Zoom Fatigue
2021. A difficult year, at best. I'm optimistic that 2022 will be better, although most of our current challenges will still be with us on January 1.
For instance, it's clear now that omicron causes less severe COVID-19 symptoms than previous variants did, but the breakthrough rate is higher. As the new year unfolds, social distancing will continue to be prevalent, and that means, among other things, more Zoom calls. (I can hear you groaning already.)
Zoom currently hosts about 300 million meeting participants per day, a 2900% increase over pre-pandemic levels. No surprise then that this dramatic spike has given rise to a phenomenon known as "Zoom fatigue."
In this newsletter I'll talk about what Zoom fatigue is, how it's measured, what causes it, and how it can be minimized. I'll be discussing studies published this week as well as earlier in the year. In the process, I'll comment on a key obstacle to public acceptance of scientific data.
What is Zoom fatigue?
Research on Zoom fatigue is new – most of the studies appeared this year – and so researchers are still in the process of trying to simultaneously define, measure, and explain it.
In theory, we should agree on the definition first, so that we can know what it is we're trying to measure and explain, but social science doesn't work that way. (For example, after more than a century of research on constructs like intelligence, creativity, motivation, and aggression, there's still no consensus among psychologists on what exactly these terms mean.)
At the moment, most definitions of Zoom fatigue treat it as a combination of physical and mental tiredness caused by videoconference participation. The physical symptoms include a general sense of fatigue, as well eye-related discomforts. The mental symptoms also include fatigue, as well as diminished motivation and undesirable emotional states such as irritability. (Sound familiar?)
Zoom fatigue can arise from any videoconferencing tool; researchers are quick to point out that the Zoom platform doesn’t appear to be unusually problematic. The phrase "Zoom fatigue" is catching on simply because of this particular app’s popularity. ("Zooming" has already become a generic verb, in the same way that "Googling" did some years ago.)
How is Zoom fatigue measured?
A standardized measure of Zoom fatigue, the Zoom Exhaustion & Fatigue Scale (ZEFS), is currently under development at Stanford's Virtual Human Interaction Lab; the first published data and preprints came online this month.
The ZEFS is open to the public. It takes about 10 minutes to complete, and at the end you receive immediate, computer-generated feedback on how much Zoom fatigue you've been experiencing compared to others who’ve completed the survey. (If you wish, you can take the ZEFS yourself here.)
The ZEFS consists of 15 questions. Three questions were selected to represent each of five dimensions. Here are the dimensions, followed by an example question from the survey:
General fatigue: "How mentally drained do you feel after video conferencing?"
Visual fatigue: "How irritated do your eyes feel after video conferencing?"
Social fatigue: "How much do you need time by yourself after video conferencing?"
Motivational fatigue: "How much do you dread having to do things after video conferencing?"
Emotional fatigue: "How moody do you feel after video conferencing?"
The scale used for responding to these questions is: 1 = “Not at all”, 2 = “Slightly”, 3 = “Moderately”, 4 = “Very”, 5 = “Extremely”.
How valid Is the measurement of Zoom fatigue?
The ZEFS is still under development, so I will tread lightly here, but it's clearly got some limitations. Here are two that illustrate something broader about obstacles to public acceptance of scientific data.
1. Redundancy
The ZEFS contains a lot of redundant, or nearly-redundant questions. For the dimension of general fatigue, two of the questions are "How exhausted do you feel...?" and "How tired do you feel...?" For the dimension of social fatigue, two of the questions are "How much do you want to be alone...?" and "How much do you need time by yourself...?" These are just a few of many examples.
Experts in survey design have identified several problems with this type of redundancy. Some respondents will become irritated at the researchers' apparent carelessness; as a result, they'll stop taking the survey (or at stop taking the questions seriously enough to give sincere responses). Some respondents will assume that the researchers intended distinctions that they (the respondents) don't quite grasp. Thus, they'll deliberately answer two similar questions differently, even though they don't see any difference between those questions.
In short, redundancy may undermine how much the ZEFS can tell us about peoples’ actual fatigue. (See Appendix 1 for another redundancy-related problem.)
2. Vagueness
The ZEFS doesn't distinguish between personal vs. professional meetings. It doesn't distinguish between high-stakes meetings (e.g., a one-on-one with your boss) vs. those where there's less to gain or lose. It doesn't reflect your role in a meeting (e.g., making a presentation vs. conversing vs. merely listening).
Because the survey fails to make these kinds of distinctions, the best answer to a lot of the questions is: It depends. How tired do you feel after videoconferencing? It depends. How much do you want to be alone after videoconferencing? It depends. In short, the survey questions ask respondents to generalize across videoconference experiences that may be radically different from each other.
What do the ZEFS limitations tell us about public resistance to science?
This limitations of the ZEFS – particularly its vagueness – illustrate one of the reasons some people fail to accept scientific findings.
That last statement must seem like a pretty big leap. Bear with me for a few paragraphs, and the connection will be clearer.
People often struggle to reconcile scientific data with their personal experiences, or with what they’ve heard from others. You know how it goes: Science tells us that smoking is bad for you, but your two-pack-a-day Uncle Johnny lived into his 90s. Science tells us that men are more aggressive than women, but your sister used to beat the snot out of you when you were kids. There are a million examples like that, and they make it challenging sometimes to bridge the gap between science and public understanding.
Smoking is bad for you. That's beyond debate. And, people like Uncle Johnny do exist. That's not debatable either. But the hazards of smoking and the longevity of Uncle Johnny aren't incompatible. The connection between smoking and health should be framed statistically: the more you smoke, the more likely you'll experience a range of problems. Likewise, although men tend to be more aggressive than women (according to most definitions of aggression), you can find exceptions. The fact that your sister is hyper-aggressive doesn't change what we know about gender and aggression; rather, it just highlights that what we know is probabilistic. If you're a guy, and your sister used to beat you up, well...you were unlucky. It doesn't change the fact that males tend to be more aggressive than females.
(As I write this, I'm thinking of the folks who refuse to get vaccinated, or mask, or distance, because they themselves – or people they know – have refused but haven't gotten sick yet.)
We can't just blame the public for the kind of misunderstanding I'm describing here; researchers bear part of the responsibility. This brings me back to the ZEFS. There are many studies in social science and other fields grounded in measures like this. When measures are insensitive to context, or survey don't give respondents the chance to say "it depends", the results get framed as generalizations that people will reject because they're quite aware of exceptions. For instance, the creators of the ZEFS found that people who report longer videoconferences also report greater fatigue. That may be true, but it’s just not a very informative or persuasive finding without more context. Does the connection between videoconferencing and fatigue depend on whether it's a personal vs. professional meeting? If it’s a meeting, does the agenda matter? How about the number of participants? And, importantly, how many exceptions to this trend are you likely to see? If you want to convince your boss that his Zoom meetings are too long, you’ll have better luck if you describe the conditions under which fatigue is greatest – and acknowledge that not everyone gets fatigued.
In short, I’m arguing that surveys (and study findings) should be framed so that people can treat the findings in a statistical way (i.e., as probably true, or more likely under certain conditions, or whatever). I believe people are capable of this. On the other hand, I think it's counterproductive to continue to present findings in an overgeneralized way, as experts often do. I wouldn't tell my boss that excessive Zooming fatigues people, and just leave it at that; I would acknowledge that it's possible to Zoom quite a bit without fatigue. But I would present the latter as an unlikely outcome, statistically speaking, and one whose likelihood approaches zero the more you Zoom. (Much the same should be applied to how we discuss preventive behavior during the pandemic. You might not get sick if you're unvaccinated, unmasked, and undistanced. We should tell people that. But we should quickly add that it's like saying you might not get hit by a vehicle if you walk from Dallas to Houston on Interstate 45. It's a true statement - you might not get hit. But what are the chances?)
What causes Zoom fatigue? (And, what can we do about it?)
Studies published up through this week have identified at least five causes of Zoom fatigue. (See here and here.) Below I’ve described each cause, followed by simple solutions that experts have proposed.
Cause 1: Excessive eye contact
Studies suggest that eye contact is more frequent during videoconferencing than face-to-face meetings, particularly when group sizes are small. A separate line of research shows that people find it stressful to be looked at. (This is one of those generalizations that comes with exceptions. Think of the Kardashians, for instance.) In short, Zooming creates implicit pressure to maintain eye contact, thereby slightly increasing the stressfulness of online meetings, at least for some people. Experts also note that the problem is exacerbated when the size of someone's face on your screen is relatively large, as if the person were a lot closer to you than they actually are.
Fix 1: Look away
That is, when videoconferencing, remind yourself to look away from your screen periodically, if it's possible to do so without seeming rude or causing distraction. Reduce the size of your Zoom window, or use an external keyboard, to reduce the size of the faces you see (and, perhaps, the extent to which a quick break in eye contact is discernible).
Cause 2: Mirror effects
Studies show that when people spend extended time in front of a mirror, they become more prone to self-doubt and self-criticism. (Again, this is a generalization rather than universally true.) Zoom researchers propose that the self-view functionality of Zoom and other videoconferencing platforms contributes to "fatigue" through an ongoing mirroring effect (and, as you can imagine, some people are more adversely affected by self-view than others).
Fix 2: Hide the self-view function
That is, at the outset of a videoconference, hide the self-view function after ensuring that your face is properly positioned.
Cause 3: Reduced mobility
Researchers have pointed out that compared to face-to-face and phone interactions, Zooming requires you to spend more time planted in one spot, in order to stay within the confines of the screen. Doing so can be stressful and tiring for some people.
Fix 3: Use an external keyboard and/or camera
By using a separate keyboard and/or web cam, the extra distance between you and your screen will allow you a bit more freedom to move around while remaining in view of other Zoomers. (If possible, you can turn your camera off for a few seconds every now and then, and get up to stretch. Or do what I do: Move around while you talk, and hope the other person understands.)
Cause 4: Greater cognitive load
Face-to-face meetings already require a lot of attention, problem-solving, and other cognitive resources. Zooming adds to the cognitive load. For example, you need to remember to keep your head positioned in the screen. You need to make sure people hear you. You need to make sure they spot important nonverbal cues, like nodding. You need to compensate for technical glitches (when you're not worrying that they might occur). These and other mini-requirements add up…
Fix 4: Use fixes 1-3
If possible, turn off your screen periodically while you listen. Look away from the screen from time to time. Accept that technological/user glitches are inevitable. If you're going to make a presentation that requires special functions (screen sharing, division of the audience into groups, etc.) persuade friends to help you practice before the actual meeting. In short, do what you can to reduce the cognitive load once the videoconference is underway.
Cause 5: Zoom latencies
A study published this week suggests that during Zoom conversations, latencies between the time a person says something and the time each speech sound is heard can disrupt the flow of conversation (and thus increase the stressfulness of online meetings).
The University of Michigan researchers who conducted this study point out that in theory, the latencies in question shouldn't be disruptive. The researchers calculated that on most Zoom calls, there's only about a 30 to 70 millisecond latency (i.e., delay) between the time a person speaks and the time each speech sound is heard. However, in ordinary conversation, the average gap between the time one person finishes speaking and another one begins is about 200 milliseconds (that is, about one fifth of a second).
Of course we sometimes start talking even before our conversational partner has finished their sentence; that 200 millisecond figure is just an average, and there's a lot of variability. But it does suggest that a 30 to 70 millisecond latency caused by Zoom shouldn't matter. And yet it does. The researchers found that responses to yes/no questions, and conversational turn-taking, were both slower when carried out over Zoom. (The researchers emphasize that the problem isn’t the Zoom latency per se, but rather the variability in latency from moment to moment. Even though a 70 millisecond delay wouldn't be perceptible under ordinary conditions, the 30-to-70-millisecond range throws us off, because our brains detect this variability, and it's consequently harder to stay in synch with the rhythm of the conversation.)
Fix 5: Awareness
Researchers haven't proposed a fix for this one yet. Perhaps just being aware of the problem – recognizing that conversation may be a little out-of-synch – and mentioning it to our Zoom partner can allow us to adjust our expectations and feel less fatigued. We're all in this together, after all.
Final thought
If you find videoconferencing a less-than-ideal experience, I encourage you to explore some of the fixes mentioned in this newsletter, as well as others that are floating around the internet. I say this because there are indications that even after the pandemic subsides, remote work, and therefore remote meetings, will continue to be much more prevalent than they were in the past.
In the meantime, I hope that your 2022 turns out however you wish, and that any surprises are good ones!
Appendix 1: Redunancy and factor loadings
The researchers claim that Zoom fatigue has different subcomponents – general, visual, social, motivational, and emotional. They obtained evidence for this claim through a statistical procedure called confirmatory factor analysis (CFA). Crudely speaking, CFA identifies groups of items for which each person's responses tend to be similar, while also distinct from their responses to other items.
In this study, CFA showed that each person responded in a similar way (i.e., consistently with themselves) to all the social fatigue questions, for example, and yet how they responded to these questions was unrelated to how they responded to other survey questions. For example, if a person expressed a strong desire to "be alone" after videoconferencing, they tended to express a strong desire for "time by yourself". If their desire for one was middling, their desire for the other tended to be middling too. But whatever they said in response to those questions was independent of, say, their visual fatigue. Among folks who reported a lot of social fatigue, some said their eyes felt good, some reported a middling visual experience, and some reported substantial visual fatigue.
Without getting into the statistical details of CFA, you can see why item redundancy is a huge problem. If one question in a group is virtually synonymous with another one, then of course peoples' responses to those questions will be identical or at least highly similar. In short, I don't think the ZEFS, in its current form, clearly identifies the different components of Zoom fatigue, much less which tend to co-occur or be independent of each other.