Missing value analysis
Describes how much data are missing and in which patterns, estimates means and covariances by EM, and tests whether data are missing completely at random (Little's MCAR test).
Menu: Analyze → Missing Value Analysis…
When should I use it?
- Before deciding between listwise deletion and multiple imputation.
What are the assumptions?
- EM assumes multivariate normal data, missing at random.
How do I run it in StatLab?
- Choose the quantitative variables; keep patterns and EM ticked.
How do I read the output?
| Output | Meaning |
|---|---|
| Missing Value Patterns | Each distinct pattern of missing (X) and its frequency. |
| EM Means / Covariances | Maximum-likelihood estimates using all available data. |
| Little's MCAR test | Significant = data are probably not MCAR. |
Worked example
On the practice survey dataset (File → Sample dataset), this command — which the dialog's Paste button also produces — runs the procedure:
PROC mvAnalysis /vars=(score age) /patterns=YES /em=YES.
Output tables: Univariate Statistics · Overall Summary of Missing Values · Missing Value Patterns · EM Means · EM Covariances · EM Correlations.
How would I report it in APA style?
Little's MCAR test was not significant, χ²(0) = 0.00, p = 1.000, consistent with data missing completely at random.
Computed by StatLab's validated procedure code when this page was built. The example only illustrates the steps; choose variables that fit your own research question.
What are common mistakes?
- Reading a non-significant MCAR test as proof that data are MCAR.