One of many tasks of OHS practitioners is to plan organizational responses to disasters. The most common kind of workplace disaster we develop plans for are building fires. In fact, the laws passed in the wake of the
Triangle Shirtwaist Fire are some of the earliest forms of OHS regulation.
In planning a response to a fire (hint: get out), it can be useful to know how people respond to rapidly evolving, high-stress, low-frequency events and why they respond that way. Last week, I ran across a 2004 article entitled “W
hy people freeze in an emergency: Temporal and cognitive constrains on survival responses”
Basically, the author looked at the literature, examined disaster inquiry reports, and interviewed a bunch of disaster survivors to verify the existence of freezing behaviour in disasters and quantify it. He concluded:
Responses to unfolding disaster can be divided broadly into three groups.
In the first group, between 10‐15% of people will remain relatively calm. They will be able to collect their thoughts quickly, their awareness of the situation will be intact, and their judgment and reasoning abilities will remain relatively unimpaired. They will be able to assess the situation, make a plan, and act on it.
The second group, comprising approximately 75% of the population, will be stunned and bewildered, showing impaired reasoning and sluggish thinking. They will behave in a reflexive, almost automatic manner.
The third group, comprising 10‐15% of the population, will tend to show a high degree of counterproductive behavior adding to their danger, such as uncontrolled weeping, confusion, screaming, and paralyzing anxiety (Leach, 2004).
This finding is important because, generally speaking, the mechanisms we create to allow people to protect or save themselves in such situations requires them to take immediate and sensible action (e.g., use fire exits, put in a life vest, open an emergency exit on a plane). If only 10-15% of people can be relied upon to do so, then these mechanisms likely won’t achieve their desired result (i.e., nobody dies).
The author attempts to explain maladaptive emergency behaviour by conjecturing that sub-optimal responses are related to our brains’ information-processing limitations. He asserts that, when faced with a novel event, our brain requires time to assess it, develop a plan, and execute it.
Disasters, which are novel and complex and involve significant stress, often unfold too quickly for us to meaningfully react. However, he says, since your brain can select among a pre-existing behaviours much faster than it can design new behaviours, training on how to respond can attenuate this effect. (This is why we do fire drills and why you get a safety briefing before every time a plane takes off.)
The explanation advanced by this article has intuitive appeal (i.e., it sounds plausible on first blush), but the question is whether the explanation is correct. Recall that the conclusion (i.e., brain too slow) is conjecture, rather than the results of any empirical testing. I spent some time looking for evidence that this conjecture was correct and didn’t find much (although this isn’t my field and maybe I looked in the wrong places; I also ran into a bunch of paywalls that I could not get past).
What I found was:
So maybe there is something to the original author’s explanation for this well documented phenomenon but YMMV. An important barrier to proving it is simulating the necessary degree of stress in an experiment.
The reason this article came up in my feed (I think on Bluesky but maybe Twitter) is that someone was likening the three-group typology to explain people’s reactions to Covid. Basically the asserted that the calm, muddled, and counterproductive groups in the 2004 disaster study are analogous to active avoiders, passive avoiders, and minimizers.
This analogy was intuitively appealing, bolstered by the seeming authority of the original study. It is useful, though, to deliberate a bit about whether the disasters that the original article looked at (e.g., a ferry sinking or a plane catching fire) is similar enough to Covid for conjectured explanation to apply. A key difference that jumps almost immediately to mind is the time scale.
Contemporary Covid behaviours are the result of a lengthy process. While Covid is a novel event, the time-scale is not the same as the disasters that the original author explored (where the speed of the disaster may have outpaced decision making).
So, while the proportion of active avoiders, passive avoiders, and minimizers may (or may not) mirror the groupings in the disaster study, the similarities between disasters and Covid are likely superficial and coincidental. Thus, we ought not put much stock in the claim that the 2004 study is in any was applicable or instructive to understanding Covid responses.
-- Bob Barnetson