Blind-test design and task wording eliminate bias in hearing studies
Evidence
In any case, it is important to eliminate all sources of error and bias in psychoacoustic tests. The ideal situation is a blind test, where subjects have as little information as possible on the sounds they are hearing. The wording of the questions or tasks is very important in this respect. — Ville Pulkki, Matti Karjalainen, Communication Acoustics- An…, 907
Categorization
Category: Causal Relationships
Related
- Valid psychoacoustic listening tests require blind conditions and carefully worded tasks. (similarity 0.91)
- Blind listening tests make subjective sound judgments objective and repeatable (similarity 0.81)
- Listening-test statistics must inform experimental design from the outset (similarity 0.80)
- Psychoacoustic knowledge rests entirely on subjective listening tests, not direct measurement (similarity 0.80)
- Engineers distrusting blind tests converted upon seeing correlations with their measurements. (similarity 0.80)
- Careful listening tests validly reveal how acoustic attributes become auditory attributes. (similarity 0.80)
- Unpredictable sound presentation prevents subjects from faking hearing-loss test results (similarity 0.80)
- Randomizing identifiers, positions, and music prevents experimenter bias in listening tests (similarity 0.79)
- Blind-tested listeners with normal hearing produce reliable, measurement-grade subjective ratings (similarity 0.79)
- Blind tests collapse personal preference into shared judgments (similarity 0.79)
- Headphone blind tests cannot use automated shuffling, so tactile cues leak into results. (similarity 0.79)
- Smaller step sizes in audiometry detect feigned hearing loss (similarity 0.79)
- Game-like testing yields less biased hearing-aid results in children than formal tasks. (similarity 0.79)
- Listener personality biases reported auditory thresholds across subjects (similarity 0.78)
- Accurate anechoic data plus blind listening tests make psychoacoustic correlations emerge quickly (similarity 0.78)