Definition
Missing data are gaps in the dataset — an unmeasured sample, an unanswered question, a sensor outage. It matters why they are missing: at random (MCAR), depending on other variables (MAR), or because of the value itself (MNAR — high incomes are not reported). That determines the remedy: dropping cases, replacement (imputation) or methods that handle missing values directly.
Data & visualisation
Dropping whole rows (listwise) is simple, but with many variables you lose a large part of the data and under MNAR you bias the result. Mean replacement underestimates variability; regression imputation, k-nearest neighbours or multiple imputation, which accounts for replacement uncertainty, are better.
Always describe the missing data first: how many, where and whether they relate to other variables. Decision trees and some ensemble methods handle missing values without imputation via surrogate splits.
In Statistica
Statistica reports the extent of missing data in Descriptive statistics, lets you choose casewise or pairwise deletion, and in the Data menu replaces missing values by mean, median, interpolation or regression estimate; tree methods in Data Mining work with missing values directly. Missing-data codes are configurable per variable.
Related terms
- OutliersAn outlier is an observation that differs conspicuously from the rest — a measurement or entry error, a different population, or a genuine rare event.
- Linear regressionLinear regression describes how a dependent variable (yield, price, consumption) depends on one or more explanatory…
- Decision treesA decision tree splits data by a sequence of simple rules ("age > 45 and income < 30 thousand") into ever more…
- Population and sampleA population is the entire group of units you want to make a claim about — all patients, every product from a line.
Knowledgebase guides
FAQ
- How much missing data is still acceptable?
- Up to 5 % is usually harmless; at 10–20 % consider imputation, and above 30 % in one variable ask whether to use it at all.
- Is mean replacement all right?
- Only for a small share of data missing at random. It reduces variance and distorts correlations; regression or multiple imputation are better.
Try it on your own data
Statistica free for 30 days
Full version, no credit card. Or get a pricing estimate in a minute.