Paired presentation

Present items in fixed pairs whose order shuffles for every participant while the members of each pair always stay together on screen.

When to use it

Use this whenever stimuli belong together in yoked units that must never be separated by randomisation: paired-associate learning (cue and target studied side by side), yoked prime and probe items, or any design where the unit of randomisation is a pair rather than a single stimulus. The pairs appear one pair per screen in a random pair order, each pair intact, and a later test probes what was studied.

Trial file

type	content	stimFormat	stim	stimScreens	stimOrder	presTime	ISI	keyboard	key	button1
instructions	Learn the word pairs. Each pair appears together for two seconds. Afterwards you will be asked which words were paired.									NEXT
learn		word	BREAD;CLOUD;NAIL;RIVER;LAMP;HORSE;GLOVE;STONE	2	random(chunk 2)	2000*4	500*3
test		word	Which word was paired with NAIL? Press R for RIVER or H for HORSE.					r h	r
test		word	Which word was paired with GLOVE? Press S for STONE or C for CLOUD.					s c	s

How it works

The study row lists eight words in stim as four authored pairs, in pair order: BREAD with CLOUD, NAIL with RIVER, and so on. stimScreens = 2 groups them two per screen, so each authored pair occupies exactly one screen, and presTime = 2000*4 with ISI = 500*3 runs the four screens at two seconds each with half-second gaps. On a learn row the trial auto-advances through the screens, so the study phase plays hands-free.

stimOrder = random(chunk 2) is what makes the pairs the unit of randomisation: consecutive units of two shuffle as wholes, and the order inside each unit is preserved. A participant might see the LAMP pair first and the BREAD pair last, but BREAD is always next to CLOUD and always on the left of it. The arrangement is drawn once per session and recorded in the stimOrder_actual results column, so the analysis knows each participant’s study order, for instance to model serial-position effects.

The test rows probe the associations as forced choices scored by key. They are ordinary single-stimulus rows and need none of the grouping columns.

Variations

• Unequal units: random(chunk 2+3+3) shuffles a pair and two triplets as wholes; the sizes must sum to the stimulus count, and stimScreens can mirror them (2+3+3) so each unit still fills one screen.
• Swap members within a pair as well: a single spec cannot chunk and swap at once, so author both member orders as separate rows and sample between them, or accept the authored inner order (the usual choice when the left item is the cue).
• Pairs across sequential single screens: keep random(chunk 2) but set stimScreens = 1, so each pair plays as two successive single-item screens (cue then target), pairs still travelling together.
• Image pairs: stimFormat = .jpg with image names in stim; everything else is unchanged.
• A response deadline on the tests: responseWindow = 5000 records timeout when no choice comes in time.