How to Delete Missing Data in SPSS: A Complete Step-by-Step Guide

Missing data is one of the most common problems in research datasets. If you ignore it, your results become unreliable. If you handle it wrong, your entire analysis can mislead you.

Knowing how to delete missing data in SPSS – or handle it properly – is a non-negotiable skill for every researcher. It affects everything from descriptive statistics to advanced regression models.

However, many students open SPSS and skip this step entirely. They run their tests and wonder why the outputs look inconsistent. Therefore, this guide walks you through every method to identify, manage, and delete missing data in SPSS, step by step.

Whether you are working on a dissertation, academic research, or a professional project, this guide gives you the clarity you need.

What Is Missing Data in SPSS?

Missing data refers to the absence of a value for one or more variables in your dataset. A respondent may have skipped a survey question. Equipment may have failed during data collection. A participant may have dropped out mid-study.

There are two types of missing values in SPSS: system-missing values and user-defined missing values.

Understanding the difference between them is critical before you decide how to delete missing data in SPSS.

• System-missing values – SPSS automatically recognises blank fields as missing and marks them with a dot (.) in the dataset.
• User-defined missing values – You assign a specific numeric code (e.g. -9 or 99) to represent missing responses. SPSS treats these as missing only after you define them in Variable View.

In addition, missing data patterns fall into three main categories:

• MCAR (Missing Completely at Random) – Missingness has no relationship to any variable in the dataset. This is the safest type for deletion.
• MAR (Missing at Random) – Missingness is related to other observed variables but not to the missing value itself.
• MNAR (Missing Not at Random) – Missingness is directly related to the value that is missing. This is the most problematic type.

Understanding the type of missingness is crucial for selecting an appropriate strategy.

Why Handling Missing Data Matters

Missing data affects every stage of your analysis. Ignoring it does not make it go away – it silently distorts your results.

Ignoring missing values can lead to biased estimates and incorrect conclusions, reduced statistical power, invalid standard errors and confidence intervals, and misleading visualisations and model diagnostics.

Moreover, different SPSS procedures handle missing data differently. Some tests automatically exclude cases with missing values. Others produce incorrect outputs if missing values are not defined properly.

Therefore, addressing missing data before you run any test is always the right approach. It protects the validity of your findings and strengthens the credibility of your research.

Before you begin, always ensure your data collection using SPSS is well-structured. Good data collection reduces missing values at the source, which is always better than fixing them later.

Step 1: Identify Missing Data in Your Dataset

Before you delete missing data in SPSS, you need to find it. SPSS provides several tools to detect missing values quickly.

Method 1 – Run Frequencies: Go to Analyze > Descriptive Statistics > Frequencies. Select all your variables and click OK. The output shows a “Missing” row for each variable. This tells you exactly how many cases have missing values for each item.

Method 2 – Run Descriptives: Go to Analyze > Descriptive Statistics > Descriptives. Check the N column in the output. If N is less than your total sample size, the difference represents missing cases.

Method 3 – Missing Value Analysis: Go to Analyze > Missing Value Analysis. This dedicated module gives you a full breakdown of missing patterns, including which variables have the most missing cases and whether the missingness is random or systematic.

For each variable, the Descriptives command uses the number of non-missing values. You can specify the missing=listwise subcommand to exclude data if there is a missing value on any variable in the list.

Running this initial check is essential. It tells you the scale of the problem before you decide which deletion or imputation method to use.

For more on structuring your research data before analysis, explore our guide on data analysis and interpretation in quantitative research. Proper data structure at the design stage minimises missing values significantly.

Step 2: Define User-Missing Values in Variable View

If your dataset uses a numeric code for missing responses (e.g. 0, -9, or 99), SPSS will not recognise these as missing unless you define them.

Here is how to define user-missing values:

1. Switch to Variable View in the SPSS Data Editor.
   
2. Find the column labelled Missing.
   
3. Click the cell for the variable you want to update.
   
4. A dialog box opens with three options:
   
   • No missing values – default setting
   • Discrete missing values – enter up to three specific codes (e.g. -9, 99)
   • Range plus one optional discrete missing value – for a range of codes
5. Enter your missing value code and click OK.
   

User-defined missing values are numeric values that need to be defined as missing for SPSS. You may use any value you choose to stand for a missing value, but be careful that you do not choose a value that already exists for the variable in the dataset.

Once defined, SPSS treats those codes as missing in all subsequent analyses. This is a foundational step before you attempt to delete missing data in SPSS.

Step 3: Choose Your Missing Data Strategy

Before deleting anything, decide on the right strategy. Your choice depends on how much data is missing and whether the missingness is random.

Here are the four main approaches:

1. Listwise Deletion (Complete Case Analysis): This removes every case that has a missing value on any variable in the analysis. It is the most straightforward method and works well when the missing data is minimal (under 5%).

2. Pairwise Deletion: This uses all available data for each calculation. Cases are only excluded for the specific pair of variables where data is missing.

By default, correlations are computed based on the number of pairs with non-missing data – this is often called pairwise deletion of missing data. You can also request listwise deletion within the CORR command using the missing=listwise subcommand, which runs the analysis only on observations with complete, valid data for all variables listed.

3. Mean Substitution: Replace each missing value with the mean of that variable. This preserves sample size but can artificially reduce variance.

4. Multiple Imputation: SPSS creates several complete datasets using statistical algorithms, runs analysis on each, and pools the results. This is the most robust approach for large amounts of missing data.

Multiple Imputation is a robust method that creates several complete datasets using algorithms and pools results to account for uncertainty.

For most student and academic research projects, listwise deletion is the most practical starting point – especially when missing data affects fewer than 10% of cases.

Understanding the full range of what data analysis tools are available helps you choose the right approach for your specific research context.

Step 4: Delete Missing Data in SPSS Using Select Cases

One of the most reliable ways to delete missing data in SPSS is using the Select Cases function. This filters out cases with missing values without permanently deleting them.

Follow these steps:

1. Go to Data > Select Cases.
   
2. Select If condition is satisfied.
   
3. Click the If button.
   
4. In the condition box, enter an expression that keeps only complete cases. For example:
   
   NOT MISSING(variable1) AND NOT MISSING(variable2)
   
5. Click Continue, then OK.
   

SPSS marks excluded cases with a diagonal line through the row number. These cases are not deleted – they are simply excluded from any analysis you run afterwards.

Moreover, if you want to permanently remove cases with missing values, follow these additional steps after selecting:

1. Go to Data > Select Cases > Delete unselected cases.
2. Click OK.
3. Save your file under a new name to preserve the original dataset.

Always save the modified file separately. You should never overwrite your original data when you delete missing data in SPSS. This ensures you can return to the full dataset if needed.

Step 5: Use the FILTER Command via Syntax

For researchers who prefer SPSS syntax, you can filter and delete missing data more precisely using the FILTER command.

Here is a basic example for removing cases with missing values on two variables:

SELECT IF (NOT MISSING(var1) AND NOT MISSING(var2)).

EXECUTE.

To permanently delete those filtered-out cases, follow with:

FILTER OFF.

SELECT IF (NOT MISSING(var1) AND NOT MISSING(var2)).

EXECUTE.

When creating or recoding variables, it is always good practice to test the resulting variables, especially for missing values. The type of missing value can make a difference in recode results, and in most cases, system-missing values are more likely to yield correct results.

Using syntax gives you a permanent record of exactly what you did and why. This is valuable when writing your methodology chapter or replicating your analysis later.

Step 6: Replace Missing Values Using the Transform Menu

Sometimes deleting missing data in SPSS is not the best option – especially when you have a large percentage of missing cases. In those situations, replacing missing values is more appropriate.

Go to Transform > Replace Missing Values. SPSS offers the following replacement methods:

• Series Mean – replaces missing values with the mean of the entire variable
• Mean of Nearby Points – replaces using the average of surrounding values
• Median of Nearby Points – replaces using the median of surrounding values
• Linear Interpolation – estimates missing values using a linear trend between neighbouring data points
• Linear Trend at Point – fits a linear trend model and predicts the missing value

Each method creates a new variable with the imputed values. Your original variable remains unchanged.

This approach is particularly useful for time-series data or repeated measures datasets where values follow a predictable trend.

For a complete walkthrough of data preparation techniques, including transformation and recoding, see our guide on how to transform data in SPSS. These skills work hand in hand with missing data management.

Step 7: Run Multiple Imputation for Large Missing Data

When more than 10–15% of your data is missing, listwise deletion is no longer safe. It removes too many cases and biases your results. Multiple Imputation is the recommended solution in this scenario.

Here is how to run it in SPSS:

1. Go to Analyze > Multiple Imputation > Impute Missing Data Values.
2. In the Variables tab, move variables with missing values into the Impute box.
3. Add predictor variables to the Predictors box.
4. In the Method tab, select Fully Conditional Specification (FCS).
5. Set the number of imputed datasets (5 is standard).
6. Click OK.

SPSS creates five complete datasets and stores them in a stacked format in your data file. An Imputation_ variable identifies which dataset each row belongs to.

After imputation, run your analysis as normal. SPSS automatically pools the results across all five datasets using Rubin’s Rules, giving you a single set of valid, robust findings.

Once the imputation is complete, SPSS provides an Imputation Summary Table showing which variables were imputed, how many values were missing, and the method used.

This technique is widely accepted in published research and is especially common in psychology, healthcare, and social science studies.

Step 8: Verify Your Missing Data Decisions

After you delete missing data in SPSS or apply imputation, always verify your results before proceeding to analysis.

Run Frequencies again on all your key variables. Check:

• N values match your expected sample size after deletion
• Missing count is now zero (or as expected)
• Distributions look reasonable and have not shifted dramatically
• New imputed variables fall within the plausible range of the original variable

In addition, document every decision you made about missing data. Your methodology section must explain:

• How much data was missing (percentage per variable)
• What type of missingness did you identify (MCAR, MAR, MNAR)
• Which strategy do you use and why
• How many cases were removed or imputed

This transparency is essential for peer review and academic submission. Reviewers will always ask about your missing data handling.

To learn how to present your analysis findings in a professional format, read our guide on how to make a data analysis report. Reporting missing data decisions clearly is a key part of any credible research report.

How Different SPSS Tests Handle Missing Data

It is important to know that SPSS procedures handle missing data differently by default. Here is a quick reference:

SPSS Procedure	Default Missing Data Handling
Descriptives	Excludes missing per variable (pairwise)
Frequencies	Excludes missing; shows separate missing count
Correlations	Pairwise deletion by default
Regression	Listwise deletion automatically
Factor Analysis	Listwise deletion automatically
ANOVA	Listwise deletion for each complete design

For regression, if values of any of the variables are missing, the entire case is excluded from the analysis – this is listwise deletion of missing data. For factor analysis, cases with missing values are deleted listwise, meaning observations with missing values on any variable in the analysis are omitted entirely.

Understanding these defaults helps you interpret your output correctly. If your N drops unexpectedly between tests, missing data handling is usually the reason.

For deeper statistical analysis after cleaning your data, explore our guide on correlation vs regression analysis to understand how missing data affects each method differently.

Common Mistakes When Deleting Missing Data in SPSS

Avoid these frequent errors that researchers make when working with missing values:

• Not defining user-missing values – SPSS treats -9 or 99 as valid data if you do not define them as missing in Variable View first
• Overusing listwise deletion – Removing too many cases when data is not MCAR, introduces bias into your results
• Forgetting to save a backup – Always keep your original file before applying any deletions or transformations
• Misidentifying missing patterns – Treating MNAR data as MCAR leads to fundamentally flawed conclusions
• Skipping verification – Never assume your deletion or imputation worked correctly; always recheck with Frequencies

Furthermore, avoid deleting variables with high missing rates without first investigating why the data is missing. The missingness itself may carry important information about your sample.

To see how missing data decisions affect downstream statistical tests, read our complete SPSS tutorial for data analysis. It covers the full workflow from data preparation to result interpretation.

Quick Checklist: How to Delete Missing Data in SPSS

Before running any analysis, work through this checklist:

•  Run Frequencies or Missing Value Analysis to identify missing cases
•  Define user-missing values in Variable View for all coded variables
•  Check the percentage of missing data per variable
•  Determine the missing data pattern (MCAR, MAR, MNAR)
•  Choose your strategy (deletion, replacement, or imputation)
•  Apply the strategy using Select Cases, Transform, or Multiple Imputation
•  Verify results using Frequencies after deletion or imputation
•  Document all decisions in your methodology section
•  Save the cleaned file under a new name to preserve the original

For hands-on practice, use a practice dataset for SPSS to apply each missing data technique before working on your actual research file.

Conclusion

Learning how to delete missing data in SPSS is not just a technical skill – it is a research integrity issue. Every decision you make about missing values shapes the validity of your findings.

Start by identifying the scale and pattern of missingness. Then choose the appropriate strategy – whether that is listwise deletion, pairwise deletion, mean substitution, or multiple imputation. Verify your results afterwards and document every step clearly.

Moreover, remember that deleting missing data is not always the right answer. Sometimes imputation protects your analysis better than removal. Therefore, understand your data before you make any decisions.

Ultimately, clean data leads to credible analysis. Take the time to handle missing values properly – your research quality depends on it.

Frequently Asked Questions