SPSS command: SPSS AIM
Attribute importance for clusters and groups (AIM)
Attribute importance profiles each group — typically each cluster from a segmentation — by testing which attributes make it different from the sample as a whole.
Free · No account · Files stay on your device · Updated 18 October 2026
What it tells you
Crosstabs runs SPSS AIM on your device: a chi-square test of each categorical attribute's category mix against the overall mix, a t test of each continuous attribute's group mean against the overall mean, and a −log10(Sig.) chart per group with an optional Bonferroni adjustment.
- Which attributes set each cluster apart, ranked by −log10(Sig.) against a critical line.
- In which direction: the group mean versus the overall mean for continuous attributes, and the group's most frequent category versus the overall one for categorical attributes.
- How strong the evidence is per attribute: chi-square or t, degrees of freedom and significance.
When to use it
Use it when
- Describing segments after a TwoStep, K-means or hierarchical cluster analysis, to write cluster profiles and names.
- Comparing any existing grouping (region, persona, plan) with the total sample on many attributes at once.
- Choosing which variables to feature in a segment summary.
Use something else when
- You want to compare groups with each other rather than with the whole sample: use one-way ANOVA or a crosstab chi-square.
- The clusters were formed from the same attributes: the tests will be significant by construction, so read them as descriptions, not as evidence.
- You want to predict group membership: use discriminant analysis or multinomial logistic regression.
- Case weights must be applied: the AIM algorithm defines none, so results are unweighted.
How to run it in Crosstabs
- 01
Open your data
Open the file with a cluster-membership or other grouping variable and switch to Statistics mode.
- 02
Choose Attribute importance
In the Classify group, pick Attribute importance.
- 03
Set the groups and attributes
Choose the Grouping variable (for example, cluster membership), then tick Categorical attributes and Continuous attributes.
- 04
Set the significance level and run
Keep Significance level at 0.05 or change it, optionally tick Bonferroni adjustment for the number of attributes, and click Run.
Worked example
Computed by the Crosstabs engine from the synthetic data below when this page was built — not typed in.
Data
180 synthetic customers already assigned to three clusters of 60 (for example by TwoStep or K-means). Plan tier (basic, plus, pro), age and monthly spend were generated to differ by cluster; region and number of store visits were generated identically in every cluster, so they should not distinguish any group. Significance level 0.05, no Bonferroni adjustment. Fixed seed.
| Attribute | Group value | Overall value | Statistic | df | Sig. | −log10(Sig.) |
|---|---|---|---|---|---|---|
| plan | basic | basic | χ² 41.932 | 2 | <.001 | 9.11 |
| region | South | South | χ² 0.813 | 3 | .846 | 0.07 |
| age | 27.07 | 41.13 | t −13.854 | 59 | <.001 | 19.48 |
| spend | 38.07 | 64.65 | t −13.314 | 59 | <.001 | 18.70 |
| visits | 6.22 | 5.80 | t 1.282 | 59 | .205 | 0.69 |
Critical value of −log10(Sig.): 1.30 (significance level 0.05).
| Attribute | Group value | Overall value | Statistic | df | Sig. | −log10(Sig.) |
|---|---|---|---|---|---|---|
| plan | plus | basic | χ² 5.142 | 2 | .076 | 1.12 |
| region | East | South | χ² 1.160 | 3 | .763 | 0.12 |
| age | 44.58 | 41.13 | t 3.432 | 59 | .001 | 2.96 |
| spend | 61.86 | 64.65 | t −0.969 | 59 | .337 | 0.47 |
| visits | 5.65 | 5.80 | t −0.446 | 59 | .657 | 0.18 |
Critical value of −log10(Sig.): 1.30 (significance level 0.05).
| Attribute | Group value | Overall value | Statistic | df | Sig. | −log10(Sig.) |
|---|---|---|---|---|---|---|
| plan | pro | basic | χ² 32.665 | 2 | <.001 | 7.09 |
| region | East | South | χ² 0.870 | 3 | .833 | 0.08 |
| age | 51.75 | 41.13 | t 11.503 | 59 | <.001 | 15.98 |
| spend | 94.03 | 64.65 | t 11.908 | 59 | <.001 | 16.60 |
| visits | 5.53 | 5.80 | t −0.877 | 59 | .384 | 0.42 |
Critical value of −log10(Sig.): 1.30 (significance level 0.05).
How to read it
In group 1 the most distinctive attribute is age (−log10(Sig.) 19.48), and 3 of the 5 attributes are above the critical value of 1.30. In group 2 the most distinctive attribute is age (−log10(Sig.) 2.96), and 1 of the 5 attributes are above the critical value of 1.30. In group 3 the most distinctive attribute is spend (−log10(Sig.) 16.60), and 3 of the 5 attributes are above the critical value of 1.30. Region and visits, which do not differ by design, stay below the critical value in every group. Read each group's row against the overall value to describe the cluster: for example, group mean spend versus the overall mean.
Output tables explained
- Attribute Importance — Group
- One table per group with its N: each attribute's group value and overall value, chi-square or t, df, significance and −log10(Sig.); the footnote gives the critical value of −log10(Sig.) for the chosen level.
- −log10(Sig.) — Group
- A bar per attribute for each group; bars beyond the critical value are significant at the chosen (optionally Bonferroni-adjusted) level.
Assumptions and notes
- Each group is compared with the whole sample, which includes the group itself; the tests describe distinctiveness rather than test a designed comparison.
- Categorical attributes: chi-square that the group's category proportions equal the overall proportions, with categories − 1 df. Continuous attributes: two-sided one-sample t of the group mean against the overall mean, with group N − 1 df.
- Missing values are deleted pairwise per attribute; cases with a missing group are left out.
- With many attributes and groups some tests will cross the line by chance; the Bonferroni option divides the level by the number of attributes.
How the results are validated
AIM is checked against SciPy on a 300-case file with three groups, two categorical and three continuous attributes and scattered missing values: the per-group chi-square statistic (with SciPy's chi-square survival function) and scipy.stats.ttest_1samp against the overall mean. Every statistic agrees to 10 decimal places, every df exactly and every p value to 1e-12. Validation evidence
Questions
How do I read −log10(Sig.)?
It turns the significance into a length: 1.30 corresponds to p = .05, 2 to p = .01 and 3 to p = .001. Longer bars mean the attribute is more distinctive for that group; bars past the critical value are significant.
Should I tick the Bonferroni adjustment?
Tick it when you screen many attributes and want to limit chance findings: the significance level is divided by the number of attributes, which raises the critical value. Leave it off for a handful of planned attributes.
Why is everything significant for my clusters?
If the clusters were built from these attributes, they differ on them by construction, and large samples make even small differences significant. Use the group-versus-overall values to judge whether a difference is big enough to matter.
What is the group value for a categorical attribute?
The group's most frequent category, shown next to the overall most frequent category. The chi-square test uses the whole category mix, not just the top category.
Is my data uploaded?
No. The tests are computed in your browser and the file stays on your device.
Related procedures
- TwoStep ClusterTWOSTEP CLUSTER
- K-Means ClusterQUICK CLUSTER
- Hierarchical ClusterCLUSTER
- One-Sample T TestT-TEST
- Crosstabs StatisticsCROSSTABS /STATISTICS
- Discriminant AnalysisDISCRIMINANT
- Free SPSS alternative: full coverage list
Run Attribute Importance (AIM) on your own data
Open a .sav, CSV or Excel file. Everything runs in your browser; the file never leaves your device.
SPSS is a trademark of IBM Corp. Crosstabs is not affiliated with or endorsed by IBM.