In social sciences, managing missing data is just as critical as collecting it. Missing data can disrupt the analysis process, reduce the reliability of results, and undermine the scientific value of your thesis. In this article, we’ll explain the types, causes, and most commonly used methods for handling missing data in a clear and accessible way.

1. What Is Missing Data?

Missing data refers to pieces of information that were intended to be collected but were not obtained for various reasons. This can occur due to unanswered survey questions, failed measurements, or data entry errors.

Examples:

• A participant leaves the age question blank
• Individuals who do not attend interviews
• Audio files not recorded due to technical issues

2. Types of Missing Data

There are three main types of missing data based on their cause:

MCAR (Missing Completely at Random)
The data is missing entirely at random. The missingness is unrelated to any variable.

Example: A section of a survey form is missing due to a printer error.

MAR (Missing at Random)
The missingness is related to another observed variable.

Example: Older participants skip questions related to technology.

MNAR (Missing Not at Random)
The missingness is related to an unobserved variable. This is the most complex type.

Example: Individuals with high income avoid answering income-related questions.

3. How to Handle Missing Data

Before analyzing, determine how much data is missing and what type it is. Then choose one of the following methods:

Listwise Deletion
Observations with missing data are excluded entirely from the analysis.

Advantage: Easy to apply
Disadvantage: Loss of data, reduced sample size

Pairwise Deletion
Only the missing data relevant to a specific analysis is excluded.

Advantage: More data is retained
Disadvantage: May lead to inconsistencies across analyses

Mean Substitution
Missing values are replaced with the mean of the variable.

Advantage: Simple and fast
Disadvantage: Reduces variance, may distort results

Regression Imputation
Missing data is predicted using other variables.

Advantage: More accurate estimates
Disadvantage: Requires complex models

Multiple Imputation
Missing data is estimated multiple times and results are combined.

Advantage: One of the most reliable methods
Disadvantage: Requires statistical knowledge and software

4. How to Present Missing Data Analysis in Your Thesis

• Report the rate of missing data. Clearly state which variables are affected and to what extent.
• Justify your chosen method. Explain why you selected it.
• Evaluate alternatives. Mention why other methods were not suitable.
• Compare results. Present analyses before and after handling missing data side by side.

5. Conclusion

Missing data is a natural part of research. What matters is managing it properly without compromising the analysis. Including missing data analysis in your thesis demonstrates academic rigor and enhances the credibility of your findings.