Replayed random order

Shuffle a block once, then replay that exact order in a later block so both passes share identical order effects.

When to use it

Use this for pre-post designs where the same items are presented twice and the change score is what matters: if each pass gets its own independent shuffle, differences between passes mix the effect you care about with order differences, while holding both passes to one fixed file order gives every participant the same sequence and invites item-order confounds across the sample. Recording the first pass’s shuffled order and replaying it in the second pass keeps the order random across participants but identical within each participant. If the two passes may legitimately differ in order, two plain shuffles (1: shuffle and 2: shuffle) are simpler.

Trial file

type	content	stimFormat	stim1	trialOrder	keyboard	button1
instructions	Rate each image from 1 (unpleasant) to 7 (pleasant).					NEXT
test		.png	forest	1: shuffle; saveOrder firstPass	1 2 3 4 5 6 7
test		.png	street	1	1 2 3 4 5 6 7
test		.png	lake	1	1 2 3 4 5 6 7
test		.png	carpark	1	1 2 3 4 5 6 7
instructions	Take a short break. You will now rate the same images again.					NEXT
test		.png	forest	2: useOrder firstPass	1 2 3 4 5 6 7
test		.png	street	2	1 2 3 4 5 6 7
test		.png	lake	2	1 2 3 4 5 6 7
test		.png	carpark	2	1 2 3 4 5 6 7

How it works

The first pass is an ordinary shuffle: its four rows share the group tag 1 in trialOrder, and the first row’s cell carries shuffle; saveOrder firstPass, so the group is shuffled and saveOrder then records the order it landed in under that name. The shuffle word is doing real work: a trialOrder group keeps its trial-file order unless a word moves it, so 1: saveOrder firstPass on its own would bank the file order and replay that, which is not a random order at all. The second pass’s rows share the tag 2, and its first row carries useOrder firstPass, which replays the saved order: rows are matched by their rank in the trial file, so the nth row of the second group lands wherever the nth row of the first group ended up. Because both groups list the images in the same file order, each participant rates the images twice in exactly the same sequence. Write each spec on the first row of its group only; the remaining rows take the bare tag.

The break screen between the passes leaves its trialOrder cell empty, which keeps it in place and splits the two passes into separate groups. That separation is required anyway: the two groups must be distinct, useOrder must reference a name saved by an earlier group, and each saveOrder name is saved exactly once. useOrder must also stand alone in its group’s cell, with no other specs alongside it. Give both groups the same number of trials; a replay onto a group of a different size is an ordering error, and any such error halts the session with an error page rather than running in a wrong order (in preview mode the rules are relaxed).

One structural constraint to respect: a saveOrder or useOrder run must stay within one randomBlock, one type and one timer-label span, because these specs are rejected when their run splits into several pools. Here every specced row is a plain test trial with no randomBlock value, so each group is a single pool. The ratings themselves need no scoring, so there is no key column: keyboard lists the digits 1 to 7 and the pressed digit is recorded as the response.

Variations

• Replay the order a third time: a later group tagged 3: useOrder firstPass in trialOrder, with the same number of rows, repeats the sequence again; a saved name can be replayed by any later group.
• The second group’s rows do not have to repeat the same stimuli: matching is by file rank, so listing paired probes in the same file order presents each probe at its partner’s first-pass position.
• Make the passes distinguishable at analysis time by tagging them pre and post in a condition1 column, which is copied into every results row.
• Jitter the pace without touching the order: an ITI value on the test rows adds a blank interval before each image.