Risk Factor Misjudgement

Description.

Components.

Time Scope

We're less likely to notice deterioration if it happens slowly compared to human timescales. This is a case of risk factor misjudgement: if we don't notice a slowly-deteriorating situation, we will treat it as having a probability of zero even though it may kill millions of people in future generations. For example, when humans started burning fossil fuels at a greater rate, there was no evidence that this could cause global warming.

Further reading.

non-linear perception of p and q

Images: aircraft crash (media or Air Crash Investigation) biasing intuition

Probability

• Small p are rounded down or assumed zero; eg, danger of driving or smoking. "It will never happen to me."
• Large p are rounded up (cf overconfidence qv Cromwell).
• black-and-white thinking (AKO oversimplification); eg,  the nuclear reactor WILL explode. link from probability, complex-behaviour qv. P overestimated because Q is high. Fear of flying likewise.
• 'framing effect': eg, 20% will die is perceived differently to 80% will recover. (Chivers 179)
• tendency to base p on what we notice or remember, such as aircraft accidents but not instances where aircraft didn’t crash. (Chivers 177ff)
• 'recency bias': recent events are over-emphasised. (Chivers 188). This could actually be beneficial if the system structure can change: underestimating-change qv.
• bias towards recent events rather than whole history (PhD 47 and Sterman 2000; cf familiarity-bias qv). Rare events, and those that could only arise from a system change (novelty qv), will have p underestimated.
• In the months after Sep 11, many Americans opted to drive long distances instead of flying. This may have led to about 2,300 extra deaths. (Chivers 189). If so, this risk misperception killed almost as many people as the event itself.
• 'frequency bias': what you've seen most often. (Chivers 188)
• p underestimated (and seemingly implausible) if novel qv because no relevant history/memory to guide instinct qv, intuition qv, or analysis.

Consequences

• 'anchoring': original assessment isn't updated on the basis of new events. (Chivers 188)
• Assuming the worst; eg, every passenger always dies when an aircraft crashes. (see .md)

Related

• cf. Optimism qv,  probability qv (risk management)
• cf underestimating-change qv, which focuses on system change over time which can increase p and/or q, leading to risk-factor-misjudgement
• cf familiarity-bias qv: if we look only short-term, we might not notice variability that indicates probability, leading us to think B&W.

---

• PhD thesis: §2.2.2.3: parse and refer.
• Underestimation:
  • Unjustified doubt, disbelief (eg, climate change denial/scepticism: trust-error qv). Underestimation of problem significance (risk=p.q). Problem can't really be bad because we're entitled to better (denial), perhaps due to cultural beliefs (cf resistance-to-change qv). See denialism.
  • Scientists will come up with something. ”...then a miracle happens...”: cf optimism qv (what is this now? overconfidence? entitlement?).
  • It's good to think critically about both sides of a theory, but it's bad to adopt a view simply that belief is more desirable (eg, avoids unpleasant actions). Belief should be based on what we believe to be probably true, not what we want to be true. confirmation-bias q.v. can be used to support underestimation qv.
  • Scepticism isn't entirely bad. The alternative is gullibility. We should not unquestioningly accept the point of view of a climate change denier, but neither should we unquestioningly accept the point of view of a climate change doomsayer — or even any academic. Everyone has biases. The most lauded academics (eg, Newton re cosmology and quantum mechanics) have sometimes been shown to be wrong, sometimes centuries later. Expecting non-experts to judge between the competing positions of those with greater knowledge is a tough call (and maybe unreasonable and fraught with risk). However, some problems (eg, global warming) can't be solved by a few academics or politicians: they will require action by the majority of the population. In addition, in democracies, the population can decide who the politicians are and what their priorities are. Even in non-democracies, regime change is possible!
  • Distinction with risk-factor-estimation: underestimation is tendency or desire to downplay problem risk due to wanting the problem to be milder or non-existent. risk-factor-misjudgement is simple misjudgement of p and/or q.
  • * Distinction with overconfidence qv: underestimation underestimates severity of problem (p&/q) whereas overconfidence overestimates likely effectiveness of actions.
  • 'What you don't know can't hurt you'. Yes, it can. Big/slow problems may not hurt me but may hurt others (including future others).
  • [was 'wishful thinking']
  • [was 'underestimation']
• Rename 'misjudgement'? Ensure unambiguous.
• [was 'risk factor estimation' or 'inaccurate' or 'black-and-white' or 'shades-of-grey']