Three Flavors of Emotion

Cyrus McCandless, Ph.D.

Many compelling uses of Implicit research stem from its ability to “measure emotion.” But most of us aren’t used to considering the diverse ways we use “emotion” to describe our own or others’ feelings, attitudes, or intentions. How we personally feel about a work of art, how we think the artist might have felt when they made it, what emotions we think the artist meant to convey, and how we expect others to react, are critical distinctions we don’t usually make in our everyday thinking about emotions. As a result, when we have a practical need to differentiate between these facets of emotion, things get complicated pretty quickly.

Let’s take a moment to consider exactly what we mean when we say we’re “measuring emotion.” We’ll touch on three fundamental, easy to understand aspects of emotion, while keeping in mind that these are just convenient starting points where we have a toehold in emotional measurement:

  1. Emotional response, aka my gut reaction to seeing a word or picture, or hearing, tasting, or smelling a product, seems intuitive. Do I like it, or dislike it? How much do I like or dislike it?At first, you probably feel you have a good grasp on your “gut feelings” about things. But it won’t take long to think of the last time you said “I’m not quite sure how I feel about that,” or the last time your mind (and words) disagreed with your gut. Anyone who’s been around marketing research for long will recognize these immediate gut reactions as exceedingly difficult to access and measure objectively. You’ll also recognize them as critical for understanding and improving predictions of real-world, real-time consumer behavior.The inadequacies of traditional explicit methods for measuring emotional valence and magnitude1 is at the center of massive investments in “neuromarketing” methods and technologies including Implicit, EEG, biometrics, eyetracking, facial coding, fMRI, and occasionally even more exotic tools.

    When we use explicit questionnaires, respondents often report their emotional responses inaccurately, can’t decide which option they [consciously] like more or less, consider their ratings relative to other ratings they’ve given in the same survey, struggle to evaluate mixed feelings, or respond to something other than what you really asked. You ask which gas station they like more—they respond based on feelings about which brand is more responsible for oil spills, instead of which gas station they’re more likely to visit when they need gas and have two equally-convenient choices. You ask whether they prefer peas or carrots—their conscious response has more to do with which one they’ve had most recently or most often. In a focus group setting, many respondents will find reasons why they love whatever you’re asking them about that day, just to be polite. And if you do business around the world, different cultural norms around feedback mean you can’t directly compare rating scales or verbatims from country to country. Even if you could, these explicit responses don’t do a good job of predicting what your respondents will really consider or decide to purchase when they go to the store—for that, you need a better method of capturing people’s “gut reactions.”

  2. Discrete emotional associations are a bit different. When we say that something is associated with an emotion, it’s more difficult to pin down precisely what we mean. If we see an image, and an emotion springs to mind, does that reflect our own emotional response to the image? Or does it reflect what we see as the content of the image, or how we think others will respond to the image? Is it related to our interest (or disinterest) in the image?A frowning emoticon brings the concept “sad” to mind, but it may not make us feel sad all by itself. Likewise, a riff from a relentlessly happy tune we may have heard one too many times—like Bobby McFerrin’s “Don’t Worry, Be Happy”—may evoke associations with happiness, but might make some of us feel irritated or even remind us of unhappy memories. If we ask whether “Let the Sunshine In” is “a happy song,” most people will consciously agree it is, because of the obvious and intentional association with happiness. And even though I’ve never intentionally listened to Pharrell Williams’ song “Happy,” I can’t suppress my brain’s automatic fill-in-the-blank response when it comes over the speakers of a passing car radio.These examples illustrate that happiness isn’t just an emotion that we feel personally. Emotion words can convey meaning, as words to communicate or describe the concept of an emotion whether it’s ours or someone else’s. They also have the ability to simply be associated with something in our memory, without being connected to any particular meaning (we might know the names of things in other languages, without understanding their meaning).

    We call these discrete emotional associations, and, to a point, you can think of them like any other association connected to a word or concept. If we want to understand the mind of the consumer, these are critically important pieces of information to have. They tell us what emotions are associated with an event or product, but they don’t, by themselves, necessarily tell us whether those emotions are a cause or consequence of perception, behavior or decision-making, and whether the named emotions are personally felt or just casually recognized as being related to the event or item in question. To understand whether a song that brings to mind the word or idea of “happiness” also makes someone feel happy (or irritated, or just bored), we also need to understand whether their gut-level emotional response to the song is positive or negative (valence), and mild or strong (magnitude). And if we want to use this knowledge of emotions to explain or predict behavior, we also need data on that behavior, and such as behavioral choice data or past purchase data, and appropriate methods and frameworks for integrating these pieces of information in valid ways.

  3. Emotional spectra: Removing ambiguity and adding precision. Sometimes it’s most useful to know not just how positive or negative someone’s response is, or which discrete emotions they associate with some stimulus, but where their response is on a carefully chosen spectrum of emotional response. That is, how strongly they associate an event or thing with one emotional concept rather than a different emotional concept (which may be the opposite of the first concept, or just an alternative to it).For example, we can understand someone’s response to a stressful situation like public speaking by understanding whether they evaluate the situation (consciously or unconsciously) as a threat or a challenge. If we know which emotional perspective someone took when confronted with this “stressor,” we can understand and explain why they responded the way they did—not just in terms of their outward behavioral or subjective emotional response, but also their physiological response.Not only that, by understanding where on the threat-challenge ‘spectrum’ someone’s response to an event lies, we can understand how to help someone change their response in a practical and useful way: much of the research around threats and challenges focuses on how to improve performance in public speaking by re-conceptualizing anxiety as excitement. When successful, this can not only improve performance, but actually change our physiological response, allowing us to become more calm, focused, and fluid in the moment.

    The threat-vs.-challenge framework for understanding stress is part of a larger theory of “cognitive appraisal”—in this case the appraisal of impending stressful events as either threats or challenges—that traces back to 20th century psychologists Arnold2 and Lazarus3. Cognitive appraisal theory is one of many tools that’s allowed us to make useful predictions and advanced our ability to explain human behaviors. And it neatly illustrates how considering data on emotional responses in terms of an appropriate spectrum can give us more useful and actionable insight into our emotional responses to things and events. Are you anxious and fearful, or excited and proud about your upcoming speaking engagement? Where is your consumer on a spectrum from satisfied to angry with your customer service? Is your diner excited or disgusted by the fresh octopus on the menu? Does a parent feel contented or guilty about the foods they buy for their child? Are voters responding out of happiness or sadness about their choices, or is it more useful to think of their response to candidates in terms of boredom vs. contempt?

    We still have a lot to learn about how emotion is generated, how and where in the brain various emotions arise, and when and how emotions are driven by or available to conscious or unconscious processes. But we’ve come a long way in our ability to measure emotions and understand how they affect behavior—and importantly, what to do about them. Having effective tools for measuring emotion, a good grasp of how to think about what you’re really measuring (e.g. whether emotions are personally felt, communicated by others, or merely associated with something), and reliable data on which emotions are and are not associated with events or products and our choices and behaviors toward them, can put you on an exceptionally productive path to understanding not just your consumer, but also yourself.

1Valence refers to the positivity or negativity of a response, for example happiness vs. sadness, or excitement vs. fear. Magnitude refers to degree of emotional arousal, or the size or intensity of emotional response. Arousal is often used as a measure of engagement, but neither arousal nor engagement should be presented as an indicator of the positivity or negativity of an emotional response, since arousal can be positive or negative in valence.

2Arnold, M. B. & Gasson, S. J. (1954). Feelings and emotions as dynamic factors in
personality integration. In M. B. Arnold (Ed.), The human person: An approach to an integral theory of personality (pp. 294-313). New York: Ronald Press.

3Lazarus, R. S. & Launier, R. (1978). Stress-related transactions between person and environment. In L. A. Pervin & M. Lewis (Eds.), Perspectives in interactional psychology (pp. 287-327). New York: Plenum Press.

1 Comment

  • Paul Marsden says:

    Useful. Thank you. Is three not simply a practical way to measure emotions, rather than a different way to conceptualize emotions? You create an emotional scale – using discrete (or dimensional) opposites and measure where a brand or anything else lies. Useful – but is it really a different flavor of emotion?

    P.S. We are really excited you are opening in London.

Leave a Reply

Your email address will not be published. Required fields are marked *