Stream, then questions

Present a sequence of stimuli on one self-advancing study trial, then ask two scored questions about it, with each stream and its questions shuffled as one unit.

When to use it

Use this when participants first watch something and only then answer questions about it: memory for a scene sequence, judgements about a set of items, two follow-ups after one exposure. Packing the questions into the presentation trial does not work, because one trial collects one response; and leaving the rows unprotected during randomisation would let a question drift away from its stream. The split here is one learn row for the presentation, which advances by itself and collects nothing, followed by one test row per question, with randomBlock holding each stream and its questions together as a shuffleable unit.

Trial file

type	content	stimFormat	stim1	stim2	stim3	presTime	ISI	randomBlock	keyboard	key	button1
instructions	You will see three objects, one after another. Then answer two questions with Y or N.										NEXT
learn		.png	apple	hammer	bicycle	1500*3	500*2	1
test		word	Did you see a hammer?					1	y n	y
test		word	Did you see a guitar?					1	y n	n
learn		.png	candle	scissors	teapot	1500*3	500*2	2
test		word	Did you see a teapot?					2	y n	y
test		word	Did you see a torch?					2	y n	n
learn		.png	drum	ladder	envelope	1500*3	500*2	3
test		word	Did you see a ladder?					3	y n	y
test		word	Did you see a whistle?					3	y n	n

type,content,stimFormat,stim1,stim2,stim3,presTime,ISI,randomBlock,keyboard,key,button1
instructions,"You will see three objects, one after another. Then answer two questions with Y or N.",,,,,,,,,,NEXT
learn,,.png,apple,hammer,bicycle,1500*3,500*2,1,,,
test,,word,Did you see a hammer?,,,,,1,y n,y,
test,,word,Did you see a guitar?,,,,,1,y n,n,
learn,,.png,candle,scissors,teapot,1500*3,500*2,2,,,
test,,word,Did you see a teapot?,,,,,2,y n,y,
test,,word,Did you see a torch?,,,,,2,y n,n,
learn,,.png,drum,ladder,envelope,1500*3,500*2,3,,,
test,,word,Did you see a ladder?,,,,,3,y n,y,
test,,word,Did you see a whistle?,,,,,3,y n,n,

How it works

The learn row is the stream. On a learn trial the response window never opens, so nothing the participant presses does anything, no response is collected, and no results row is written; the trial advances on its own as each presTime entry elapses. 1500*3 gives the three images 1500 ms each, and 500*2 puts a 500 ms blank between them, so the stream plays hands-free from start to finish.

The two test rows that follow are ordinary scored trials: a question as a word stimulus, y n on keyboard, and the correct answer in key. Their results rows carry the accuracy and reaction time for each question.

randomBlock is what keeps each triplet together: the learn row and its two questions share a block number, so the three units shuffle as wholes and the questions always follow their own stream. randomBlock is the right tool here rather than a trialOrder structure spec, because spec runs split into separate pools wherever type changes, and this unit deliberately mixes learn and test rows.

Variations

Ask for typed recall instead of yes/no: make a question row responseType = box with the expected answer in key, and the typed text is scored against it.
Longer or uneven streams: add stim4, stim5, … and extend presTime/ISI per screen (1500;1000;2000 times each screen individually).
Randomise the question order within a unit too: give the two question rows of a block the same random tag (for example 5 for the first unit, 6 for the second); they then swap order among themselves while randomBlock still moves the unit as a whole.
For a single question after each stream, this is the same pattern with one test row; for questions about a stream shown segment-by-segment under participant control, see the Self-paced reading recipe.