Methods for the Analysis of Choices and Evaluations

Step-by-step creation of data for Options Analysis of an individual

Using data such as that described in the webpage on collecting individual uncertainty data , you can use the steps below to create a series of option appeal uncertainty distributions for an individual. This data can then be used for a Options Analysis of an individual.

  1. Startup Excel, click on the Tools menu and then on the Data Analysis item. (If Data Analysis is not present, click on the Add-Ins item and in the dialog box check Analysis Toolpak to install it.)

  2. Within the Data Analysis list scroll down to Random Number Generation and select it.

  3. In the Random Number Generator dialog: Insert the number of options in the Number of Variables box. Insert the number of scenario cases you wish to analyse in the Number of Random Numbers box. Choose, say, 500 for relatively stable results that will run reasonably quickly. From the Distribution drop-down box select Normal. Then select your output range on your Excel sheet and click OK. (Just leave the mean and standard deviation specifications as they are. These will be OK.)

  4. You can insert any names you want at the top of each option distribution. (Later, Options Analysis will insert default names if they are unnamed.)

  5. Start up Options Analysis from the Tools Menu and from the Data Specification and Analysis Options Dialog: Select the appeals input range. Check or uncheck the Labels box, as appropriate. For Data Adjustment be sure to check the No Adjustment option. Amongst the analysis options, only Functional Analysis should be checked. Click the Next Step.

  6. In Functional Analysis - Basic Dialog: Check Change Analysis and the Global Impact analysis option. For Statistical Tests check None. Click the Next Step.

  7. In the Functional Analysis - Simulated Change Dialog: Reset Functional Change option to clear the settings. Now, using the Parameter Change Table, enter the individual's parameters for up to three options. (If you have more than three options, you will need to rerun this procedure to change the remaining options.) All parameter entries for each option should be completed. If you do not have specific requirements for skewness or kurtosis, then enter zero in each case. (Note that transformation to extreme skewness or kurtosis may well fail. There are technical limits to the distortion possible. Keeping skewness or kurtosis transformations within plus or minus 0.75 will usually work OK.) Lastly, check the Change Inter-Correlations check box and click Next Step.

  8. In the Functional Analysis - Simulated Parameter Change - Correlation Dialog: Unless you have specific requirements for correlations, it is recommended that you randomise all option ratings at this stage. You can carry out any further changes to correlations in the course of later Options Analysis. Click on Next Step. If you are using a very small sample, e.g. 10, there may be difficulties in randomising to the criterion level. You may need to loosen the criterion level if this problem arises, or, try drawing another sample.

  9. In the Analysis Summary Dialog: It is essential to check the box for Analysis Data for your Functional Change Analysis and to check the request for "Both" statistics and casewise output at the bottom of the dialog. Then click Next Step and allow the analysis to proceed.

  10. Ignore all the resulting output except for the data to be found on the FunctChangeDat sheet under the heading Scenario Appeal. This is the data which becomes your raw input for Options Analysis. Above the data is a summary of the parameter settings which you can check against your requirements. To examine the inter-correlations you will need to run a Fundamentals Analysis. (If you needed to configure more than three options, you will need to repeat the above process until you have changed the parameters for the remaining options.)


Collecting individual uncertainty data

Notes on uncertainty, probability, risk and Options Analysis

Notes index


Copyright 2012 Corporate R & D Ltd
All rights reserved. Email: