Latest / Safe As Podcast / E73: When business indicators become the problem

Transcript

E73: When business indicators become the problem · Safe As Podcast

• 0:00
  Imagine the moment a new parent first holds their child.
• 0:03
  The warmth, scent, emotion, and then they're handed a number, an
• 0:08
  APKA score. And just like the bitterness
• 0:10
  score of a craft beer, modern life is increasingly distilled
• 0:13
  into digits. But here's an unsettling
• 0:16
  proposition. Turning equality into a number
• 0:19
  doesn't just describe reality, it changes how we value it.
• 0:23
  Today's paper is from Chang ET al 2024 titled Does Counting
• 0:27
  Change What Counts? Quantification fixation biases
• 0:30
  decision making published in Proceedings of the National
• 0:33
  Academy of Sciences. It describes psychological
• 0:36
  phenomenon called quantification fixation.
• 0:38
  It's the tendency to privilege numbers, favouring options that
• 0:42
  score higher numerically, even when qualitative factors matter
• 0:46
  just as much. When we count something, we
• 0:49
  inadvertently make it count more in our minds.
• 0:52
  To test this bias, researchers ran 21 pre registered
• 0:55
  experiments with over 23,000 participants.
• 0:58
  They examine decisions across managerial, policy and consumer
• 1:02
  context, from hiring interns and promoting engineers to choosing
• 1:07
  hotels and donating to charities.
• 1:10
  In each scenario, participants faced a trade off, like choosing
• 1:13
  between a candidate with high grades or one with richer
• 1:16
  experience. The researchers randomised
• 1:19
  whether attributes were shown as numbers or as verbal
• 1:22
  descriptions or visual icons. The results were consistent.
• 1:26
  People shifted towards whichever option dominated numerically.
• 1:30
  In one experiment, participants were more likely to choose a
• 1:33
  higher rated or expensive hotel when the rating appeared as a
• 1:37
  number like 5.0 rather than as an icon of five stars.
• 1:42
  Even in incentive compatible tasks, where bonus pay depended
• 1:46
  on making the right hiring choice, participants still
• 1:49
  overweighted quantified scores, even when doing so reduce their
• 1:53
  earnings. Why does this happen?
• 1:55
  The answer lies in what the researchers call comparison
• 1:57
  fluency. Numbers are uniquely comparable.
• 2:00
  We can see exactly how much larger one value is than
• 2:03
  another. This creates A subjective sense
• 2:06
  of ease. We don't just see numbers as
• 2:08
  more precise, we find them easier to use when making trade
• 2:12
  offs. This bias is strongest in people
• 2:14
  high and subjective numeracy, those most comfortable with
• 2:17
  numbers. Conversely, when numbers are
• 2:20
  made disfluent presented in harder to process formats like
• 2:24
  fractions, the effect disappears.
• 2:26
  So what can we make of the findings?
• 2:28
  For those who design decision environments or lead teams,
• 2:32
  these findings have some implications.
• 2:34
  Audit your choice architecture. If one metric is quantified and
• 2:38
  others aren't, decision makers will overweigh the numbers.
• 2:42
  Balance the fluency. Make qualitative attributes
• 2:45
  easier to compare using clear standards or shared scales.
• 2:48
  Beware metric fixation. Cantable things can create a
• 2:52
  richer but less measurable experience, like on a board
• 2:55
  report. And my favourite, use strategic
• 2:57
  disfluency. Slightly complicating a number
• 3:00
  can force deeper thinking and reduce fixation.
• 3:03
  Turn things into fractions and other numbers that require more
• 3:06
  cognitive processing. The study noted some
• 3:08
  limitations. The experiments don't fully
• 3:10
  disentangle actual computational lays from the felt A's of using
• 3:15
  numbers. In short, numbers act like a
• 3:17
  high powered lens, sharpening focus while blurring everything
• 3:21
  else. That's it on safe as.
• 3:23
  Thanks everyone. Please remember to like and
• 3:25
  subscribe, and visit safetyinsights.org for more
• 3:28
  research and links in comment if you'd like to support the
• 3:31
  channel.