3. The task inclusion test

Each task has a three-part inclusion test that defines the boundary of the task: what the participant does, what the experimenter varies, and what is recorded.

3.1 Structure

The test is three lists on the task record:

"inclusion_test": {
  "procedure": [
    "A continuous stream of stimuli is presented.",
    "Participants indicate when the current item matches the one presented N items back."
  ],
  "manipulations": [
    "Load level (N = 0, 1, 2, 3)",
    "Stimulus type (letters, locations, faces)",
    "Lure trials (N+-1 matches)",
    "Dual N-back (two simultaneous streams)"
  ],
  "measurements": [
    "Hit rate, false alarm rate, d-prime",
    "RT",
    "Load-dependent accuracy decline"
  ]
}

  • procedure - what the participant does, as ordered steps. Each step is a sentence naming an observable event: a presentation, a response, an instruction. The order is the order of a trial or block.

  • manipulations - what the experimenter varies: one independent variable per item, with its typical levels in parentheses.

  • measurements - what is recorded: one dependent variable or signal per item.

Every list has at least one item. The steps of the procedure are the level at which the task meets CogPO’s conditions and the Catalog’s facets: each step names a stimulus, a response or an instruction.

3.2 Using the inclusion test

A specific experiment is an instance of a given task if and only if:

  1. Its procedure matches the procedure steps (the participant performs the described sequence of actions)

  2. It manipulates at least one variable from the manipulations list

  3. It records at least one measure from the measurements list

The procedure is the most important part. Two experiments with different manipulations or measurements but the same procedure are instances of the same task. Two experiments with different procedures are different tasks, even if they target the same cognitive construct.