Sign in
x
☰

Blank, skipped, and missing values

Control how missing values appear

Updated at 2019-06-26 14:46:21 UTC
Decorative illustration showing three puzzle pieces - one purple-gray with speckled texture on the left, one white in the middle, and one blue-teal on the right. The pieces are arranged at angles with small decorative shapes scattered around them, representing the concept of missing or incomplete data.

Protobi uses the label "[NA]" to represent missing values.  This acronym can be interpreted as "Not available," "Not applicable," or "No answer."   

For some analyses it may make more sense to ignore missing values and consider only respondents who actually answered the question.  The [NA] button in the toolbar toggles whether missing values are shown in charts or not.

Missing values can occur when:

  • Someone skips a question due to a skip pattern
  • Survey logic doesn't force an answer
  • A profile variable has no recorded value
  • The analyst removes outliers for individual questions

Simple example

Consider this hypothetical survey with two questions:

  •  S1:  What is your medical specialty?
  •  S2: [IF NURSE] What grade nurse are you?

Question S2 is asked only of respondents who indicated they are Practice Nurse in question S1. 

Results would appear as follows, where 40% of all respondents are nurses, and 24% of all respondents are Nursing Equivalent band:

Protobi survey results showing two questions. Question S1 "What is your specialty" displays 60% General Practitioner and 40% Practice Nurse (N=100). Question S2 "What grade nurse are you?" shows [NA] at 60%, Equivalent band 7+ at 24%, Equivalent band 6 at 14%, Equivalent band 5 at 2%, and Equivalent band 4 or below at 0% (N=100). Each response is shown with a horizontal blue bar chart.

Exclude missing values

In the toolbar, press the [NA] button to toggle whether missing values are included in the basis or not.  This applies globally to all elements (except those with specific settings.)

Here, the basis is now N=40, reflecting answers only from nurses, and 60% of all nurses are Equivalent band 7+.

Blue rectangular button labeled "[NA]" from the Protobi toolbar. The button uses white text on a medium blue background and appears as a toggle control for showing or hiding missing values in the analysis. Protobi chart showing question S2 "What grade nurse are you?" with missing values excluded. Results show Equivalent band 7+ at 60%, Equivalent band 6 at 35%, Equivalent band 5 at 5%, and Equivalent band 4 or below at 0%. The sample size is now N=40, showing only those who answered the question. Each response displays as a horizontal blue bar chart.

Specific settings

You can override the global setting for one or more elements by pressing the edit icon, choosing "More properties..." from the context menu, and selecting a value for "showMissing."

  • Yes will always include missing values 

  • No will always exclude missing values 

  • Default will defer to the global setting per the toolbar

    Protobi Advanced properties dialog box showing configuration options for an element. The dialog displays multiple fields including key, displayKey, title, footnote, type, field, element, children, string, col, split, scale, denominator, weight, and maxValue. The showMissing property is highlighted with a red border, showing a dropdown menu set to "(default)" value.

Recode missing values

Alternatively, your analysis may require missing values to be recoded to another value, such as zero ("zero filling") or a known answer from another question.  See the tutorial on replacing missing values.  

Video Tutorial

Missing values