Q. What hard evidence is there that using implicit measures will lead to better decisions and better commercial ROI?
Q. What's the science behind Implicit Reaction Speed (IRS) testing?
In a double blind study carried out by us for a major personal care brand and aimed at assessing which of several techniques could best predict the uptake of a brand extension, the post-launch results showed that consumers' implicit, but not explicit, responses predicted market uptake.
In a recent combined biometric and implicit study we conducted for American Express on the way in which customer service impacts on consumers' nonconscious emotions, the PR story that emerged had over 15 million media take-ups - a huge ROI given their objectives.
A recent study that we conducted with trained sensory panelists showed that they are better at discriminating the specific aromas contained within a liquid compound when they perform the evaluation implicitly rather than explicitly. This is one of the few scenarios in which the individual components are known with 100% accuracy, thus providing the ideal way of validating the implicit methodology.
The weight of evidence that is being amassed from both academic and commercial sectors indicates that the integration of conscious and nonconscious reactions provide substantially higher predictive power of subsequent consumer purchasing behaviour than either alone.
Importantly, where the two measures are found to be discrepant, implicit responses are often found to be the more accurate predictor of subsequent behaviour. In circumstances whereby implicit and explicit responses are aligned, each provides a point of validation for the other (in other words, we can conclude that few social influences are acting on decisions or stated attitudes in this case).
In a recent study of voting intentions, analysis that combined both explicit and implicit reactions predicted the results with significantly greater accuracy than when explicit and implicit reactions were analysed separately.
There have been numerous academic papers published on the validity of implicit responses as predictors of subsequent behaviour. In a longitudinal study published in 2012, it was shown that implicit, but not explicit, reactions were able to predict the success of songs in the American charts, see below for a quote from the author:
"When the research subjects were asked to rate the songs on a scale of one to five, their answers did not correlate with future sales of the songs. That result may be due to the complicated cognitive process involved in rating something", the author, Berns, speculates. "You have to stop and think, and your thoughts may be coloured by whatever biases you have, and how you feel about revealing your preference. [But] you really can't fake the implicit brain responses while you're listening to the song," he adds. "That taps into a raw reaction."
Published academic articles, including from the Neurosense team, show that implicit methods predict behaviour with greater accuracy than explicitly stated attitudes on topics including but not limited to:
- Food choice
- Brand choice
- Alcohol consumption behaviour
- Subsequent behaviour (e.g., speeding on a driving simulator)
- Relationship duration
- Job switch
- Adoption of environmentally friendly packaging
- Liking of an advert and brand loyalty
Q. What's the advantage of IRS testing over other neuromarketing measures or biometrics?
There has been over 30 years of academic research in cognitive psychology on the use of implicit measures, including articles from academics, practitioners, and ‘gurus’ in the field. There is also a rapidly increasing number of academic papers on the predictive validity of the approach.
It is well established that much of what drives our behaviour goes on below our conscious awareness. An important implication of this is that if people can’t access the causes and hence explain their behaviour then why should we give them a questionnaire to complete?
An alternative is to bypass the need for conscious self-awareness and go in indirectly. This is where implicit testing comes in. We can measure how someone feels by presenting them with words and images and seeing how quickly they react. The so-called implicit association test and the affective priming test are two tests that have been extensively assessed and validated scientifically. These form the basis of what we do at Neurosense.
Q. What is implicit testing and how is it applied to market research?
Neurosense has spent over 15 years comparing implicit measures against almost every available alternative technique for measuring human emotional responses (from fMRI, EEG, IAT, priming, eye-tracking, to biometric measures). What we discovered was that our implicit response time methods correlated repeatedly with the results from the direct fMRI imaging techniques (thus validating the scalable reaction time tools). This meant it was no longer necessary to continue to use expensive brain scanning equipment to address questions that could be just as easily and more efficiently answered using scalable, practical and cost effective implicit response time testing.
In terms of eye-tracking, in one study, we compared reaction times in our N-ACT test with eye movements recorded at the same time. Reaction time, but not, eye-movements, was able to discriminate between respondents such that reaction time gave a better indication of where respondents were looking than actual records of eye fixations. Although this seems to be a curious finding, it is less so when you think about what actually happens in visual attention. Normally our attention is directed exactly where we are looking – however, something can quite easily grab our attention through our peripheral vision and when it does we often turn our head towards it and our eyes fixate on it. This is something psychologists call covert attention – we have the ability to look straight ahead but focus our attention on something at the side or periphery of our visual field. This being the case, eye-fixations measured by eye-tracking devices cannot capture covert attention. Reaction times, on the other hand, are based on attention capture (whether in the centre of the visual field or in the periphery). This can explain why our N-ACT can have greater validity than eye-tracking.
Biometrics can be very informative as to the emotional state of a respondent being exposed to various products, adverts, and services. They are especially useful in PR studies when one wishes to explain to a general audience the effect a particular experience can have on a person physiologically – we all know what a rapid increase in heart rate feels like, for example.
Two problems with biometrics are that (a) they are virtually impossible to do online and (b) they tend to illustrate the amount of emotional arousal someone is experiencing but not what emotion they are feeling. For example, both fear and excitement can yield similar physiological responses but they are qualitatively different – most of us (horror movie fans aside) would rather feel excited than afraid!
Q. What differentiates Neurosense's methodology from others in terms of the science and the outputs?
It has been suggested that our emotions and the history of our personal experiences (which lie somewhere in the nonconsicous regions of our minds) tend to drive our decisions. We all know that we have these feelings – some people call them gut feelings, others call them intuitions, impulses, suspicions or something we just know. These feelings are very difficult access, all we know is that we like some things and dislike others. How we have come to like some things and why we like them are embedded too deeply in our nonconsicous. We find it very difficult to put those feelings into words.
Accessing these feelings requires specifically designed neuro-scientific tools and over the last 15 years Neurosense has developed the largest range of proprietary tools in the world to help clients access consumers' non-conscious emotions and feelings. These tools help guide, amongst other areas, new product development, brand positioning, marketing and communication strategies, advertising, and creative campaigns.
Implicit response time tests are being used by many world leading brands to gain greater predictive insight into a very broad range of questions, including:
- Which packaging design is most effective at signalling product benefits?
- What are the components of a pack design that are crucial for identifying and recognizing the brand on shelf, amongst clutter?
- What is the relative effectiveness of different media platforms for communicating specific messages?
- How is my brand being perceived against competitors and how is it being perceived periodically throughout the year?
- Which of several versions of copy will work best for my brand?
- Which version of ad is going to work best for my brand?
- Who is the most effective and plausible endorser for my product?
- Which logo design do my customers prefer and which do prospects prefer?
- What is my brand’s archetype?
Q. What do people mean when they refer to System 1 and System 2 processes?
The tests we have developed have been specifically designed for commercial market research. Our tests provide huge flexibility in terms of the questions that can be addressed, the platforms they can be taken on, and whether tests are run online, at a central location, or in people’s homes.
Our flagship BrainLink™ tool is based instead on the paradigm of affective priming, which has a very robust scientific basis, with validation in FMRI and 100's of publications. Affective priming is the most compatible implicit test with our knowledge of how the mind works. It is also compatible with theories of how the brain stores and processes knowledge and emotion.
Neurosense has a database of hundreds of individual reaction times collected over several thousand tests and use this to develop the attributes which benchmark a sector.
Neurosense tests are TRULY implicit – many companies say they offer true implicit but actually offer what we call fast explicit. What they do is ask respondents to make speeded evaluations – say which brand they prefer very quickly, or say which of two brands the more strongly associate with, say, modern. That kind of test is not true implicit. The reason why it isn’t true implicit is that a respondent could lie, fake, guess, or choose randomly – they can do this because there is no right or wrong answer. True implicit tests work differently – respondents merely detect the presence or absence of a target (e.g., brand logo) which has a right or wrong answer. By cleverly flashing distracting words and images on the screen, we can see how these affect target detection in terms of reaction time. When we do this we can know how someone feels about the target (which could be a brand, a pack, a brand claim, and so on). Moreover, we can characterise who they feel towards it on an array of word or picture attributes.
The result is a clear picture of how consumers feel about the brand, product or service, and we have not asked them a single question, nor have we asked them to evaluate anything!
Q. If we used BrainLink to test a new design against the current pack, would the test tell us which one is preferred?
We have to remember that most contemporary psychologists are cognitive psychologists who believe that humans are processors of information. What does this mean? It means that things we see, touch, taste, smell, including internal physiological feelings (e.g., butterflies in the stomach) are all pieces of information about something going on in the world (or inside us). All the time we are bombarded with such sensory information, yet most of the time we appear to deal with (or process) this information without much effort. However, there are times when we try to focus on one thing (one source of information, such as what our friend is saying) and ignore everything else (all other noises, sounds, speech, as well as things we can see, touch and taste), for example when we are at a party trying to hear what our friend is saying amidst everything else going on at the same time.
The ease at which we can process information, depends upon how well it is learned. For example, young infants are new to all sorts of skills that they will later be able to do with ease, such as standing up without falling, walking, running, jumping, reading, speaking, reasoning, arguing effectively, making moral judgements, knowing what they like and don’t like and so on. When something is new, we need to focus all of our attention and effort on it. Similarly, when we make decisions that are new to us (or ones we don’t make very often), such as which house to buy, which holiday to go on, and so on, we need to reflect, evaluate, ponder, and examine the pros and cons. However, when we are faced with decisions we make all the time, such as whether we like a face or not, and tasks we do effortlessly like driving a car, reading text, listening to music, listening and understanding what someone is saying on the radio, making sense of what we are looking at when thumbing through a magazine, we can do this effortlessly. It’s easy for nearly all adults to do these things, without even thinking about them.
The easy effortless things we do we tend to refer to as processes involving System 1 and the things we do that need a lot of thought and deliberation are said to involve system 2.
Now, there are numerous assumptions that are made about the two systems, and these are shown in the table below
||Reflective, analytic processing
|Done by all animals
|Can do several in parallel
||Can only do one at a time
|Dependent upon experience
||Dependent upon intellect
|Type I reasoning
||Type II reasoning
|Everyday words to describe System 1
||Everyday words to describe System 2
|Feeling in one’s bones
The above is not even an exhaustive list!
What is important for consumer research is the fact that judgements about a product, brand, or service, can be made independently by both systems. So, we might love the look of that expensive new car in the showroom, but on careful reflection, we may decide that we can’t afford it. There are many more examples, such as finding somewhere to live – often we can have a very strong feeling for one property even though we know that on paper, another property would be a more sensible purchase.
The fact is that we make instant, automatic decisions about things every day of our lives. We do this even if we are not voluntarily trying to do it, an evaluation just happens! Consumer researchers have identified that some purchasing decisions can be based more these automatic evaluations coming from System 1 than the slower, more elaborated evaluations of System 2. If a market researcher asks shoppers which brand they prefer and why, they will provide a System 2 answer. But clearly, that is not the whole picture, and may even be completely incorrect, because System 1’s major influence in a decision cannot be so easily verbalised. This is unfortunate, especially when many consumer psychologists have established that many, and possibly even the majority, of consumer decisions are made through System 1 and not through System 2.
A general rule would be: Asking a question is only good if people know the answer (when the System 1 response is the same as the System 2 response).
Fortunately, implicit technology is designed to read directly from System 1 rather than from System 2.
Q. Is the BrainLink test based on just word attributes or can you use images?
BrainLink does not only give a single answer about which pack is preferred (an overall engagement score) but also why it is preferred. If we had say 40 attributes, there may be 30 reasons why the new pack is preferred and 10 reasons why the current pack is preferred over the new design. The point is that we can characterise each pack against the set of attributes, saying which pack is most strongly associated with each attribute and by how much. The results will also provide the basis for creative to improve on the design, if that is required.
Q. Isn’t BrainLink sensitive to handedness and age?
Our test can use words or visuals or combinations of both. We usually use visuals as the target to look out for and words as attributes/primes. However, if you were testing a larger number of products, you might want to use them as picture primes and attribute dimensions as targets.
Q. How do you choose the attributes for a BrainLink test?
Yes! Reaction times are dependent on handedness (usually right handed people can react faster with their right hand than with their left hand, and vice versa for left-handed people) and in general the older we get, the slower we respond. Our analysis algorithms run through the data to assess handedness, overall speed of responses, and error rates (among other factors), and we apply transformation functions and correction factors to control for these. The result is that our findings aren’t biased by handedness, age, error rates, and similar.
Q. Which languages can BrainLink be tested with and are you able to test people who don’t have an internet connection?
In most studies, the primes or attributes are words or phrases. Our selection of attributes is based on (a) the type of product being tested (e.g., pack, claim, brand logo, etc), (b) the market segment (c) a generic model of brand equity we use, and (d) our extensive database which has thousands of attributes and a measure of how discriminating they are. We also allow clients to suggest attributes they wish to test. We can advise whether a particular attribute is likely to yield interesting results.
Q. What evidence is there that implicit reaction time tests can predict anything that explicit methods can’t?
Neurosense has run tests all over the world and in numerous languages, including Spanish, French, Italian, Portuguese, Mandarin Chinese, Thai, Hebrew, German, Afrikaans, Dutch, Swedish, Korean, Danish, Vietnamese, Bahsa, Japanese, Nigerian, Malay, and Isizulu. We can test respondents online, at a central location, or via a home visit. Neurosense has desktop versions that can be installed on a PC or a laptop so that no internet connection is required to take the test. At the end of a session of testing, the interviewer can upload the data online when they get back to the office.
Q. How fast is an implicit response?
In terms of the predictive validity of implicit versus explicit methods, there are numerous examples in the peer-reviewed literature demonstrating that in many circumstances (including in the case of consumer choice) implicit attitudes are better predictors of subsequent behaviour than explicit responses provided at the same time. For example, Steinman and Karpinski (2008) found that implicit but not explicit attitudes towards the brand GAP predicted GAP patronage and buying intentions. Brunel, Tietje and Greenwald (2004) showed that implicit methods can detect attitudes about brands that explicit measures cannot (e.g., the effects of the same or different race of the advocates of a brand).
Other research highlights the subtleties of attitudes that implicit methods are able to detect that are missed by explicit methods. For example, Priluck and Till (2009) found that explicit and implicit measures were both good at detecting attitudinal differences between brands when the difference was large or obvious, but only implicit methods could detect differences when they are less obvious. Other research shows that implicit methods in a consumer context are difficult to fake - Chan and Sengupta (2010) found that while the claims of an advertisement were dismissed, implicit responses revealed favourable attitudes to the brand.
A very convincing study of the benefits of assessing implicit attitudes is from Vianello et al. (2010) in which college students were given explicit and implicit measures of conscientiousness. Half were further told to imagine that they were being tested for their ideal job (good income, low effort, and so on) and half were not. Those with the job-story had higher explicit measures of conscientiousness (showing that they could give biased answers), but both groups scored about the same on the implicit measure – which shows that the implicit measure was not easy to fake.
Of course, most of our evidence comes from our own consultancy and these have not been published. However, if check out our Case Studies, there are examples of how well implicit measures were able to make very useful and accurate predictions.
People often ask me how fast a response needs to be, to be considered implicit. We always say that implicit processes are fast, but just how fast are they?
But first, let’s ask how fast can people react anyway?
As a teenager, when I got my first digital watch, I used to play with the stopwatch function. You press the same button to start the timer as you do to stop the timer. By pressing the button twice and as fast as possible, you can get a measure of how quickly you can tap your finger. A good score for me was around 0.13 seconds (130 ms). On rare occasions I could get 0.1 (100 ms) and my personal best is 0.09 (90 ms). So tapping a button as fast as you can (this is not even reacting to anything) can take about 100 ms. In terms of reacting to something, the time it takes to detect whether a target is a word or a nonword (e.g., scrambled letters that don’t make a word) is around 480 ms (Dijksterhuis et al., 1998). So the quickest responses we might expect from any reaction time test might be somewhere between 100 and 500 ms. Turning to implicit responses, we might conclude that fast is 500ms or less and slow (and hence not implicit) is anything else above 500 ms. But this is still just a guess, I haven’t proved anything. In fact, I have spent some time searching academic article databases for an answer to the question ‘How fast is implicit?’ and have yet to read anything that gives me a figure.
But are we even asking the right question?
Some so-called implicit tests make an assumption about this figure, or even avoid the question altogether.
Let’s consider an example from one kind of test in which respondents are asked to make a speeded evaluative response. The question might be ‘Which product do you prefer this one or that one?’ or ‘This brand is high quality” press the E key if you agree and the I key if you disagree.
The kind of test I have just been describing is a speeded evaluative decision test – henceforth referred to as a Fast Explicit test. The full test goes something like this: it is completed over a number of trials, e.g., 40. Two brands appear on the screen, one to the left and one to the right. A word (some brand attribute, such as ‘Innovative’) appears midway between the two brands and the respondent has to indicate (by way of a key press) which brand the word most applies to in their view. Respondents are encouraged to answer quickly and not think about it too much. So, respondents decide as quickly as they can between the two brands for around 40 word attributes. The computer records the decision and how long it took for the response, measured in milliseconds.
In analysing the results one method is to do a median split on the responses, having removed outliers [a median split involves finding the mid-point of all reaction times and calling anything above this, ‘slow’ and anything below this, ‘fast’] . By doing this, the reaction times are divided into two groups, those below the median reaction time and those above. The median reaction time might be around 750 ms or it might be 2000 ms, depending upon how slow or fast the sample as a whole responded. The data are analysed in two ways – one analysis for responses less than the median value [the ‘fast’ responses] and another one for responses greater than the median value [the ‘slow’ responses]. In fact, given that the median can vary between studies, the terms fast and slow are misnomers, and should instead be ‘faster responses for this sample’ and ‘slower responses for this sample’ – they are not absolute values. The median split method is really a ‘cop-out’ and doesn’t address the ‘how fast is implicit?’ question. To the novel eye, this method appears to provide both implicit and explicit data. But it doesn’t and I will tell you why.
The first reason is that there is no objective measure that tells you the point at which an implicit reaction time becomes explicit. You just have to invent one, just make one up, or, in the above example, let the data decide! Without truly knowing the cut-off point - the figure that indicates the maximum reaction time for an implicit response to be ‘truly’ implicit - you are just making it up. It isn’t scientific at all, it’s purely subjective.
There is an even bigger problem for this fast explicit method, which is that there is no way of knowing whether a respondent on any given trial is trying to respond in the way they have been asked to. There are a number of strategies they could be using to get through the test quickly (such as complete random responses, alternate left brand-right brand responses, or some other combination too difficult to detect by a computer algorithm).
Another problem arises from the fact that most of us are ‘cognitive misers’ – if there is a short cut to a problem then we will take it, rather than the alternative of giving the problem an extensive and tiresome amount of thought. One short cut is to decide that you prefer Brand A over Brand B and so all you need to make sure you select Brand A most of the time. So, overall preference for one brand can lead to a generalised response for that brand.
So where do we go from here? And what is a true implicit test?
True implicit tests are those where there is no evaluative response, there is just a correct answer on each trial – press one key if you see this brand or this kind of word and another key if you see a different brand or a different kind of word. Only one response is correct and you can only proceed if you get it right. The best strategy through this test is to give the correct response on every trial. You might think then that the best strategy would be to press both keys at the same time or both keys as fast as you can. In our tests we detect this kind of key pressing and give warnings or exclude them from our survey. In a true implicit test one compares reaction times under a ‘congruent’ condition with those in an ‘incongruent’ condition, there is no such thing in a fast explicit test.
In a true implicit test, respondents are looking out for ‘targets’, usually two brands. When they see brand A they press one key and when they see brand B they press another key. By flashing word attributes (such as Quality, Trust, and so on) on the screen before the target, we influence how quickly they are detecting the two brand targets. We can calculate by how much a word might speed up or slow down detection of either target, and hence infer the respondent’s attitudes and feelings to the two brands separately. Now, this is a far more impressive technique than the simple reaction time test of the fast explicit technique. The true implicit test can measure attitudes and feelings without respondents even knowing they are being measured, and without even asking them anything. How do we know this? Because people make spontaneous evaluations very rapidly all of the time. Everything we see, touch, taste, and hear, we very rapidly determine how we feel about it, even without any conscious intention to do so. We can’t help ourselves. Implicit reaction time tests pick this up and in the way I just described.
So, we see that the question ‘How fast is implicit?’ is really the wrong question to ask in the first place, because if you are asking that question then you are actually thinking about a fast explicit test not a true implicit test.
Doctor Implicit, 28th July 2015