experiment sandbox subject-level experiments

Cdb758075f7b54ea4ef646898497ecf3?s=47 cjlortie
October 06, 2019

experiment sandbox subject-level experiments

Life is an experiment of one. Within-subject designs are a crucial form of scientific inquiry. Simple design decisions improve change detection capacities.



October 06, 2019


  1. @cjlortie experiment sandbox resource: experimental design 4 the life sciences

  2. subject-level experiments

  3. life is an experiment of one

  4. within-subject designs [cross-over/repeated-measures] typically use sequential treatments to explore outcomes

    on same individual
  5. A to B change is the key response

  6. sequential treatments are an efficacious design, manipulative, & emulate natural

  7. limitations 1. temporal-ordering effects 2. carry-over effects

  8. solutions for each challenge

  9. counterbalancing ensures that the order of treatments is varied to

    encompass all permutations
  10. to redress carry-over effects design washout periods or buffers in

  11. test for reversibility, i.e. the capacity for subjects to return

    to pre-intervention states
  12. sequence effect depends on timing first-order and higher-order carry-overs

  13. rare that subjects return to baseline conditions

  14. better design focus n = 1 delta = 2

  15. measure change and collapse measures when possible to avoid pseudoreplication

  16. delta > 4 treatments switch from each first order(ing) to

  17. ideal strategy is random permutations of all order

  18. one block model typically and focus on sequence (of attack/intervention)

    to precipitate changes
  19. typically employ sequencing and not simultaneous treatments and can lead

    to more protracted experimentation
  20. consider coding time as a factors in models