Missing values can break your statistical models before you even run them. When you import a spreadsheet into Jamovi, empty cells or blank text strings often turn into NaN (Not a Number) values or system-missing designations. If you don't handle these gaps correctly, your sample sizes shrink, your standard errors skew, and your conclusions lose validity.
Jamovi handles missing data automatically by excluding cases listwise in most analyses, but that default behavior isn't always what you want. Understanding how Jamovi treats missing values, how to specify missing data codes, and when to filter or transform variables gives you control over your output. Let's look at how to manage missing data properly without distorting your findings.
How Jamovi Detects and Treats Missing Values
Jamovi reads missing data in two main ways: true blank cells in your spreadsheet and user-defined missing values. When you import a CSV or Excel file, blank cells are instantly treated as missing by the software.
By default, Jamovi uses listwise deletion (sometimes called complete-case analysis) for most statistical procedures. If a participant has a missing value on any single variable included in your model, Jamovi drops that entire row from the analysis.
- Listwise deletion benefits: Keeps calculations simple and ensures every test uses the exact same sample size.
- Listwise deletion drawbacks: Throws away usable data from other variables and reduces statistical power, especially if you have multiple variables with scattered gaps.
You can inspect your dataset's missing patterns by looking at the summary statistics pane or by setting up filters to isolate incomplete rows.
Setting Up Missing Values in the Jamovi Variable Editor
Often, datasets contain numeric codes representing missing answers—such as -99, 999, or 8 for "Prefer not to say". If you leave these numbers as active data, Jamovi will calculate means and standard deviations that include those code numbers, which ruins your descriptive statistics.
To fix this:
- Double-click the column header in the Jamovi data spreadsheet to open the Variable Editor.
- Locate the Missing field for that specific variable.
- Enter the specific numeric code (like
-99) that represents your missing data.
Once defined, Jamovi treats those coded numbers as true missing values. They won't factor into your math, and Jamovi will exclude them automatically during your analyses.
Filtering and Subsetting Missing Data
Before running imputations or advanced models, you often need to isolate rows with missing values to inspect why they are missing. You can use Jamovi's computed variables and filters to manage this.
To create a filter that isolates complete cases:
- Navigate to the Data tab and click Filter.
- Write a logical expression using Jamovi functions to check for missing values. For instance, you can filter rows where key variables are not missing.
- Jamovi will temporarily hide rows that contain missing values in your specified columns, letting you run comparisons between complete and incomplete cases.
Checking whether your data is Missing Completely at Random (MCAR) helps you decide whether you can safely drop missing rows or if you need to take alternative steps. If data is missing systematically—such as high-income earners refusing to report their salary—listwise deletion will introduce severe bias into your estimates.
Common Mistakes When Dealing with Missing Data
When analysts encounter NaN values in Jamovi, they often make quick mistakes that distort their results.
- Replacing blanks with zero: Substituting zero for a missing test score or income value drags down your mean and flattens your variance. Zero is a real number, not a missing value.
- Ignoring pattern types: Deleting rows without checking if the missingness is random leads to biased samples.
- Over-imputing small samples: Filling in half of a tiny dataset with estimated values creates artificial consistency that doesn't reflect reality.
If you want to read more about the underlying design of this open-source software, check out the Jamovi Project website. For a broader academic perspective on how missing data impacts research integrity, read guidelines from the American Psychological Association.
Frequently Asked Questions
Does Jamovi support multiple imputation out of the box?
Base Jamovi focuses primarily on listwise deletion and standard descriptive filtering. For advanced multiple imputation techniques, users often install specialized Jamovi modules (such as snowRMM or similar packages via the Jamovi library) or export the dataset to R to run packages like mice.
Why is Jamovi showing NaN in my calculated variables?
If you create a computed variable using a formula and one of the input variables contains a missing value, Jamovi will return NaN for that row. You can manage this by using conditional logic within your computed variable formulas to check for missing inputs first.
How do I know if my missing data is random?
You can run chi-square tests of independence or t-tests comparing participants with complete data against those with missing data on other demographic variables. If significant differences appear, your data is likely not missing at random.
Can I export datasets with NaN values back to Excel?
Yes. When you export your Jamovi dataset back to CSV or Excel format, Jamovi exports missing values as blank cells, which keeps your exported files clean for other spreadsheet software.
