Quantitative Bypassing
My grandparents recently shared with me a piece by Charles Eisenstein called On Political Bypassing (text / audio). He begins with the John Weiwood definition of spiritual bypassing: “the tendency to use spiritual ideas and practices to sidestep or avoid facing unresolved emotional issues, psychological wounds, and unfinished developmental tasks.” Eisenstein proposes his own phenomenon of political bypassing, which he describes in parallel structure as, “the tendency to use political narratives and framing in order avoid facing unresolved emotional issues, psychological wounds, and unfinished developmental tasks, or to assert social dominance.”
We could swap many adjectives in for spiritual or political and find a useful concept for the ways in which a particular worldview, mode of epistemics, source of identity, or aspect of our society can extend its reach beyond its sphere of utility and become detrimental and blinding.
Though to my knowledge a new term, quantitative bypassing is certainly an oft-expressed (or at least oft-implied) critique of schools of thought like rationalism, positivism, and effective altruism which privilege the empirical, the quantifiable, and the numerical. “In order to assert social dominance” seems the most relevant of the purposes for this particular form of bypassing, and to this we might also add: to assert intellectual dominance, to avoid the fear of the unknown, and to repress the discomfort of emotion and feeling.
This matter is far from theoretical for me: I run a program called The Mathematics of Morality where we help talented math and STEM students magnify the positive impact of their lives and careers on the most important problems in the world. One of the modern patron saints of understanding the world through math, Nate Silver (of fivethirtyeight fame), recently described quite well on an 80,000 Hours podcast episode what I hope our participants take away from much of the mathematical modeling we teach and practice: “A model is a tool or a device. It’s important not to use the model as necessarily being oracular. It’s a disciplining mechanism to force you to think through questions more rigorously.”
Because I do worry that, as the concept of quantitative bypassing suggests, our emphasis on the mathematical de-privileges in the experience of these participants the also-important modes of knowing, understanding, and deciding that are decidedly non-numerical.
And yet two things can be true at the same time:
- We need more quantitative reasoning in decision making.
- Among those who primarily reason quantitatively, we need more diversified ways of understanding the world.
This is to say that I find the critiques offered by the concept of quantitative bypassing accurate and useful directionally, but only from a certain place and only to a certain point.
Can something be both ironic and wholly unsurprising? Let’s try: I best understand the coexistence of the two points above through a mathematical example that I just recently referenced in my own 10th grade math classes. Chapter 1 of pop-mathematician Jordan Ellenberg’s book “How Not to be Wrong: The Power of Mathematical Thinking” begins with the following anecdote: “A few years ago, in the heat of the battle over the Affordable Care Act, Daniel J. Mitchell of the libertarian Cato Institute posted a blog entry with the provocative title Why is Obama Trying to Make America More Like Sweden when Swedes Are Trying to Be Less Like Sweden?” Ellenberg lambasts Mitchell for a presumably intentional (and if not, woefully ignorant) mischaracterization of what is of course a non-linear relationship between “Swedishness” and prosperity. It is perfectly compatible for both Sweden to need to become more like the US and for the US to need to become more like Sweden:
I use this example not to evoke some sentiment of Aristotelian affection for perfect moderation. Instead, I just mean something more simple: Prescriptions should depend on initial conditions.
So in a course built to emphasize and expand the role of math in personal, moral, and professional decision making — one that’s selected for students who are already highly predisposed to quantitative reasoning — there is good reason to put up some guardrails to limit the domain in which quantitative thinking should dominate. Yet I can maintain the need for a course like this for more young thinkers. The real phenomenon of quantitative bypassing is not a case against quantitative reasoning, just a case for doing so with more care and reflection.
Have I properly struck this balance? Certainly not perfectly, so I welcome your comments and thoughts on the section of our curriculum called “Quantifying Carefully,” which concludes our Week 2: “The mathematics of what matters most.”
Get Jonah Boucher’s stories in your inbox
Join Medium for free to get updates from this writer.
“Let’s consider two famous quotes:
- “Not everything that can be counted counts, and not everything that counts can be counted.” -Often misattributed to Einstein though the true origin is unknown.
- “All models are wrong, but some are useful.” -British statistician George Box
We’ve made the case in this section that you can quantify anything, but often only with rough approximations or simplifying estimates. The first quote reminds us that these quantifications are necessarily incomplete and imperfect, but the second emphasizes that this process can be useful anyway. You can, for example, use estimation to establish upper or lower bounds on the quantities you care about, and you can find dramatic order of magnitude differences between alternatives that are strong evidence for a real underlying difference even if the assessment tools are but a crude look.
This process of quantification is notoriously fraught with cognitive traps, risks, and biases. Consider this perhaps all-too-relevant example:
If you’re a teenager in the UK, you’re probably used to your school’s performance being ranked in league tables based on exam results. The pressure to perform can be immense, not just for you, but for your teachers and school administrators as well.
When the government introduced school league tables in the 1990s to promote accountability and improve educational standards, the goal was to create transparency and motivate schools to perform better. In many ways, they have likely done just this. However, the intense focus on these rankings led to some unintended consequences.
To climb up the rankings, schools started to game the system. For instance, some schools focused disproportionately on students who were on the C/D grade borderline to boost their overall pass rates, while neglecting both higher achievers and those struggling the most. This led to a narrow focus on exam preparation, often at the expense of a well-rounded education.
In some extreme cases, schools manipulated enrollment practices to exclude lower-performing students from taking exams, thereby artificially inflating their performance metrics. These actions compromised the integrity of the educational system and did not genuinely reflect the students’ needs or potential. The metric itself had become the end-game, not the actual result the metric was there to measure.
[Sources: BBC Viewpoints, “What gets measured gets done”, The Guardian]
This was a classic case of Campbell’s Law, which says that the more any quantitative indicator is used for social decision-making, the more it will be subject to corruption pressures that can distort the social processes it is intended to monitor.
Let’s mention a few other biases so common they’ve also earned their own name, and then you can learn more about examples of each in the links below.
- Simpson’s Paradox is when trends that appear in different groups of data disappear or reverse when the groups are combined. Over-reliance on aggregated metrics can obscure important subgroup variations and lead to incorrect conclusions.
- False Precision refers to quantification giving a false sense of certainty. In reality, many quantified measures are based on estimates, assumptions, and imperfect data. Over-relying on these numbers can lead to overconfidence in decisions, especially when estimations leave the hands of those who originally made them and appreciate their limitations.
- Overfitting is when, in statistical modelling, a model is too complex and captures noise rather than the underlying pattern. This can result from an overemphasis on optimizing quantitative metrics.
- Confirmation bias is where individuals seek out or interpret data in a way that confirms their pre-existing beliefs. By selectively choosing metrics that support their views, people can reinforce their biases.
Despite these potential pitfalls, the utility of quantitative estimates in moral decision-making is still vast! Quantitative estimates allow for systematic comparisons and informed decision-making, even in the context of uncertainty. We’ve seen how quantitative methods have been instrumental in fields such as public health, where estimating the spread of diseases can help allocate resources effectively and save lives. Similarly, in environmental science, models predicting climate change impacts guide crucial policy decisions. In economics, cost-benefit analyses, which rely on quantitative estimates, are fundamental in evaluating the trade-offs of different policy options. In each of these fields uncertainty introduces the risk of biased estimation, but experts and organizations can systematically reduce these potential hazards with rigorous practice. And as we saw through the Scale/Neglectedness/Tractability framework, we can even use quantitative reasoning to make sense of which problems in our world are most effective for us to focus on in the first place.
So, when used judiciously, estimation tools provide powerful insights that are indispensable in navigating the complex moral landscapes of our time.
I believe there’s quite a lot to add and reinterpret here based on the left brain — right brain metahistory work of Iain McGilchrist (who I’ve written on before in Valueception and Your attention, please) in The Master and His Emissary. Any defense of left-brain modes of understanding in a world dominated — according to McGilchrist, to our detriment — by this very mode of thinking is bound to miss important aspects of the water we’re swimming in. I would love to hear from others with clearer views of this water or deeper knowledge about McGilchrist’s work, and anybody curious about or with thoughts on the Mathematics of Morality course!
