intuition vs. analysis

3 key features of intuitive decision-making

• process dominated by subconscious

• see situation as a whole and not in parts

• connected with your emotions; something does/doesn’t feel right

intuitive processing can be

remarkably accurate

thin slice judgements

tiny, quick judgments can be surprisingly accurate

unconscious processing can be useful for making decisions

easier to pick jams

unconscious thought theory

• conscious perform well under simple conditions and poorly under demanding ones

• unconscious performed consistently well

our unconscious can

impact decisions in all sorts of ways (not only positively)

priming effects

• being exposed to a concept can impact

  • social judgment

  • evaluations

  • behavior

“moneyball approach”

relying on a linear model so pitfalls of “human” judgment won’t get in the way of a good choice

linear models are vastly superior

linear model dramatically outperforms admission committee ratings as a predictor of program success

intuition vs. analysis

despite era of “big data” many major decisions still made based on instinct

when is intuition useful?

when based on experience in seeing relevant patterns… not the same as just a gut instinct

keeping intuition in check

• self checking

  • don’t be fooled by intuition (confirmation bias, dissonance, self-serving biases, etc.)

  • continual feedback…willingness to adjust

• “fact-based culture”

data has to come from somewhere

source can be incorrect or misleading

someone still has to interpret the data

data interpretation is more of an art than a science, even experts, given the same daaset, might reach opposing results

lots of data doesn’t mean you’re “right”

biases in algorithm creation process + algorithms reflect existing biases

is the goal always efficiency or “success”

yeah

building “quantitative intuition”

• what data was collected?

• what data am i not seeing” (surveyor bias)

• can i trust the data and analyses?

• putting the data into context

• pressure test the analyses - what happens when you change key numbers

  • if it looks wrong, it probably is

  • if it looks right, it may still be wrong