Preference Elicitation task

Also colloquially known as called "Voting", this task is about choosing the relative value or weight between a number of options. On every trial, a given number of Stimulus (images, text, etc) options are displayed, and by clicking or right-clicking on the stimulus, the participant can respectively give or take away one vote for that option. The option's relative weight is displayed as a bar above the stimulus.

Constraint types

no constraint

There is no overall constraint and the value of votes is always 1. Examples: separate binary probability elicitation, range voting

L0 constraint

The 1-hot or k-hot choice of a subset, any option is either selected or deselected.

L1 constraint

Here all options' weights together sum up to a single value. Examples: probability distribution, resource allocation, budget allocation

L2 constraint

As the L1 constraint, but additional votes' value diminishes quadraticly. Examples: quadratic voting, quadratic funding, quadratic preference elicitation

rank constraint (no ties)

In this case, a vote for an option directly switches its rank with the next highest ranking option. Examples: ranked-preference voting, instant runoff voting

rank constraint (with ties)

Similar to the previous, but this allows two or more options to hold the same rank.