Anna Janska is a Ph.D. candidate in cognitive psychology and a research consultant at Point-Blank International Marketing Research and Consultancy, Berlin.
As qualitative researchers, we repeatedly have to fend for our space. Once in a while we find ourselves in a conversation with someone who does not have a clear understanding of qual and who will question us as to how we can ensure that our sample is representative, that our data are valid, since the procedure is not 100-percent identical between groups/interviews/ethnos, if our data are replicable, if we do any statistical testing, etc., etc. In short: how qual can adhere to the standards of quant. By now, we have our answer ready: Qual has its own rules and standards, based on thoroughly understanding what our interlocutors think and what makes them tick.
These days, this very notion is being threatened by neuromarketing.
Neuromarketing argues that we cannot understand what people think by just asking them, as participants cannot articulate why they make decisions or prefer one product over another. Instead, these processes happen on a hidden, unconscious layer. So neuromarketing in a way circumvents the participants as informants and asks the origin, their brains, directly.
Neuromarketing is still a young field and accounts of its limits and usages still differ significantly. The largest common denominator is: Neuromarketing is the application of neuro-imaging techniques to marketing research questions. It is a spin-off of neuroscience, which itself is a multi-disciplinary field uniting scientists from biology, psychology, medicine, physics, chemistry, mathematics, linguistics and informatics. What they have in common is their interest in some characteristic or function of the brain.
As old as humankind
Our fascination with our brains is probably as old as humankind. Earliest evidence for surgery on living humans’ skulls dates back to the Mesolithic Age: In the Dnieper Rapids region in the Ukraine, skulls between 8,000 to 9,000 years old have been excavated that had surgically-created holes in them, probably by drilling in the skull. Remarkably, these patients survived surgery long enough for the bone to start healing around the holes!
In ancient Greece, Hippocrates thought that all emotional, sensory and cognitive functions are rooted in the brain. Aristotle, however, thought that the main function of the brain was to cool the blood. The blood, he thought, was in turn warmed by the heart, inside of which the soul lay and where all sensations were located.
Fast-forward to the Renaissance: The invention of hydraulic machines lent itself to the belief that the human body functioned in the same manner. René Descartes introduced mind-body dualism: The body is a physical entity working like a machine, while the mind is a separate, non-physical entity which, as a consequence, cannot be understood by studying the human body. It was the mind, he reasoned, that distinguished humans from animals. Neuroscience today in fact assumes a monoism, rather than a dualism, postulating that physical investigation can give us insights into the mind; the mind and the brain are one, since “the mind is what the brain does,” as the psychologist Stephen Pinker puts it. Thus conversely, by understanding the brain, we should be able to infer what is going on in the mind.
By the end of the nineteenth century the brain had been dissected extensively. It was observed that all human brains had comparable structures, which gave rise to the suspicion that each bit of brain was home to a specific function.
Localization of brain functions made good progress by observing how behavior is affected when a part of the brain is damaged. One such famous case is that of Phineas Gage, a 25-year-old railway worker who suffered a tragic brain injury in Vermont in 1848. In an explosion, a metal spike shot through his skull and brain. Miraculously, he survived and was physically and mentally fully functional. His personality, however, had been changed dramatically. Once a responsible, balanced, earnest young man, Gage became impulsive and irresponsible and did not adhere to social norms, so that his friends felt that he was no longer himself. This accident provided evidence that our personality sits in our brains.
Brain localization became more concrete through Paul Broca. His famous patient, Monsieur Leborgne, whom he met in 1861, could only say the syllable “tan” for several years but could vary the intonation with which he did so. Broca understood this case as a window to understanding the localization of speech. An autopsy revealed damage to Leborgne’s left frontal lobe. Over the next years Broca documented several similar cases with comparable symptoms and damage to the left side of the brain. This marked a turning point in localizing the function of speech in the brain (and led to Broca having a region of the frontal lobe named after him).
Shortly after, Carl Wernicke observed that damages to other parts of the brain could also affect language, however, it did not affect speech production. Broca’s area, he argued, is responsible for the physical movements necessary to produce speech. Other areas involved in language are areas of understanding speech and selecting the right words. Patients with impairments in these areas might speak fluently, yet the words they string together do not make any sense.
These observations show that specific functions, such as language, do not “live” in one specific region of the brain. Instead, they rely on the workings of several connected areas. In fact, many brain regions even have more than one function.
Although we have just breezed through 8,000 years of human interest in the brain, neuroscience is a fairly young science that has only emerged as a distinct discipline in the late 1950s-early 1960s. It is not a coincidence that the last 50 years have seen a surge in interest and discoveries in the brain: Revolutionary discoveries about the brain’s structure and function have been made possible by the inventions of new methods: EEG, MEG, PET, fMRI and TMS. Up until that point, neuroscientists were dependent on accidents to localize human brain function. Sophie Scott, neuroscience professor at University College London, argues that if we map our quest towards knowing everything there is to know about the human brain onto a 24-hour day, we are only at the very beginning, maybe two minutes into that day. As one of neuroscience’s newer daughter disciplines, neuromarketing is probably only a second into that day.
Mixed bag of methods
The methods neuromarketing employs are as heterogeneous as the discipline itself. Neuromarketing for Dummies (you know a topic has entered mainstream when there is a for Dummies book on it) names a mixed bag of methods, such as EEG, fMRI, eye-tracking, skin conductance, heart rate measurements and response times. These physical response measurements vary in the immediacy by which they monitor changes in mental states.
The method which probably has captured the imagination of the general public most is functional magnetic resonance imaging (fMRI), as it generates images that allow us to see brain activation. In fact, David McCabe and Alan Castel found that people rated articles illustrated with fMRI scans as more credible in their scientific reasoning than the very same articles accompanied by topographical maps, bar charts, news pictures or no pictures at all.
FMRI uses the electromagnetic properties of blood to depict brain function: the protons inside the atoms in our body spin (i.e., they rotate around their own axis). You can picture this movement like the Earth turning around its axis. This spinning motion creates a tiny magnetic field, which MRI scanners can pick up. Oxygenated blood has different magnetic properties than de-oxygenated blood, so the two can easily be distinguished. BOLD (blood oxygen level-dependent) fMRI or BOLD contrast relies on the assumption that activated brain areas need more oxygen and so oxygenated blood rushes into them. By identifying areas that are supplied with lots of oxygenated blood, areas of increased activation are identified.
You probably have come across images of such activation patterns, where colorful blobs are superimposed on images of black-and-white brains. The areas that are highlighted are areas that are more active in one condition compared to the control condition (or the baseline, i.e., when a participant lies in the scanner and does not perform the task). In other words, the blobs you see in an fMRI picture are not the only areas that active in your brain; they are rather areas in the brain that are significantly more active in one experimental condition than in another (thus the name BOLD contrast).
To illustrate how this method can be applied to questions of consumer behavior, let us look at a study Plassmann and colleagues published in 2008 in the Proceedings of the National Academy of Sciences. They used fMRI to investigate how the price of a bottle of wine influences the level of enjoyment a participant derives from tasting it. Experimental participants were told that they would be sampling five different wines. Participants lay in an MRT scanner, which is a large metal tube, with their head fixated, as it is imperative for fMRI image generation that all movement is avoided. The price of the wine was shown as a projection inside the participant’s visual field and then a little bit of wine was squirted into a participant’s mouth through a straw. In addition to the fMRI measurements, participants were asked to indicate by button presses how much they enjoyed each wine they tasted. In fact, participants tasted only three wines, not five, as announced. Two of the three wines were tasted twice – once labeled as a cheap wine and once as a more expensive wine.
As we would probably expect from our own research practice, participants rated the enjoyment of the wine significantly higher if the price tag on the tasted wine was higher. Neuroscience can shed some light on why that is, even when the participants themselves are unable to tell. Enjoyment correlates with activation increases in the medial orbitofrontal cortex (mOFC), an area associated with the experience of pleasure. Areas immediately responsible for taste perception, however, did not show any difference like that. So the expectation of enjoyment effects the actual experience of enjoyment but not taste perception itself. Thus, neuroscience offers us an answer to why that qual cannot attain.
Herald the death of qual?
Does neuromarketing herald the death of qual? Hardly. As with quant and qual, there are two different standards for good research at work, each of which has its own affordances.
The more precise a method is and the smaller the distinctions it makes are, the more control over the experimental parameters are needed to ensure that effects found are not just a fluke. Physiological measurements, as used in neuromarketing, are small changes that do not speak for themselves, but rely on a number of processing steps and statistical analyses. Just conducting an fMRI experiment to “see which areas light up” is highly problematic. If a large number of comparisons are made, chances of seeing a difference increase, just because of the sheer quantity of comparisons. This can lead one to falsely assume that two conditions are different when they are not. A typical approach to compensate for that risk is to acknowledge it by being stricter when evaluating differences between conditions. So for every additional investigated contrast, the difference between conditions needs to be a bit more pronounced to be considered a real effect.
This issue is demonstrated by an experiment published in the Journal of Serendipitous and Unexpected Results. Craig Bennett and his colleagues conducted the experiment to caution against the dangers of negligent use of statistical evaluation. The experimental setup and reporting procedure of the fMRI experiment were according to standard protocol, while the statistical evaluation was somewhat dubious, though not unprecedented. A large number of statistical comparisons were run without adjusting for it.
It is the delicate detail of their choice of experimental participant that emphasized the absurdity of conclusions drawn from such analyses: A dead fish. The researchers put a dead salmon in an fMRI scanner and told it to guess what emotions humans on projected pictures were experiencing. Their results showed that some brain regions in the dead salmon were more active during this task than when the fish just lay in the scanner and relaxed. Obviously, the scientists did not discover cognition in a dead fish but revealed fishy statistics, instead.
Thus, it makes sense to only conduct a few hypothesis-guided comparisons. A hypothesis is an educated guess, such as: More-expensive wine is enjoyed more than cheaper wine. An experiment is then conducted to see if the gathered data support or falsify the hypotheses. To do this, the experimenter needs to define how “is enjoyed more” would manifest itself in fMRI data. In the wine example, this was activation changes in the mOFC. Activation of this area is compared across experimental conditions. The hypotheses are then answered as a series of yes/no questions: Is Area X in Condition Y more active than in Condition Z? Or does activation in Area X correlate with Parameter Y (e.g., price)?
One additional safety net to avoid spurious results is to gather additional behavioral responses to help with the interpretation of neuroscientific data, such as the enjoyment ratings used in the wine study. It would be an oversimplification to assume that only by observing activation in a single area can conclusions about a participant’s mental state be drawn. While complex functions like language are not constrained to a single area (but are rather the result of the collaboration of several areas), one area often has more than a single function, so that its involvement cannot tell us anything definitive about a participant’s mental state.
Author Martin Lindstrom told us a few years ago that we “loved” our iPhones – “literally.” In an fMRI study participants were shown videos of iPhones or played the ring tone. Activation in the insular cortex was found, which is “associated with feelings of love and compassion. The subjects’ brains responded to the sound of their phones as they would respond to the presence or proximity of a girlfriend, boyfriend or family member.” However the insular cortex has more than one function. The insula does lots of things, only one of which is experiencing feelings of love. Other emotions it is involved in are happiness, fear and disgust. As such, one could just as well infer that you are disgusted by your iPhone.
Qual is not subject to equally strict constraints, as it does not record miniscule changes in physical responses. Instead, it is storytelling that qual excels at. It does not need to eliminate as much variation between participants as possible but instead revels in it. Qual can look at the bigger picture and create the story as a meaning-infused entity.
Let’s return to the wine-tasting example. Qual fails to disentangle expectation of enjoyment and taste, so indeed, it cannot answer that question of why a particular wine is enjoyed. Yet, qual still has the card of con-textualization up its sleeve: It can investigate how a wine is enjoyed. It can meet participants in their natural habitat and understand the situations that generate rituals. What are situations that call for a bottle of wine? What are the places where a more expensive bottle of wine is bought? How is it served? What makes a special evening for the individual?
While neuromarketing attempts to get closer to the real answer by asking a participant’s body, qual approaches the consumer by entering their living situations and their realities.
The other trump of qual is its openness – it is free of the constraints imposed by standardization. Thus qual cannot only explore variation across a dimension (such as price) but set out to explore the dimensions that shape consumers’ worlds. Qual is open for serendipity and cherishes the element of surprise that is at the heart of new knowledge. A stance of openness and empathy in the researcher allows an interaction with the consumer at the eye level, which lets the consumer show the researcher their truth, rather than attempt to answer questions to specific fragments of it. The researcher can thus understand the consumer as a complete human being rather than a data point. We must not forget that this empathy is also desperately needed in order to further marketing decisions, as effective marketing must address the human consumers who use the products and services. The currency of qual is thus not so much a consumer’s perception but their (inter)actions with their surroundings and the contexts that shape them.
Both want to understand
The overlap between neuromarketing and qual is that they both want to understand the consumer. The angle, however, is a completely different one, so that they should be seen as informing, rather than cannibalizing, each other.
Neuroscientific methods can greatly help us to understand the human mind. How are decisions made? What aspects play into the decision-making process? What weight is given to individual factors of the decision? Knowledge of how the human brain, memory and decision-making work can truly help us to sidestep errors in our research design and in our interpretation of data. For that, however, there is no need for us qual researchers to conduct studies ourselves. Neuromarketing can in turn rely on findings of qual that identified relevant dimensions to explore further. There is real potential for a mutually enriching relationship here. I think this is the beginning of a beautiful friendship.