Unofficial Guide to Creating Collective Mental Models with 3CM and Anthropac

Step 6: Name the Clusters

Now that you have at least a preliminary partition selected, you want to determine what terms are contained in each cluster so that you can assign names to the clusters. What seems like an easy step can be difficult because the MDS maps often contain overlapping clusters that make it hard to tell which cluster a term belongs to. This step provides a chart indicating the terms contained within each cluster at each partition level.

The majority of tools you need for 3CM analysis are in the Windows version of Anthropac, but the next feature requires the DOS version which cannot run on Windows 7 or above without a DOS emulator such as DosBox.

To begin, open DosBox.
For this next step to work, you must have installed Anthropac to a folder directly under C:. I installed Anthropac to a folder called “ap” under C:.
In the DosBox window, after the Z:\> prompt, type mount s c:\ap, where “s” is an letter that is not used as a drive on your computer and “ap” is the folder directly in your C: drive that contains Anthropac. Press Enter and the following message will appear: Drive S is mounted as a local directory c:\ap\
At the Z:\> prompt type s:, where “s” is the drive letter selected in the previous step. Press Enter. Now an S:\> prompt appears.
At the S:\> prompt, type apx, where “apx” is the program file name for Anthropac. Press Enter and the Anthropac DOS program should launch as shown below.
Press any key to move past the intro screen and into the program.
Go to Data > Import > DL. Press Enter.
Make sure your .txt data file is saved within your Anthropac folder on the C: drive with a short filename. Then, for Input ascii data file, enter the file name including the .txt extension. Press F10.

You will see the data file load as shown below. In addition, two files will be added to your Anthropac program folder. They will have the same file name as your data file, but with the extensions .##D and .##H.

Press ESC to return to the main page.
Back on the main page, navigate to Tools > Clustering > Hierarchical. Press Enter.
For Input dataset:, enter S:\DATA, where “S” is the drive letter selected in the step 9 and “DATA” is your data file name without extension. Press F10.
The resultant icicle plot appears on the screen.
The icicle plot is also saved as OUTPUT.LOG in the Anthropac program file. Each time you run a new icicle plot, OUTPUT.LOG will be overwritten, so I like to make a copy of the file and rename it.

Below is the output file for the sample data set. The results show 43 rows, which corresponds with the 43 partitions. At the top are the numbers for each of the 57 terms used in the 3CM activity. On the bottommost row, you see Xs all the way across – this row is the equivalent of _partition1 where all of the terms were contained in a single cluster. The next row up is the equivalent of _partition2 where there was one cluster of 56 terms (represented by the Xs in the icicle) and one cluster of 1 term (represented by the dot for term 11 in the icicle plot, which is “Community Zoning). One more row up is the equivalent of _partition 3. Here we see a break in the Xs between terms 55 and 9 – this division and the dot for term 11 indicate three clusters. As we continue up row by row, we can see where clusters divide into smaller clusters.
Counting up from the bottom, identify the partition you selected in the last step of the previous section, then use the term numbers at the top to see what terms are in each cluster. Based on this information, you can assign names to each cluster. I suggest having several people perform the naming independently and then compare the results to help minimize the intrusion of your personal biases.

Once you have performed this analysis on the partition you originally selected, you may want to look at some of the nearby partitions to see if they may be better suited for your analysis.

Step 7: Analyze the Results