Subtitle Exit Lab · about 25–30 minutes

What exactly do subtitles
do instead of your listening?

EarShift finds the moment where natural English breaks down, reveals your specific trap, and tests the same skill again without text.

ONE LINE · THREE PASSES

01No textan honest first attempt

02Audio X-rayEnglish support explains the sound

03Caption offanswer again without support

Honestly: EarShift has not yet shown that it improves film comprehension or reduces subtitle dependence. This version first measures a difficulty, then tests short-term change on new material.

24 signals for the map4 + 4 controlled probes and film scenes before and after2–4 full learning loops without script duplication

RESEARCH DEMO · 01

Who is conducting this study: EarShift is an independent product research project for training spoken-English perception.

With your consent, we will store an EarShift code, background responses, task results, response times, replays, and final feedback. If you came through Prolific, we will separately store your Prolific participant, study, and session IDs so we can match the response to the submission and pay you correctly.

We do not ask for your name, email, phone number, or exact address.
A Prolific ID is pseudonymous and will not appear in public results.
The order of parallel forms A/B is assigned automatically from the participant code so their difficulty can be compared across groups.
Benefits for film comprehension or reduced subtitle dependence have not yet been established.
This is not a medical or hearing assessment.
Participation is voluntary; you can close the page before finishing.

How the data is used and protected

Purpose

To evaluate EarShift's clarity, usability, and research value. This is not a hearing test, medical assessment, or certified language exam.

Data we store

An EarShift participant code; for Prolific participants, separately stored PROLIFIC_PID, STUDY_ID, and SESSION_ID; first language, English level and experience, country and subtitle habits; task answers, accuracy, response time, replays, recording and speaker metadata; A/B form assignment, before/after results, training answers, and final feedback.

Use and storage

Data is stored in EarShift's research database on OpenAI Sites and Cloudflare infrastructure. We use it for product analysis and may include only de-identified aggregate findings in documentation, presentations, or funding applications. We do not sell responses or use them for advertising.

Retention and deletion

Data will be kept for up to 24 months, then deleted or irreversibly de-identified. To request access or deletion before aggregation, message the researcher through Prolific Messages and include the EarShift code shown on the next and final screens.

Contact

For questions, complaints, or data requests, contact the EarShift researcher through Prolific Messages in this study's listing.

I am 18 or older. I have read the data information and voluntarily consent to participate.

What exactly do subtitlesdo instead of your listening?

What exactly do subtitles
do instead of your listening?