inVibe logo

The Value of Longitudinal Voice Tracking Research

By Beth Baldys

Thu Apr 09 2020

Emotions are, by their very definition, irrational. And yet, we’re taught to believe emotions are highly predictable: Good things make us happy. Bad things make us angry. Scary things make us afraid. Straightforward, right?

So, it would seem reasonable to hypothesize that, as the coronavirus spread throughout the U.S. between January and March — bringing with it increasing levels of illness, death and financial hardship — our speech emotion recognition tools would detect steadily-rising levels of fear in the general population.

And yet — that’s not at all what we heard.

Instead, the current crisis triggered some to panic, others to fume, and more than a few to scoff. To make matters more complicated, our analysis revealed not only a diversity of emotions across individuals, but a broad mix of emotions within the same person (see chart below).

Which begs the question: why is it that a single, shared global experience such as COVID-19 creates such a diverse spectrum of emotional responses?

How COVID Attacks Our Mental Models

As humans, we have a basic instinct to categorize. By organizing the people, objects, and ideas that exist around us, we simplify our understanding of the world so that we can more effectively and efficiently grapple with it. Our experiences are processed in similar fashion. When faced with a novel situation, people subconsciously scan their memories for the closest equivalent experience to what they are experiencing, and then react in kind.

The problem, of course, is that most of us have never experienced anything quite like the current COVID-19 crisis. This novelty throws a significant wrench into our brain’s ability to readily retrieve emotions from its well-categorized index of past experiences. Instead, our brains enter a near-continuous state of “recalculating” — shuffling through multiple past experiences and emotions as it attempts to process the ever-evolving barrage of new information.

"It's Reminding Me of a Plague from the 15th Century."

Early on in inVibe’s COVID-19 Voice Tracker research, we observed respondents actively trying to anchor the coronavirus to anything that was remotely recognizable or relatable. Early on, many people compared and contrasted COVID-19 to the seasonal flu, describing its symptoms either as “worse than” or “not as bad.” Over time, comparisons to other outbreaks began to surface (e.g., H1N1, HIV/AIDs, Ebola, etc.) yet these did not necessarily correlate with an increase in fear, as one might expect. These “reference events” were oftentimes highly dependent on the respondent’s personal association with those experiences.

inVibe’s speech emotion recognition (SER) analysis of research participants’ voice responses plots their emotional reactions on a two-dimensional Valence/Activation chart, where valence represents an individual respondent’s relative “positivity” or “negativity” toward the topic being discussed.

As mentioned earlier, emotion is not a static state. In fact, as the coronavirus has become less and less “novel,” we are able to measure an increased confidence in the emotional reactions individuals exhibit over time. Their language reveals stronger, more negative positions being taken, rather than hesitant or mixed emotions. In addition, opinions expressed previously were frequently reflected upon with mild embarrassment as minds changed.

How Longitudinal Voice Tracking Studies Can Help Organizations

Longitudinal Voice Tracking studies allow us to more precisely interpret an individual’s changing landscape of emotional reactions based on both the novelty and duration of the experience. By more precisely understanding how these emotions evolve, organizations can better tailor their communications and interventions over time.

Researchers have come to realize that emotions are primary drivers of behavior, yet there is still a limited understanding of how or why. Social scientists call this the “hot-cold empathy gap”: our tendency to underestimate the influence that a visceral, emotional state can have on a person’s decision-making. As our analysis shows, in an unfamiliar situation, our mental models lose all reliable signal.

So what should organizations be doing in these times of heightened emotion and uncertainty? Our recommendation: they should harness the power of voice and emotion to better serve their end customers. Here are some things to consider when embarking on a voice research project:

  • Don’t expect logical reactions to emotional moments
  • Keep listening as longitudinal data provides a more complete picture of your audience
  • Develop helpful mental models to guide thinking
  • Ensure your healthcare “journey maps” take into account an individual’s previous experience — not just the stage they are at in their disease
  • Enlist the assistance of listening experts

This week’s newsletter is part of inVibe’s “Voice of a Virus” series. We will continue to publish results and findings from our longitudinal voice tracker as we learn more. We are hopeful that the learnings and observations help people contextualize these unprecedented events.

If you would like to learn more or ensure you are on our mailing list please drop us a line at (949) 438–4836 or an email at, or better still — let’s Zoom, Skype, Facetime, or Hangout.

We hope you are safe, healthy, and surrounded by care and kindness.

Thanks for reading!

Be sure to subscribe to stay up to date on the latest news & research coming from the experts at inVibe Labs.

Recently Published

  • Cover Image for Imagine: Gaining Deep and Meaningful Insights from just a Few Questions

    Imagine: Gaining Deep and Meaningful Insights from just a Few Questions

    The why’s behind inVibe’s voice response collection process

    By Tripp Maloney

    Wed Apr 13 2022

  • /voice

    1. Why Voice
    2. Sociolinguistic Analysis
    3. Speech Emotion Recognition
    4. Actionable Insights
    5. Whitepapers
    6. The Patient Voice


    2022 inVibe - All Rights Reserved