--------------------------------------
Toy Example: Step-by-Step Instructions
--------------------------------------
1. Upload the original data CSV file ("example-colors.csv") to the MF
Calculator website.
2. Choose the parameters for the analysis (N Factor and the weights
for Metric and Frequency Similarity); in the published example, we
chose N = 5 and gave equal weights to Metric and Frequency
similarity (50).
3. Generate the MF similarity matrix, by clicking on the name of the
database.
4. Click in "MF similarity" to visualize the Overall Metric Frequency
Similarity on the screen.
5. Click in "Display Graph" to obtain the additive similarity tree
that represents the MF matrix.
-----------------------------------------------
Illustrative Example: Step-by-Step Instructions
-----------------------------------------------
1. Upload the original data CSV file ("nle-inventory-data.csv") in MF
Calculator website.
2. Choose the parameter for the analysis (N Factor and the weights
for Metric and Frequency Similarity); in the published example, we
chose N = 25 and gave equal weight to Metric and Frequency
similarity (5).
3. Generate the MF similarity matrix, by clicking on the name of the
database.
4. Click on "MF similarity" to visualize the Overall Metric Frequency
Similarity on the screen.
5. Download the MF similarity matrix by clicking on one of the icons
under "Export to" (there are three formats available for the
exportation file; in the published example we used the XLS format).
6. Convert the exported MF similarity matrix into a format compatible
with the statistical software that you will use for the analysis;
to analyze the illustrative example, we used SPSS software (please,
check the FAQ page on the MF calculator site, for potential
difficulties in converting the MF matrix file to a SPSS similarity
matrix file).
7. Proceed with the cluster analysis; in the illustrative example, we
chose the average linkage within groups agglomerative method for
building the dendrogram (please, check the SPSS commands below for
the exact procedures used in the illustrative example).
SPSS commands used for analyzing the published example
* Hierarchical cluster analysis using the average linkage within
groups agglomerative method (WAVERAGE algorithm).
* Cluster membership for k = 2 clusters.
* SPSS similarity input matrix located in "c:\MFanalysis" folder.
CLUSTER C001 to C100
/METHOD=WAVERAGE
/PLOT=DENDROGRAM
/PRINT SCHEDULE CLUSTER(2)
/MATRIX=IN('c:\MFanalysis\nle-mf-mat-n25.sav').
* End of SPSS commands
Note: To interpret the clusters found in the illustrative example, we
further analyzed data using the following three steps:
a) We inspected visually the dendrogram and make a subjective decision
about the number of relevant clusters to be considered (we chose k
= 2 clusters).
b) From the output provided by SPSS (cluster membership table) we
extracted the cluster membership for each participant and input it
into the original data base.
c) In order to interpret the meaning of the clusters found, we
compared clusters using the information available (we used the mean
scores obtained in each dimension of the NLE Inventory and the
number of Negative Live Events reported by each participant;
comparisons involved Cohen's d effect size and t-tests for
independent samples; the file
"nle-inventory-data+clustermembership.csv" contains all the
information necessary to reproduce the illustrative example
analysis).