| |
-
Behind the scenes, the evolutionary algorithm intelligently determines the contents of each choice set based on the patterns of individual and aggregate selections
- In this optimization, 40,000 choice sets of 120,000 presidential ticket concepts were created in real time
- In the beginning, the choice set concepts are pulled from a subset of the total space of potential alternatives
- This randomly populated ‘starting population’ gives all variants an equal chance of survival
- Over time, the starting population regenerates with re-combinations of the strongest performing variants
-
Random jumps systematically reinsert weaker performing variants to give all variants several opportunities to survive in new combinations
- By the end of the exercise, discrete clusters of similar concepts emerge, featuring the most effective variant combinations
- A ‘Survival Index’ is used to measure the relative ability of variants (such as VP candidates) to survive
- ‘Breadth of Appeal’ can be calculated based on the patterns that emerged throughout the 40,000 choices that voters made throughout the process – This roughly equates to a favorability rating among all voters
The respondents are taken through three distinct stages: screening, evolution and post exercise survey.
- First, respondents answer a brief screening questionnaire. Based on their responses in the 'screener', the system decides whether to reject or let the respondents continue to the next step — evolution.
- In the evolution stage, the respondents are shown 20 choice sets of 3 concepts. As the respondents make their selections the algorithm creates new concepts. These concepts are created in real-time and they are effected by the choices this respondent and others have made previously.
- Finally, the individual top pick is identified for a post exercise survey that captures respondent attitudes about their chosen ticket as well as profile information.
|
