ChoiceAnalyst FAQ

1. What is ChoiceAnalyst?

When you are confronted with a decision that requires you to choose the best choice amongst a group of options, and you need to judge these options based on a list of criteria, you have a formidable task in front of you. ChoiceAnalyst is a program which assists you in analyzing and making those decisions.

ChoiceAnalyst accomplishes this by taking that large, complex decision and breaking it down into a series of the simplest decisions you could make. While considering just 1 of the criterion, you are presented a random pair of your options and all you need to do is, on a sliding scale, decide which one is better and by about how much. You continue to make these very simple two choice comparisons for all of your criteria and then ChoiceAnalyst mathematically calculates the best choice for you.

2. How does ChoiceAnalyst help me make the best decision?

ChoiceAnalyst provides the fundamental ability to create decision models. ChoiceAnalyst facilitates a structured process in creating these models in a simple, effective,and straightforward way.

However ChoiceAnalyst goes beyond just structuring the decision process with a decision model, its calculations are derived from decades of research in operation research mathematics. The decision models facilitated by ChoiceAnalyst places you in control of the definition of your decision model. However, once you define your decision model, ChoiceAnalyst will create data derived from that definition and ask you to make simple two choice comparisons to arrive at a solution that best optimizes your criterion for the best choice. The process of using ChoiceAnalyst in defining your decision model, evaluating your decision model, and calculating your best decision against the decision model data is supported on both subjective and objective basis in a very simple and straightforward way. This is what sets ChoiceAnalyst apart — its overall simplicity.

3. Is ChoiceAnalyst the same as or based on Saaty's Analytical Hierarchy Process?

ChoiceAnalyst is not based on the analytical hierarchy method of Saaty. ChoiceAnalyst takes a very direct approach using a gradient free derivative and makes use of probability and statistic functions in generating a sampling of generated pairs. The statistical sampling is calculated based on a user supplied (or default) confidence level and confidence interval model option. A best choice solution is found using numerical iterations that refine calculations until a solution is found that best satisfies the decision maker and criteria weights along with the paired preference weight matrix. The number of iterations to reach convergence and therefore a solution is dependent on the particular model involved.

4. Why does the value of weights go as high as 9999?

Since decision maker or criteria weights are relative to one another, a weight range from 0-9999 allows a weight difference as small as 1/10 of one percent between decision makers or criteria. If the user needs relative difference in influence between decision makers or criteria on the choice ranking solution as small as 1/10 of one percent, then weights should differ from one another by one. If the user needs a relative difference in influence between decision makers or criteria on the choice ranking solution of one percent, then weights should differ from one another by ten.

Overall the degrees of freedom in choosing weight differences is wide ranging to support what ever is required for a given decision model.

5. What if I don't know the values to set for weighting Criteria or Decision Makers?

Build a model in which the 'Choices' you enter are the list of Criteria that you want to use in your pending decision model. Then, enter just one Criterion, 'More important to me'. You can simply list yourself as the only DecisionMaker, or if the pending Model is going to have multiple DecisionMakers, you could include them in this model also if you want their opinion on the Criteria weighting.

Once you go through the pairings of your Choices (the listed criteria) and made your selections, then when ChoiceAnalyst calculates the results of this model, you can take the percentages in the '% Rank' column and use them as the weights for the Criteria in your pending model. You can also take this same approach if you are building a model with several DecisionMakers and you are not sure what weighting to give them. You could even use multiple criteria for this one, such as, 'Has more experience in this area' or 'Has more knowledge on this issue', etc.

6. How can I collaborate with the other Decision Makers listed in a model?

ChoiceAnalyst stores all model data in an xml file. If multiple decision makers are created the decision model xml file will store these decision makers. Once the model is stored to a file, each decision maker exchanges the model file as completed by them with the next decision maker in the model. Each decision maker can evaluate paired preferences (comparing 2 choices against a given criterion), save the data to the xml model file and provide the xml file to the next decision maker that would reside on another computer. There are myriad ways of exchanging data from 1 computer to another and all the means of achieving this can't be listed here, however some common methods of data exchange from one computer to another are through a shared file server, e-mail with data file attachments, or a portable flash drive.

7. What if I have more than ten choices I'm trying to choose between?

Let's assume that you have a decision problem that contains 20 Choices. All you would need to do is first, create two models using the same criteria and decision makers and the same weighting for each. List 10 unique Choices in one and the second 10 unique Choices in the other. Now simply go through both models. When the results are calculated, build a final model. Again, in this model leave the criteria and decision makers the same along with the weights assigned to each.

For your Choices in this model, take the TOP 5 Choices from each of the initial models and enter those 10 into this model. Once you go through the pairings for this model and calculate the results, you will have the best choice reflected from the original 20 Choices.

8. What options can I change that impacts the decision processing?

Chapter 8 in the ChoiceAnalyst User Guide, "Solution Calculation, Display and 'What If' Analysis" describes those options, such as, Confidence Level, Confidence Interval, Preference Scale, etc.

9. What are the system requirements for running ChoiceAnalyst?

ChoiceAnalyst is designed for the Windows 7, Windows Vista and Windows XP desktop operating systems and will conform to the minimum memory requirements of these platforms. ChoiceAnalyst was designed to have very low RAM requirements, loads very quickly, and uses very little resources during its computations. ChoiceAnalyst has also been tested using WINE with Linux and Parallels on the Mac, however support will not be provided if it fails to function on these systems.

The installation will require at least 20 megabytes of disk space which is a very small footprint by todays standards. During its operation ChoiceAnalyst creates 2 types of files, decision model XML files and PDF report files. Even the largest decision model files themselves are generally less than 500K while the PDF files are also less than 500K. The PDF files are generated from within ChoiceAnalyst and require no third-party tools, printer drivers, or the latest version of Adobe Reader. ChoiceAnalyst is a very self contained application.

Adobe Reader 4.0 and above is required to read the generated PDF files although ChoiceAnalyst can generate PDF files even if a PDF reader is not available. ChoiceAnalyst also uses the optional compression feature of the Adobe PDF specification so the resulting PDF files that are generated from a model are very small. Overall ChoiceAnalyst was designed to run a wide variety of environments and has very low supportability requirements. The Adobe Reader typically comes pre-installed with the Windows OS provided with the computer at the time of purchase; however, if one wishes to install the product or upgrade, Adobe Reader is available at no cost by going to www.adobe.com

10. Why should I rely on the decision results calculated?

ChoiceAnalyst calculations are based on a solid foundation of years of mathematical research and application, accurately computing a solution based on data that you enter.

There are three primary inputs to the decision calculation process. The first is how one assigns decision makers' weights relative to each other, the second is how one assigns criteria weights also relative to each other, and finally the third input is how two choice comparisions are weighted relative to each other along a number scale. The process is very intuitive and relies on the user to provide these inputs.

ChoiceAnalyst gathers this data from a numerical database and calculates the best decision, the next best decision and so on in ranking order. You can rely on ChoiceAnalyst because it sits on a solid mathematical foundation and it works with you to provide the most optimal decision given the outcome that you specify in terms of criteria. ChoiceAnalyst facilitates the decision model process because it tracks all the data, comparisions, and calculations for you, so you can concentrate on specifying the model and the judgements regarding what is the best choice for only two choices at a time.

See the question "How does ChoiceAnalyst help me make the best decision?" for more information on how ChoiceAnalyst enables effective decision making.

11. Why is selecting between pairs a good methodology to use?

Paired preferences techniques have been available for some years and have been used effectively in a wide variety of decision modeling capabilities. They provide a very simple means of characterizing a user preferences of one choice over another, and allow for optimizing the pairs in such a way that one can reduce the total number of pairs that require evaluation by the user. This effectively shortens the time and effort to provide the data required for a best choice calculation.

12. Does ChoiceAnalyst support "What-If" analysis?

What-If analysis is useful to compare impacts of weight changes for either criteria or decision makers or both. What-If analysis also considers paired preference weight or score changes to see the impacts of changing the model. The weights of decision makers, criteria, and paired preference scores (weights) can have an impact on changing the best choice calculation and therefore changing what is the best choice for a given decision.

What-If analysis is supported by design with three facilities in ChoiceAnalyst. First It can be used with multiple calculation windows, second by creating a report to a PDF file for a more permanent storage and third by loading more than one ChoiceAnalyst application for copies of models and comparing the differences between the different application windows.

Consider a scenario in which you are a decision maker in a decision model but not the only one. If you want to see the outcome of a decision if only you were the only decision maker, then set other decision makers weights temporarily to zero and see what happens. This decision now reflects only your paired preference scores. Or if you want to see what would happen if a criteria were put aside, then set its weight to zero and compare the BestChoice calculation before and after. Finally you could change your paired preference scores and see if the model calculations are different. They may not be different if the change to the paired preference scores does not have a big enough impact.

When going through these What-If analysis, it is best to generate a report after each so you have a permanent history of each analysis to compare. However one can just visit the multiple calculations windows to see the impacts of the changes

13. Why not trust my intuition instead of ChoiceAnalyst?

ChoiceAnalyst is not making the decision in place of you, it's simply organizing a sequence of decisions you've already made. It computes the ranking of choices based on both objective and subjective evaluations made by you, such as weights for each decision maker and your preference scores for each choice pair.

ChoiceAnalyst doesn't replace intuition, it supplements and amplifies it.

14. Are there examples that show how to use ChoiceAnalyst?

ChoiceAnalyst includes a number of different example models, as well as a comprehensive QuickStart Guide that will help you become productive using the software right away. There are also several examples you can watch in our video tutorial.