quality responses. 

Open the QC dialog

Select an element and press Shift + Q + C on your keyboard to bring up the QC dialog.

You can also select the element, then go to the "Advanced" button. Scroll to find the "QC check (beta)" option.

QC a single element

Single elements have three pre-set QC options. These checks are usually applied to numeric questions where you want to flag and cap outlier responses. 

QC checks are stored in a special element

After you add a QC check, it creates a new element called "$flags". Every time you add a new QC check, it's added to this element. 3% of respondents are flagged as Q1 > 50 and those that do not meet the criteria are shown as [NA]. 

QC a group

There are more QC options for a group of questions, for instance there is a preset to check for straight-liners.  

Straight-line check

Select "All children are equal (aka straightline)" to flag respondents that entered the same answer for all questions in a group. 

Result:

Remove flagged respondents

Shift + press on flagged respondents to exclude them from the sample. Then set a permanent filter on the question to exclude the flagged responses.


Create your own custom flags

You can define a condition to create custom flags. Q3 asks how many Condition Y patients are on a gamma antagonist, and S8 asks how many Condition Y patients the respondent manages. 


You can set a custom condition that flags respondents that entered a greater value for Q3 than S8. Custom conditions can reference any element, you can even create the QC in an element, and not include it in the condition. 

Javascript operators 

The custom condition text box will accept conditional statements written with Javascript operators. Below is a list of to create operators you can use to create custom conditions. 

Arithmetic

+

Addition
- Subtraction
/ Division
* Multiplication

Comparison

== Equal
!= Not equal
> Greater than
>= Greater than or equal to
< Less than
<= Less than or equal to
=== Equal to, and the same data type (Identical)
!== Not identical

Logical 

&& AND
|| OR


Advanced support

There are complex cases where you might want to flag respondents, not limited to what's shown in this tutorial. Contact us at support@protobi.com to discuss your specific needs.

apply the filter

Was this article helpful?