Projective Mapping Workbook

Projective Mapping Workbook

Overview

Projective Mapping data can be analyzed in Compusense by using the Projective mapping workbook. This report, centered around MFA (Multiple Factor Analysis), includes multiple sheets to allow you to explore your results in depth.

The data can also be exported for analysis in XLSTAT or in SensoMineR.

In this workflow we will focus on the analysis within Compusense.


Generate The Report

To generate the Projective mapping workbook:
  1. In the  Results  area of your test, filter the results if necessary by clicking Filters.

  2. Select Reports > Create report.

  3. In 1. Select report type , select Projective mapping workbook.  

  4. In 2 Select options , select the sheets you wish to include in the report. All sheets are included by default.

    Select from the following options (details about individual sheets are provided in the Report Details section):

    1. Decimal places. Default selection comes from the selection previously made when generating this report or when generating the Standard report. Choose from zero (0) to four (4) to meet your analysis needs.

    2. A separate sheet will be generated when any of the options below are selected:
      1. Summary statistics. 
      2. MFA summary. 
      3. RV coefficient. 
      4. Panelist coordinates. 
      5. Correlation circle. 
      6. Consensus configuration. 
      7. Partial inertia. 
      8. Confidence ellipses. 

    3. Minimum citations. The default value is 3 while the maximum value is 10. This threshold excludes from the analysis any tags with fewer occurrences than the selected value. To include all tags, set the value to 0 (zero).
      Notes
      If a tag is used an equal number of times in every sample, it will be automatically excluded from the analysis due to the lack of variability between samples.

    4. Alpha. The default Alpha value is 0.05. Set the value that matches your company standards.

    5. Panelist identifier. Default selection is Sample set number, which masks panelist information. Select Panelist code if that better suits your reporting needs.


  5. In 3 Select questions , select the Projective mapping questions that you wish to include in the report. Only Projective mapping questions are compatible with this report.

  6. The .xlsx is the only export type. Click Create my report. 

  7. Click the down arrow, save the report to a location on your computer or network drive and open it.


Projective Mapping Report Details    

Review below the information found on each sheet in the report.  


Summary Statistics 


The Summary statistics sheet displays the Sample set number or Panelist code (depending on what you selected before generating the report), Minimum, Maximum, Range, Mean, and Standard deviation for each variable.

This sheet also displays the complete list of tags, their associated tag numbers, and the the number of times each tag was used in the test. Tag numbers are assigned sequentially to each unique tag during data collection. Tag numbers are useful in plots for reducing visual clutter when tags are lengthy.

As shown in the example image above, both spelling and letter case impact the tag counts. For example, we can see "crisp", "Crispy", and "crispy" as unique tags. Although we can assume that different people were trying to use the same term, they did not create the tags the same way, and so each of these terms is treated as a unique tag. The tags list can help you decide how to handle these instances. You might want to increase your citations threshold and thus exclude such tags. Alternatively, you may clone your test by keeping results in the clone. You can then edit raw results to update spelling and/or letter case before regenerating the report.

The citations threshold set prior to generating the report is displayed below the tags list. Tags that were cited less number of times than the threshold value will not be included in the analysis.

Notes
If a tag is used an equal number of times in every sample even if it meets the Minimum citations threshold, it will be automatically excluded from the analysis due to the lack of variability between samples.

MFA Summary 



The MFA (Multiple Factor Analysis) summary allows you to review inertia extracted within the MFA. Inertia extracted represents the amount of variance explained by the dimensions.

RV Coefficients 



The RV coefficients sheet represents the panel consensus configuration vs. each assessor configuration (after optimal scaling). The coefficient ranges from 0-1 with 0 meaning no correlation and 1 meaning strong correlation (i.e. what is the individual’s correlation to the panel).


Panelist coordinates


The Panelist coordinates sheet provides one graph per sample set. These graphs use panelists' raw data to plot the placement of products on the mapping area. The panelists are displayed with green markers while panel is displayed with black markers.


Correlation Circle Graphs

The Correlation circle sheet provides 3 consensus graphs: 
  • Correlation circle. Plots panelist and tag correlation.



  • Assessors. Plots panelist correlation.  Scroll down the sheet to find this plot.



  • Tags. Plots tag correlation. Only tags that were used more times than the citation threshold value set before generating the report will be included in the plot. Scroll down the sheet to find this plot.


Consensus Configuration

The consensus configuration sheet provides 2 graphs:

  1. Projection of panelists on consensus configuration. Plots the consensus sample configuration and individual panelist configuration.


  2. Consensus sample configuration. Plots the consensus sample configuration. Scroll down the sheet to find this plot.



Partial Inertia


The Partial inertia sheet provides the plots of inertia extracted for each sample set and the tags, where tags are grouped together. The partial inertia measures how much each specific variable (i.e. sample sets and tags) contributes to the structure of the data. From this graph, you can determine which variables drive the dimension. 

Only tags that met the citation threshold value set before generating the report will be included in the plot.


Confidence Ellipses



The Confidence ellipses plot illustrates the stability of the configuration. The confidence intervals are derived from data generated using virtual panels.

The product configuration is based on true panel results. Virtual panel means are Procrustes‑transformed onto the real configuration. The coordinates from all virtual panels are then used to generate confidence ellipses, which assess the uncertainty of the real product configuration.

This results in 95% confidence ellipses for the product configuration. Large ellipses with substantial overlap indicate low configuration stability, whereas smaller ellipses with minimal overlap indicate high stability.

    • Related Articles

    • Projective Mapping .r Export

      Overview Projective Mapping data can be analyzed directly in Compusense and it can be exported for analysis in XLSTAT or in SensoMineR. In this workflow we will focus on exporting data and analyzing it in SensoMineR. Select the Right Export Options ...

    • Standard Test: Projective Mapping

      Overview For Projective Mapping, panelists are asked to organize multiple samples in a two-dimensional space according to perceived similarities and differences. Panelists may be asked to tag the samples, for example, to associate a series of samples ...

    • PCA Workbook

      Overview PCA (Principal Component Analysis) allows for a better understanding of the multi-dimensional relationships among samples and attributes. The visual representation allows for an easier interpretation of these relationships. The PCA is ...

    • TI Workbook

      Overview The Time Intensity (TI) test data can be analyzed directly in Compusense by generating the TI Workbook report. On this page you will learn how to generate the report and what various options exist to help you meet your analysis needs. ...

    • Analysis Across Tests: Pivot Table

      Overview The Pivot table feature enables you to review descriptive statistics of data sets from multiple compatible tests in a live and interactive way. The data transformations are immediate and the table can be exported as an Excel workbook or a ...