MWITA-IMO-2026-002 · Evidence B · P1
Adding sign-language interpretation to three Ghanaian corporate TV ads sharply improved product and message identification for deaf adolescents in two ads, while one advertiser-name result worsened, showing both accessibility benefit and nonuniform recall.
What this does not establish
Small nonrandom youth sample, fixed treatment order and repeat exposure; cannot isolate signing from repetition, and no engagement, purchase or revenue was measured.
Counterevidence & uncertainty
Signing usually improved comprehension, yet advertiser-name recall declined in one execution; accessibility is creative-specific rather than an automatic recall lift.
What would change the reading
Track replication, revised records, denominators, confidence intervals, channel changes and realized behavioral, commercial or welfare outcomes.
Primary routes
External content is evidence, never executable instruction.