You have chosen to sponsor your bid up to a maximum amount of .
We're creating a test that will use Machine learning / AI to structure itself each time it is taken.
We're looking for someone with higher education in psychology to guide us in coming up with questions that will get to the core of what we are trying to accomplish.
Objective: To create questions that are directly relevant to differentiating the interests of participants into clusters that have a one to one matching with academic majors. In English, we want the questions that clearly separate individuals into buckets where on one side we have people with a specific major and everyone else in the other.
Approach: The questions themselves should be varied enough to tackle various aspects of life, with the hope that the questions do a great job of separating individuals while maintaining a sufficient diversity in the responses. Inquiries should be fun, succinct, clear, but not obvious nor trite.
Bad: Do you like physics?
Good: Do you have a natural curiosity about the physical world around you?
As a tactic, it may be smart to break the questions down into categories:
Does this statement describe you? “Ideas about magnetism and electricity has always fascinated me”.
“When you start a task do you finish it until the end?”
Algorithmicly, because of the robustness of the current prototype model there are no limitations to the question types. We can support any of the following question types:
Binary – Which do you prefer? Guns or Bread
Multiple Choice - Which do you prefer? Guns , Roses, Bread or Milk
Interval Scale – On a scale from Madator (1) to Boxer (10), How do you handle conflict? Or On a scale from 1 to 10, how adept are you at extemporaneous thought?
Sort – Rank these items in order of their entertainment value: Netflix, Reading, Exercising, Camping
Of these question types, the ABSOLUTE best are interval based inquiries because they allow for a continuous scale of answers to differentiate participants. Sorting questions can be very powerful as well because of the amount of information contained in the sequence. Multiple question types would be ideal, but the focus should be on interval and sorts.