Skip to content

break

Experiments Surveys & forms

break turns the row it is on into a deferred break screen. The row is lifted out of the normal trial sequence and displayed exactly once, when its scheduling condition is met: either a number of trials has elapsed or a certain amount of wall time has passed. Scheduling is checked only between trials, never mid trial.

The break value says WHEN the break appears, not how long it lasts. The break row itself is rendered like any ordinary trial row, so its duration and appearance come from the row’s own columns (content, presTime, button1, keyboard, etc.). To end the break, the participant completes that row like any other trial; the experiment then resumes from the interrupted position.

Put the number of trials after which the break should appear in the cell: 10 shows the break row at the first opportunity after 10 trials have run.

Time-based breaks add a unit:

  • Ns — after N seconds (90s).
  • Nm — after N minutes (2m).
  • NmMs — minutes plus seconds (1m30s is 90 seconds).

Only the first run of digits in the cell counts, so keep the cell to exactly one schedule value. The clock for time-based breaks starts when the experiment page loads, unless you restart it mid-experiment via the then column (see Works with).

ValueBehavior
empty cellOrdinary row, stays in the trial flow
bare integer N >= 1Break shown at the first opportunity after N trials have run
NsBreak shown at the first trial boundary after N seconds of clock time
NmSame, minutes
NmMsSame, combined
0 or 0sRow is treated as a break row (so it is skipped in the normal flow) but never scheduled, so it never displays
no digits at allNot a break row; behaves like an empty cell

Leaving the column out, or the cell empty, keeps the row as a normal trial. Any cell whose first run of digits parses becomes a break schedule; a cell with no digits at all is treated exactly like an empty cell.

  • Row content: whatever the break row defines (content, title, presTime, buttons, keyboard, feedback…) is presented normally when the break fires. A presTime makes the break auto end; a button makes it participant paced.
  • then on the break row: after the break, flow resumes at the then destination instead of the interrupted trial.
  • then: timer: start on any other row restarts the break clock when that row runs. If any row in the file carries such a then, ALL time-based breaks wait until it has run, and a time break never fires on the very trial that started the clock.
  • subjectGroup on the break row: after group allocation, a break row belonging to a group the participant was not allocated to is dropped; rows whose group cell is empty or 0 always stay.
  • Randomization / picking: if the break row was excluded from the run (for example by randomPick or subject-group filtering), the break is silently skipped and the flow continues.
  • Trial counting: every displayed trial of any type counts once toward trial-count breaks (instructions included); the displayed break itself does not count toward the next break’s threshold.
  • End of experiment: a final break row keeps the experiment alive until it has displayed; the experiment only ends on the last trial when that row schedules nothing further.
  • Preview mode: breaks are not deferred in the previewer; break rows simply display inline in file order.
  1. Break after 10 trials, participant paced. The break row can sit anywhere in the file (commonly at the end); it is pulled out of the normal order:
typecontentbreakbutton1
teststimulus trial 1
test...
instructionTake a short rest. Press Continue when ready.10Continue

After the 10th trial has run, the instruction row appears; clicking Continue resumes with the trial that was about to run.

  1. Timed break 2 minutes in, auto ending after 30 s on screen:
typecontentbreakpresTime
instructionPlease rest for 30 seconds.2m30000

The break appears at the first trial boundary after 2 minutes of run time and disappears after its own 30000 ms presTime.

  1. Clock started mid experiment via then:
typecontentthenbreak
instructionPractice done. The timed block starts now.timer: start
instructionHalf way rest.90s

The 90s break counts from the moment the first row’s then executed, not from page load.

  • Decimals do not work: 1.5m reads only the leading 1 and becomes a break after 1 trial.
  • At most one break fires per trial transition; when several are due, time-based breaks are checked before trial-count breaks, and the rest wait for the next transition.
  • A time break fires at the first trial boundary after its threshold; a long trial delays it.
  • 0 or 0s creates a row that is skipped forever and never shown (see Options).
  • A stray | in the cell is read as a seconds unit (5| is 5 seconds); keep the cell to a clean value.
  • Under randomization, the bookkeeping that resumes the flow around neighbouring break rows can misjudge; avoid relying on two break rows sitting adjacent in a randomized section.
  • While a break is on screen, other due breaks wait; they queue up rather than nest.