Hello Jason,
The method labeling of the dendrogram, where ‘Average
Linkage (Between Groups)’ is printed in the title regardless of the
actual linking method used, has been filed as a bug for Statistics 17. It is
expected to be fixed in version 18. I’ve pasted a related resolution
from the Knowledgebase at http://support.spss.com
below my signature.
David Matheson
SPSS Statistical Support
*********************
Resolution number: 83549 Created on: Apr 30 2009 Last Reviewed on: May 5 2009
Problem Subject:
Cluster labels results as between groups averaging although within groups
averaging requested
Problem Description:
I have been using the CLUSTER procedure for hierarchical cluster analyses
in PASW Statistics 17. I clustered the same set of variables with several
linkage methods to see whether the cluster memberships would change. Regardless
of the linkage method chosen, the Dendrogram was titled "Dendrogram Using
Average Linkage (Between Groups)". The dendrograms actually differ across
runs with different linkage methods and reflect the Agglomeration schedule for
each run. The linkage method that is reported in the "Case Processing
Summary" footnote reflects the linkage method chosen for each run.
Resolution Subject: Mislabeled
linkage method in dendrogram title filed as defect for Statistics 17
Resolution Description:
The method labeling of the dendrogram, where 'Average Linkage (Between Groups)'
is printed in the title regardless of the actual linking method used, has been
filed as a bug for Statistics 17. It is expected to be fixed in version 18.0.
The dendrogram does reflect the linkage method that was chosen by the user,
despite the mislabeled title.
From: SPSSX(r)
Discussion [mailto:[hidden email]] On
Behalf Of Beckstead, Jason
Sent: Thursday, April 30, 2009
9:06 AM
To: [hidden email]
Subject: v 17.01 mislabels
dendgrograms created in CLUSTER procedure
I have been using SPSS CLUSTER procedure for hierarchical cluster
analyses for many years but only recently had the need to use the procedure in
v 17. I am concerned because if the user requests within-groups averaging
method, the appropriate algorithm seems to be applied but resulting figure is
mislabeled "Between-groups" . I don't know what other
"improvements" have been made to the procedure in v 17, but I am now
hesitant to trust the results.
_____________________________________________________________
Jason W. Beckstead, Ph.D.
Associate Professor/Quantitative Methodologist
University of South Florida College
of Nursing
12901 Bruce B.
Downs Blvd., MDC22, Tampa, FL 33612, USA
Statistical Editor, International Journal of Nursing Studies
phone: +1.813.974.7667 fax:
+1.813.974.5418
International Journal of Nursing Studies
Locked
