Intervention Roulette

Idea I've named based on a bit of Tim Leary history...

I remember some reference to him doing a comparison of psychotherapy to lots of other interventions (including voodoo) and concluding that in pretty much all of them got the same results (1/3 get better, 1/3 stay the same, 1/3 get worse).

• in Arno Ruthofer's Think Yourself, Question Authority: Leary quit his post at Berkeley because he felt that he was “practicing a profession that didn’t seem to work” (Flashbacks: 16). Psychology still had not developed a way of significantly and predictably changing human behavior. Leary’s studies showed that one third of the patients who received psychotherapy got better, one third got worse and one third stayed the same.
• If we assume that every intervention works for some people (note that without more info it could be that all interventions work for the same 1/3 of people, but I choose to assume these are random/independent outcomes), and that every person would find that some interventions would work, then your personal change strategy would be Little-Bet Iterative A-B Test-ing (Quantified Self for Experiment, Thinking in Bets) a variety of intervention experiments, right? (Self Improvement)
  • Beware Placebo/Hawthorne Effect!
  • also beware Side Effects
  • also want process to cap very-bad outcomes (Risk Management; Coaching?)
• Public Policy version (not the same): Piece Meal Social Engineering

Could odds be improved with a combination of

• Personality Tests
• Quantified Self + Thinking in Bets
• Crowd-Sourcing Big Data like Patients Like Me?