Diagnose the pattern and mechanism of missing values — and identify the best strategy for handling them.
Missing Data Analysis diagnoses the extent, pattern, and mechanism of missing values in a dataset — a critical step before any analysis, as the wrong handling of missing data can seriously bias results. The three missing data mechanisms are MCAR (Missing Completely At Random — missingness is unrelated to any variable), MAR (Missing At Random — missingness depends on observed variables), and MNAR (Missing Not At Random — missingness depends on the missing value itself).
StatRyx computes the percentage of missing values per variable and overall, creates a missingness heatmap and pattern matrix, performs Little's MCAR test, and assesses whether missingness is related to observed variables. Based on this diagnosis, StatRyx recommends the appropriate handling strategy: listwise deletion (MCAR, small %), mean/median imputation, or multiple imputation by chained equations (MICE) for MAR data.
Start by describing how much is missing per variable and per participant, then ask why. Data are MCAR (missing completely at random) when missingness is unrelated to anything; MAR when it depends on other observed variables; MNAR when it depends on the missing value itself. Little's MCAR test checks the first; a non-significant result is consistent with MCAR. Listwise deletion is only unbiased under MCAR, while multiple imputation and full-information maximum likelihood remain valid under MAR.
| Variable | n Missing | % Missing | MCAR p-value |
|---|---|---|---|
| Anxiety | 8 | 3.5% | .421 |
| Depression | 15 | 6.6% | .187 |
| Social Support | 4 | 1.8% | .834 |
Missing data analysis revealed that 4.2% of values were missing across all variables (range: 0–12% per variable). Little's MCAR test was non-significant, χ²(48) = 51.3, p = .342, supporting the MCAR assumption. Missing values were handled via listwise deletion, retaining 94.1% of the original sample (N = 213).
Yes. You can run it without an account (3 free analyses), and a free account adds 10 more — a one-time allowance, no payment needed. Every result includes the APA 7 write-up. Paid plans (Student from $9/month) add more analyses and Word, PDF and Excel export.
MAR: missingness is related to other observed variables (e.g., older participants are more likely to skip the income question, but income itself doesn't cause skipping). MNAR: missingness is related to the missing value itself (e.g., people with very high income skip the income question). MNAR is the most problematic and hardest to detect.
Multiple imputation creates several complete datasets by filling in missing values with plausible values drawn from a predictive distribution. Analyses are run on each imputed dataset and results are pooled using Rubin's rules. It is the gold standard for MAR data.
Report the amount (overall and the range per variable), the result of Little's MCAR test, and the method used: "4.2% of values were missing (0–12% per variable). Little's MCAR test was not significant, χ²(48) = 52.7, p = .30, so the data were treated as missing completely at random."
There is no universal cutoff. Under 5% rarely matters for most analyses; above 10% the handling method starts to change results; above 40% on a key variable, any method relies heavily on its assumptions. The pattern (why data are missing) matters more than the percentage.
Run a Missing Data Analysis on your data →