Myth-spotting

03 Oct 2018 Tags: Soft skills, Course design

As part of a larger project, I’m reading many articles from the management literature. This is a very different kind of writing and argument than I’m used to in the technical side of computing. The management literature (including the literature on management of computing personnel and projects) is also prone to fads, to such a degree that Harvard Business Review has an article on spotting management fads.

In addition to fads, management theory is also prone to myths. Where a fad is often closely tied to its major proponent, who typically derives a steady consulting revenue from it, myths are more widely accepted. Where the instigation of a fad can be precisely traced, typically to a publication such as The One-Minute Manager, the source of a myth is impossible to locate. It’s just always seemed to be there, a piece of received wisdom, and only on close examination does the viewer start to question it: “Everyone accepts it—but has any one demonstrated it?”

An example myth

An example of a widely-accepted myth is the so-called “Yerkes-Dodson Law”. I say so-called because neither Yerkes nor Dodson codified it into a law. As Michael Corbett recounts in his excellent paper, “Cold Comfort Firm: Lean Organisation and the Empirical Mirage of the Comfort Zone” (whose broader argument I will return to in a later post), their experiments were published in 1908 but it was only in 1957 that Broadhurst published a paper codifying their results as a broader “law”. The original data from Yerkes and Dodson in fact contradicts the typical presentation as a U-shaped graph (contrast the “Original” and “Hebbian” versions of the graph on Wikipedia).

Yet despite this longstanding inaccuracy, management theories, such as performance management, continue to base their recommendations on Yerkes and Dodson’s “law” and its supposed support for a “comfort zone”. What is it about such myths that ensures their persistence, even in the face of well-grounded contradictory evidence, and is there anything about the structure of such claims that suggests it might be a myth?

Why do we want to believe myths?

Myths persist because they serve a need: They help people navigate complex questions by providing easy answers that appear on their face to be valid. When first told the typical (mis)presentation of the Yerkes-Dodson results, that moderate stress encourages highest performance, while too little and too much stress reduce performance, it makes sense. We have typically had experiences where the right stakes—neither trivial nor life-threatening but the possibility of meaningful gains—improved our focus. We have also probably had days where we weren’t motivated to budge from the couch and others where the prospect of failure shrank our creativity or even froze us outright.

The problem with a myth is that it only conforms to experience but it does not actually explain it. The myth lays out a pattern into which we can place a few experiences selected to match. The cycle is quick and subtle: The myth provides a pattern, we locate a few experiences within that pattern, which we then interpret as validating the pattern set out by the myth.

This approach differs from how we interpret actual, well-demonstrated results in social psychology, such as the fundamental attribution error. In those cases, the original supporting evidence is a broad series of experiments, covering a wide range of circumstances. When we then hear of the principle and organize some of our own experiences according to the pattern, we are sense-making, integrating the principle into our experience, testing it out. Although we may be testing whether it is appropriate to our life and how it may fit with our established principles and values, we are not validating it. The validation preceded our learning about the result, during the period of testing in the literature.

The marks of a myth

Superficially, both well-grounded results and myths will appear to be supported by the literature. The definitive test is to do a detailed review of the primary literature but this is impractical: We do not have time to track down the sources of every claim, however widely-believed. On the other hand, brute contrarianism, the automatic gainsaying of all received knowledge, is a dead end as well. Are there strong surface indicators of a myth that might warn us not to take a claim as stated?

I believe such markers exist. This post is a draft of my approaches to testing a broad, widely-repeated principle to assess the likelihood that it can guide successful action or is just a well-meaning platitude.

Mark 1: Broad, vague terms

Myths traffic in broad, vague terms. In fact, myths probably require terms that sweep the horizon, encompassing the largest possible range. This provides an essential advantage: The myth appears to apply to any situation, all depending upon how the viewer interprets its key terms.

Consider the Yerkes-Dodson “law”. Although it is frequently described in the management literature as the effect of changes in “stress” on “performance”, Yerkes and Dodson’s original work used the more technical psychological concept of arousal, which has measurable indirect correlates in heart rate and blood pressure, as well as more direct correlates in activation of the ascending reticular activating system. In actual fact, Yerkes and Dodson did not measure any correlates at all, instead choosing to vary the external stimulus, an electrical shock.

Varying the external stimulus was an entirely acceptable choice for Yerkes and Dodson, given their goals and the technology of the time. The problem enters when this experimental operationalization is glossed as “stress” by those who want to apply it to management theory. Although there is face value in associating greater “stress” with an increase in electrical shock, the actual relationship is complex. In commmon use, “stress” refers to a complex mix of physiological and psychological characteristics. The measurements recorded by Yerkes and Dodson provide no way to assess the level of “stress” experienced by the mice in their experiments. Perhaps it was a linear relationship, perhaps logarithmic, perhaps quadratic, perhaps its form varied piecewise over several distinct ranges. There is no way to know. This ambiguity works to the advantage of the myth-makers, as the listener interprets the terms within their existing biases.

Mark 2: Conflating description and judgement

If the first part of a myth is framing the relationships between variables denoted by broad terms, the second part exploits the connotations of those terms.

Mythmakers not only prefer terms sufficiently vague to be broadly applied, they also prefer to use those terms in a manner that conflates description and judgement. Consider the term “stress”. On the one hand, that term is descriptive, characterizing a form of experience, one that can even be operationalized as measurable correlates. But it also, especially in contemporary usage, has an emotional valence, a connotation of undesirability. Stress is something to avoid, a situation to minimize or from which to be rescued, and in sufficient quantity is even a risk to your health.

To a management myth-maker, this conflation of description and judgement is not a problem but a resource. The descriptive element implies an objective, measurable, “scientific” aspect, while the unpleasant connotation gets the listener’s attention. “We’re talking about stress after all, and who wants too much of that?”. Better still if it can be exploited for contrarian spin, such as “Stress can be a good for you—here’s why!”.

Combining denotation and connotation serves another purpose for the mythmaker as well: It reduces a complex relationship to a single dimension. But this post is already long enough, so I will save that part for the next post.