A comical guide to Missing Not At Random (MNAR)

Recall that Missing Not At Random (MNAR) is a type of missing data mechanism where the probability of missingness is related to the unobserved data itself. Here are some more examples of MNAR:

1. Medical Studies: Patients with more severe symptoms are less likely to attend follow-up appointments because their condition makes it difficult for them to do so. Therefore, missing follow-up data is related to the severity of the condition, which is unobserved.
2. Survey Responses: In a survey about income, individuals with higher incomes might be less likely to report their income due to privacy concerns. Consequently, the missing income data is related to the actual income levels, which are not observed.
3. Educational Testing: Students with poor academic performance might be more likely to skip certain sections of a standardized test. Thus, the missing test scores are related to the students’ actual performance, which is unobserved.
4. Customer Feedback: Customers who are extremely dissatisfied with a product may choose not to provide feedback because they don’t believe it will make a difference. Therefore, the missing feedback data is related to the level of dissatisfaction, which is unobserved.
5. Employee Surveys: Employees who are highly disengaged or dissatisfied with their jobs might be less likely to complete an engagement survey. Hence, the missing survey data is related to the level of disengagement or dissatisfaction, which is unobserved.
6. Clinical Trials: Participants experiencing severe side effects from a medication may drop out of a clinical trial. Thus, the missing data on side effects is related to the severity of the side effects, which are unobserved.

In each of these examples, the missingness of the data is directly related to the values of the data that are missing, making it MNAR.