Word Reading Comprehension - Image Selection

Platform: HTML · German · verified 77d ago

This is a task in the LEAPS Zurich study. Click to know more.

Local data export available

This task saves data locally in your browser and offers a download at the end of the session. For production use, you will need your own server-side persistence.

Description

Pixi.js + TypeScript educational game built for the LEAPS Zurich learning-progress study. The child sees a German word (e.g. "Tor") or a short sentence (e.g. "Der Bus ist blau.") together with four candidate images and taps the one that matches; the target text is also played as audio for support. The task ships as a multi-scene flow — two animated demo trials, three practice trials with feedback, a transition, then 60 main test items, finishing on an end scene that bundles the data. The test scene caps at 7 minutes and shows a 'stop frog' early-exit overlay after 30 seconds of inactivity. Each trial logs the four shown images, every press/release event, every audio replay, the final selection, and per-trial timestamps.

Contact & Lab

Lab

Developmental Cognitive Neuroscience Team

kjpd.uzh.ch

Principal investigator

Silvia Brem

silvia.brem@kjpd.uzh.ch

University of Zurich

Task contact

Iliana.Karapidis@kjpd.uzh.ch

Contributors

Iliana Karapidis

University of Zurich

Conceptualization

Supervision

Nicoletta Klotz

University of Zurich

Data Curation

Investigation

Hugo Farajallah

University of Geneva

Software

Citation

No citation available — contact the task author.

Credits

No credits provided.

Domains

Vocabulary

Reading

Psycholinguistics

Part of suite:

LEAPS Zurich

Target Population

Age Categories

primary

Clinical Status

non-clinical

Additional Notes

German-speaking primary-school children, ages 6-8 (LEAPS Zurich cohort).

Languages

German

Run under

Engine

HTML / JavaScript

Format

Online + in-lab

Modalities

behavioral

Task Outputs

What it measures

Reading comprehension at the word and short-sentence level: how well a child maps written German text onto a depicted referent. Audio support disentangles decoding from comprehension when needed.

Output formats

json

One AppResults JSON per session. In browser preview, the file downloads automatically at the end. In LEAPS Android deployment, the same payload is forwarded via window.System.sendData to the LEAPS backend. Top-level shape: { practice: SceneResult | StopFrogResult | null, test: SceneResult | StopFrogResult | null }.

Experimental design

Trial-based

60 trials

Within-subject: scene

Conditions

scene

demo

practice

test

— Demo (2 animated story items, no scoring), practice (3 items with feedback), test (60 scored items, 7-min cap).

Key measures for analysis

Test accuracy (% correct on the 60 test items)
Practice accuracy (% correct on the 3 practice items)
Mean response time per trial (ms)
Audio replays per trial (proxy for decoding difficulty)
Selection-revision count (selections.length per trial: how often the child changed their mind)
Stop-frog rate (proportion of sessions terminated early by inactivity / explicit stop)

Recorded variables

Variable	Role	Type	Description	Unit
`practice`	dependent	object	SceneResult for the practice phase, or a StopFrogResult if the child exited early. Null if practice was skipped via routing.	-
`test`	dependent	object	SceneResult for the test phase, or a StopFrogResult if the child exited early (7-min timeout, 30 s of inactivity, or pressed the stop frog). Null if skipped.	-
`answers`	dependent	array	Inside each SceneResult: an AnswerResult per trial (see fields below).	-
`totalCorrect`	dependent	number	Inside each SceneResult: count of correct answers in that scene.	-
`totalTime`	metadata	number	Inside each SceneResult: total time spent in the scene.	ms
`reference`	independent	string	AnswerResult field: the target text shown to the child (a German word or short sentence).	-
`trialIndex`	metadata	number	AnswerResult field: 0-based trial index within the scene.	-
`startTimestamp`	metadata	number	AnswerResult field: timestamp when the trial started.	ms
`endTimestamp`	metadata	number	AnswerResult field: timestamp when the green-arrow submit was clicked.	ms
`answerOptions`	independent	array	AnswerResult field: the four images shown, each with imageFilename, isTarget (true for the correct option), and choiceIndex (1=top-left, 2=top-right, 3=bottom-left, 4=bottom-right).	-
`audioPlaybacks`	metadata	array	AnswerResult field: every time the audio of the target was played, with timestamp and audioFile. A single trial can have several entries (the child can replay).	-
`selections`	dependent	array	AnswerResult field: every click on an option, with timestamp, choiceIndex, and isFinal (true for the one submitted).	-
`isCorrect`	dependent	boolean	AnswerResult field: whether the final selection matched the target.	-
`responseTime`	dependent	number	AnswerResult field: time from trial start to final-submit click.	ms