Dendrogram (hierarchical clustering)
Shows how cases join into clusters step by step. The default is SPSS’s between-groups (average) linkage on squared Euclidean distance with the distance axis rescaled 0–25. Large data sets (up to 4,000 cases) are clustered quickly; static charts show the last 250 clusters, and the Chart Builder can scroll through every case.
Menu: Graphs → More charts → Dendrogram (hierarchical clusters)…
When should I use it?
- Exploring natural groupings and choosing the number of clusters.
What are the assumptions?
- Scale clustering variables; standardize them if they use different units.
How do I run it in StatLab?
- Choose the clustering variables (and a label variable).
- Choose the method and whether to standardize.
- Choose how the distance axis is scaled.
How do I read the output?
| Output | Meaning |
|---|---|
| Branches | Cases or clusters joining at the distance on the axis; long vertical gaps suggest a good number of clusters. |
Worked example
On the practice survey dataset (File → Sample dataset), this command — which the dialog's Paste button also produces — runs the procedure:
PROC vizDendrogram /vars=(score age) /label='' /linkage=average /standardize=NO /rescale=spss.
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?
- Clustering variables on very different scales without standardizing.
- Treating the tree as proof that clusters exist.