METHODS
How we measure what listeners feel
People are poor witnesses to their own reactions. Our methods read the reaction itself.
OVERVIEW
Implicit association testing
By the time an opinion is spoken, the moment it describes has already passed through memory, language, and the wish to sound sensible. Implicit association testing skips the interview. It measures the speed of simple sorting decisions, in milliseconds, while a stimulus is present.
When two ideas sit together easily in someone's mind, sorting them together is fast. When they conflict, the hands hesitate before the mouth ever would. Adapted for audio, the method reveals what a sonic logo, a voice, or a soundscape actually evokes.
HOW AN IAT EXPERIMENT WORKS
04
The index
Primed response times are compared with respondents' baselines, and standardized into an index from −100 to +100. Zero indicates neutral affect (i.e., no net affect).
01
Baseline sort
Respondents sort images of different emotions as either positive or negative, while being timed. This sets a baseline that normalizes for individual cognitive speed.
02
Primed sort
The same timed sorting task is repeated under intermittent priming from the audio and/or visual assets being tested.
03
Interference
Subconscious response to the primes is irrepressible, making each sorting decision easier or harder. Differences in response time from the baseline reflect affective response.
01
Emotional Appeal
How positively or negatively listeners respond to a stimulus, scored from −100 to +100 against each respondent's own cognitive baseline. Result reflects both valence and intensity of response.
02
Strength of Attribute Association
How strongly a stimulus is linked to an intended attribute, such as energetic, innovative, or trustworthy. For UX and UI sounds, we measure this as Strength of Intended Meaning.
03
Long-Term Memory Encoding
How durably a stimulus and its associations persist in long-term memory after exposure ends. This is different from simple recall, which measures short-term or working memory.
THE METRICS
What an IAT experiment measures
An IAT study can comprise up to three separate implicit experiments, each producing a distinct measure of brand impact. The appropriate combination is selected according to the research objectives.
Most audio testing focuses on explicit assessments of the sound itself. Do people like it? Do people recall it?
Our approach measures the subconscious impact of sound on the brand. Does it make the brand easier to choose? Harder to forget? More distinctly itself?
THE WIDER TOOLKIT
One question, the right instrument
Implicit association testing comprises most of our work, but some questions need a different instrument. The toolkit divides into two families. Neurometric techniques such as EEG and fMRI record activity in the brain itself. Biometric techniques read the body's response to sense experiences through the eyes, the face, or sensors that track signals like skin conductance and heart rate. Each of the tiles below notes where the technique fits best and where it falls short.
CORE METHOD
Implicit association testing
Reaction-time measurement of association, adapted for audio.
BEST FOR Emotional appeal, attribute association, and memory encoding at market scale, with large online samples.
LIMITS Reads association strength for a whole stimulus rather than second-by-second attention within it.
BIOMETRIC
Eye tracking
Camera-based measurement of gaze position and movement, showing where visual attention lands and how long it holds.
BEST FOR How sound directs visual attention in advertising, interfaces, and on-screen or on-shelf environments.
LIMITS Attention carries no verdict. A long look can mean interest or confusion, and the method needs a visual stimulus to read.
NEUROMETRIC
EEG
Electrical activity at the scalp, read millisecond by millisecond as sound unfolds.
BEST FOR Tracing engagement moment by moment across a timeline, down to which second of a track gains or loses the listener.
LIMITS Lab-based and small-sample, sensitive to movement, and imprecise about where in the brain activity originates.
BIOMETRIC
Facial coding of emotion
Automated reading of facial muscle movements from video, classified into emotional responses frame by frame.
BEST FOR Continuous emotional valence while people watch advertising or use a product.
LIMITS Passive listening produces few expressions, and expressiveness varies across individuals and cultures.
NEUROMETRIC
fMRI
Where in the brain a response happens, mapped through blood flow.
BEST FOR Localizing which neural systems a sound engages, when that question justifies the cost.
LIMITS Expensive, slow in temporal resolution, and run inside a loud scanner that is a poor listening environment.
BIOMETRIC
Sensor-based measures
Skin conductance, heart rate, and related signals that index physiological arousal as an experience unfolds.
BEST FOR Locating arousal peaks and measuring intensity through a piece of audio or film.
LIMITS Arousal carries no valence. A spike can be delight or irritation, so sensors pair best with measures that read direction.