In the Fall of 2020, my colleague, Prof. Immaculata De Vivo of the Harvard School of Public Health, and I wrote an essay about the public perception of risk and uncertainty, especially with regard to COVID-19. In this post, we are gathering comments from students in the Spring 2021 edition of "GenEd 1112: The Past and Present of the Future," an undergraduate course I teach at Harvard. Students were asked to read the essay, and then comment here on which part(s) of the discussion they expect would be most illuminating for non-quantitatively-inclined readers --and/or to suggest another framing of the issues discussed that would be more effective.
top of page
bottom of page
The parts of the essay that give actual, quantitative examples about each scenario would likely be the most helpful for non-quantitatively-inclined readers. Simply telling someone what the difference between risk and uncertainty is not very useful in a vacuum, but having one control "variable" (i.e., the level of risk) helps to illustrate it greatly. That is to say, showing how the moon landing has high risk because the astronaughts might die in any number of ways, and how a rare disease would have low risk if it doesn't kill often, but both have high uncertainty due to the large unknowns of a moon landing or the disease's erratic behavior, provides extremely helpful contrast for demonstrating the difference between the terms.
In the essay, the part that was most illuminating was explaining that uncertainty is inherent in risk assessment, and it's not just about whether something will happen or won't happen, but also about the degree of certainty regarding the outcome. In thinking about COVID-19, the uncertainty associated with the pandemic relates to some of the other notable events mentioned in the essay. Some of those examples used were Russian Roulette, lying on a couch, and the Apollo 11 mission, which helped illustrate the different levels of risk and uncertainty people encounter in their lives.
I expect the following parts of the discussion to be the most illuminating for non-quantitatively-inclined readers. First, the contrasting examples given of high risk/low uncertainty (Russian roulette), low risk/low uncertainty (lying on the couch), high risk/high uncertainty (Apollo 11), and low risk/high uncertainty ( a new disease that behaves erratically), illustrate how the concepts of risk and uncertainty work together clearly. I think another part of the discussion that provides further useful information to readers is how the COVID-19 mortality risk is put into perspective by comparing it to the overall annual death rate in the US, and to past pandemics. This helps readers understand the true risk level of the pandemic, and does not raise as much of an emotional reaction to the statistics of death tolls presented, although we should still be wary of them. Understanding how we have gotten used to a low-risk world and the way in which people misunderstand uncertainty is crucial to non-quantitively-inclined readers.
The part of the discussion that was most illuminating for me (a non-quantitatively-inclined reader) was the clear way uncertainty and risk - and the differences between them - were explained. Using examples like the Apollo missions (high risk AND high uncertainty) and russian roulette (high risk, low uncertainty) really put things into perspective. Therefore, the high uncertainty nature of the COVID-19 projections became much more understandable in this framework.
Hello,
I found this reading to be thought-provoking. Many people in society will spew random statistics they saw or heard on the news without considering the factors that may influence this data. This is what non-quantitative thinkers will look at; due to this, they do not grasp the true complexity of the situation and what it takes to resolve such an event. For the Russian roulette analogy, many people may think, "Yes, there is a 17% chance of a round being fired, causing almost certain death," but those are not the true odds. I say this because the essay mentioned that revolvers misfire or malfunction, which can affect 1 in 6 chances of a live round going off. This is uncertainty because we are, in a sense, in the dark about the millions of factors that could have affected the weapon, person's hand movement, etc., which means without this information, we will never know the true odds of death in any given Russian roulette round.
Best,
Joey Cano
For non-quantitatively-inclined readers, I would expect the introduction to be the most illuminating. The analogy to Russian Roulette and the deer hunting were helpful for setting up the conversation on risk and uncertainty. By bringing up this example throughout the post, the article provides a relatable way to understand these complex concepts. Using the Apollo 11 voyage as an example for high-risk and high-uncertainty was another helpful analogy made for non-quantitatively-inclined readers. Using COVID-19 was also another good example that grounded the reader in the concepts of data and uncertainty.
I think that the discussion on the uncertainty around the risk of COVID would be very helpful for non-quantitatively-inclined readers. When you are estimating something like risk, you have an average estimate of the risk and some degree of uncertainty around your estimate. This article’s discussion on COVID risk does a great job of illustrating how with few data points, we can still find an estimate of risk but we will have a larger uncertainty. This is because we did not know enough about the virus (early in the pandemic) to be certain if our best estimate of risk is close to the true value. As we see more cases of COVID, we can say with more certainty what the actual risk of the virus is since it is far less likely that we are just seeing a non-reflective sample of how dangerous COVID is. Knowing this concept is important in statistics since all estimates have some degree of uncertainty and understanding that uncertainty can help a lot in interpreting the statistical values we see in the world.
Readers who are not as familiar with quantitative concepts probably would find the section distinguishing "risk" and "uncertainty" especially helpful and insightful. This part of the discussion demonstrates how risks can be quantified, whereas uncertainties often cannot. Using accessible scenarios, such as playing Russian Roulette versus encountering unpredictable challenges in hunting, would help readers clarify the implications of each in real-world decision-making. Understanding this distinction can profoundly impact how individuals and organizations prepare for future challenges, emphasizing the importance of flexibility and resilience when handling unknowns.
The parts of the discussion that I expect would be the most illuminating for non-quantitatively inclined readers would be the introduction with analogies and the description of human perception of risk. The initial introduction, comparing the certainty of outcomes in hunting and Russian Roulette, can help non-quantitative readers grasp the fundamental differences between risk and uncertainty because analogies provide a relatable way to understand complex concepts. The section discussing how humans typically perceive risk, categorizing it into "it will happen," "it won't happen," and "it might happen," can resonate with non-quantitative readers because it calls awareness to something that they do without thinking.
I found this essay to be extremely accessible to those non-quantitatively-inclined, simply because hardly any numbers/math are mentioned -- when they are, they're (usually) representing a probability (essentially a well-know fraction with which most are familiar). On the difference between risk and uncertainty, the Deer Hunter "one shot" example versus the Russian Roulette game was a perfect example used to highlight such differences. The reason for this is that both examples fall on the opposite ends of an uncertainty spectrum, so their differences are easier to recognize. Its fairly straightforward to understand you risk an almost 1/6 chance of death from Russian roulette, and most people can understand how the uncertainty of this event is practically zero. This is simply because most people understand that if you shoot yourself in the head with a loaded gun, you'll most likely die. On the other hand, in the case of deer hunting and killing a deer with one shot, most people understand that there are higher levels of uncertainty associated with each risk. This is because there are more factors (weather, area, skill of hunter, rifle type, etc.) involved when it comes to hunting, rather than simply holding a gun to you head and pulling the trigger.
For non-quantitatively inclined readers, the discussion on how fear and uncertainty shape perceptions of risk, especially in the context of COVID-19, would be particularly enlightening. It delves into how the novelty and invisibility of the virus, coupled with initial data scarcity, led to heightened anxiety and exaggerated perceptions of risk. By highlighting the emotional reactions triggered by the pandemic and their impact on public responses and policy decisions, it provides valuable insights into the complexities of risk assessment. This segment underscores the importance of clear communication about known risks and uncertainties, as well as the need for transparency in decision-making processes, to navigate uncertain situations more effectively. Ultimately, it offers a pathway towards fostering resilience and informed decision-making in the face of uncertainty.
In order to better understand a high-risk high reward situation, you can change the framing to one of a roller coaster simulation. Imagine A new type of roller coaster gets introduced at an amusement park. It promises safety and excitement. Being the first of its kind, predicting safety is difficult. Rides can break down sometimes. Without past examples of this specific ride, calculating risks proves challenging. This lack of information makes trying the new ride uncertain regarding potential risk. Trying a new rollercoaster and being an early reviewer can be fun, but it may also come with alot of dangers
I believe that the distinction between risk and uncertainty will be the most illuminating for the quantitatively non-inclined; especially the example of a (essentially) 0 uncertainty game with known risk odds. Obviously, the example of Russian Roulette in The Deer Hunter is a grizzly way of conveying such truths, but given the statistical principles of the game are very transparent and easy to conceptualize, its use allows the reader to focus on the math behind risk as opposed to spending time understanding the specific mechanism of a hypothetical example.
As a non-quantitatively inclined reader, I found the example from The Deer Hunter to be the most illuminating. Although I had no prior knowledge of the movie, these simple and easy-to-understand examples helped me comprehend the difference between risk and uncertainty without a high understanding of statistics or data science. Simple examples like this make the essay more accessible to a broader audience that does not need to be an expert in the topic to understand the message the writer is trying to send.
I believe the parts that were most illuminating for non-quantitatively-inclined readers was the comparison between the low-risk low-uncertainty situations in contrast to the high-risk in general (both high uncertainty and low). But, for this example, let's keep it at high-risk low-uncertainty. People are generally good at making judgements on things that are obvious, so, when presented with the option of lying on a couch, versus participating in a round of Russian Roulette, then people would be highly inclined to take the couch option, even though they know the uncertainty of Russian Roulette. Even without knowing exactly the "odds" as the essay mentioned, non-quantitatively-inclined people will be good at measuring risk without the need to pay too much detail for uncertainty.
As a non-quantitatively inclined person myself, I think the use of the film The Deer Hunter was a good rhetorical device for breaking down the difference between uncertainty and risk and then walking through what different combinations of these concepts entail. For example, it made a lot of sense to me that the Russian Roulette game had a high amount of risk, but a low amount of uncertainty. The risk refers to the chance of dying while playing the game, while the uncertainty refers to the chance of the risk calculation being off. Because of how close of a shot the Russian Roulette game provides, it is unlikely that there will be much uncertainty. Continuing with this line of thinking, the hunting of a deer could be viewed as a high risk, high uncertainty situation. If we assume the hunter is pretty good and only takes shots that they have a good chance of hitting, as from my understanding most hunters do, then we say there is a high risk. But, similar to the Apollo example, there is high amount of uncertainty, over a longer distance with more variables, it becomes much more challenging to evaluate the exact risk.
1. Which part(s) of the discussion do you expect would be most illuminating for non-quantitatively-inclined readers?
I think that throughout the discussion, many clarifying points are made to make the piece more accessible to non-qualitatively-inclined readers, such as clarifying the fraction's numerator and denominator, or expliciting that 1-in-a-hundred is 1%. Also, the mathematical calculations that are written out are very useful and help the reader follow along and not feel lost.
1. Which part(s) of the discussion do you expect would be most illuminating for non-quantitatively-inclined readers?
I think the examples used for high-risk and high-uncertainty versus low-risk and low-uncertainty situations are helpful for illuminating the concepts herein for readers who are not quantitatively oriented. While numerical summations of risk and uncertainty may be opaque to these readers, I suspect they would have a strong reaction to their comfort level in playing Russian Roulette versus lying on the couch. I also think the example of Covid-19 is very effective for evoking these feelings as well — people can likely viscerally recall how they felt about the uncertainty of that situation, putting into context the importance of grasping uncertainty when dealing with potentially dangerous situations.
2. Suggest another framing of the issues discussed that would be more effective.
I wasn’t familiar with the Deer Hunter, honestly, and I think a large chunk of the potential audience for this paper likely is not as well. A fun example for risk that is familiar to younger audiences might be Avengers: Infinity War, in which Thanos’ famous snap wipes out half of all life. Here, as in Russian Roulette, there is extreme precision in both uncertainty and risk — you will without exception die should you be “selected” by the snap, and you have an exactly 50% chance of being selected. This might not be the perfect example, as technically you are not opting into this, but perhaps you could put the reader in the shoes of Thanos, who himself had that same level of risk and uncertainty when he chose to take the snap. Maybe this is a bit cheesy, but I do think it could be a fun and accessible way of explaining these concepts!
The essay on ‘uncertain risks’ provides qualitative and quantitative insight on risk vs uncertainty. Here, risk is defined as the fraction of number of people affected divided by the people that could have been affected. The part of the essay that would be most illuminating for non-quantitatively-inclined readers is the discussion about uncertainty in the case of russian roulette. Here, uncertainty is illustrated through a very logical way, as a percentage of times a gun would not work or a person would not die. This is a non-quantitative explanation of a usually quantitative-heavy subject of uncertainty.
Another framing that might be more effective in its own way is diving deeper into how figures for uncertainty are arrived at, how the uncertainty of 370k people is arrived at. Is it constructed based on previous similar pandemics or current, ongoing data?
I thought most of the article was geared well to non-quantitatively-geared readers. Throughout, it uses examples, never getting too theoretical without concrete, real-world examples. The movie reference was particularly helpful, although the article left me on a cliff hanger because I don't know whether or not the guy survived Russian Roulette or not (and I didn't click on the link because the movie sounds like it has some upsetting themes, so I didn't want to read the full plot). The one example that I think would be a bit hard to follow if you were not used to quantitative thinking was the Covid one. Since Covid is so recent, it makes it hard to look back on in the ways that I could with the other examples, and my mind got bogged down with so many connotations of Covid risk that it was hard to focus on learning the content. However, overall, I thought the article did a good job of providing information to a non-quantitatively-geared audience.