neuromarketing.com.au

AOI 6 · Putting it together

Reading a neuromarketing claim: eight questions

A neuromarketing claim is only as strong as the study behind it: what was measured, in whom, how well it predicted anything, and whether people agreed to take part. The eight questions below come from the research and the Australian rules described in the other five guides, so each one links back to its source.

General informationThese questions help you read a claim; they are not legal advice. For advertising claims, the official place to check is the ACCC’s guidance on false or misleading claims.

An invented example, for illustration only

Our EEG study1 proves8 that shoppers’ brains2 love3 the new pack.

Not a real claim or a real study. The small numbers point to the questions below that each phrase raises.

The eight questions

  1. What was measured, and with what?

    Each instrument records something different. fMRI can show where activity is but works in steps of seconds; EEG catches changes in milliseconds but sees little of the deep brain. Eye tracking records where people looked. A claim should name the method; see what neuromarketing is.

  2. How many people, and who were they?

    Two of the studies described in this guide were small: 26 adults aged 18 to 26 in one purchase study, and 30 students in one eye-tracking study. Ariely and Berns listed generalising from a small group to a large population among the field’s ethical issues.

  3. Is one brain region carrying the whole claim?

    Reading “liking” from activity in a single region is reverse inference, and one region takes part in many processes. On Ariely and Berns’s reading of the evidence, there is no single “buy button”.

  4. How much better than chance, and than simply asking?

    In the 2007 purchase study, brain measures predicted buying about 60% of the time, against 50% for chance, and added to what people’s own ratings predicted. Ask for the figure, and for what it is being compared with.

  5. Do other studies agree?

    A 2022 review of 174 EEG studies found some measures more consistent than others, but limited consistency overall, with mixed results for each measure. One study is a starting point, not a settled fact.

  6. What were people shown, and could someone repeat it?

    A 2026 review warned that when the images or products shown to people are unique, poorly documented or out of reach, this makes it hard for other researchers to repeat the study. Ask whether the materials are described well enough to test again.

  7. For eye tracking: is the method reported, and does looking mean understanding?

    A careful report gives the tracker, sampling frequency, calibration, uncertainty and data loss. And attention is not comprehension: in one experiment, happy faces during a spoken health warning reduced how well viewers understood it. See eye tracking and attention.

  8. Were people asked properly, and can the claim be proved?

    The National Statement’s guiding principle is voluntary participation based on enough information and understanding. Under the Privacy Act, collecting sensitive information, such as biometric information used for automated verification or identification, generally needs consent. And in advertising, a business must be able to prove a claim that its product has a particular quality or benefit. See consent in research, biometric data and privacy and persuasion and the law.

The invented claim, read with the questions

“EEG study” (question 1): which EEG measure, and why that one? A 2022 review found frontal alpha asymmetry and the late positive potential the most consistent markers of preference, but even those gave mixed results.

“Shoppers’ brains” (question 2): how many shoppers, and who? A small group tells you about that group.

“Love” (question 3): this is the reverse-inference step, reading an emotion from a signal. Ask what the signal was, and what else it could mean.

“Proves” (question 8): if this sentence appears in an advertisement, it is a claim about the product’s qualities, and the ACCC expects a business to be prepared to prove its claims.