Enum: AssignmentStructure

How the cohorts a study compares came about. Open vocabulary: write the source's own wording in the field when nothing here fits.

The first question is whether there are arms at all. Two cohorts are in parallel in the ordinary sense of the word, so a study that assigned nobody to anything reads as parallel on the plain meaning and is wrong. observational_cohorts is the value for it, and it has no arms.

URI: neurostudy_ex:AssignmentStructure

Permissible Values

Value	Meaning	Description
observational_cohorts	None	Two or more cohorts that were not assigned to anything: they differ by a ...
parallel	None	Each arm is a separate cohort, and no participant is in more than one, so the...
crossover	None	Every participant passes through several arms, in an order that was randomize...
within_subject	None	Every participant is measured on several occasions around one intervention, w...
single_group	None	One cohort measured once, with no arms and no repeated occasions to struc...

Identifier and Mapping Information

Schema Source

• from schema: https://neurostuff.org/schemas/neuroimaging-study-extraction

LinkML Source

name: AssignmentStructure
description: 'How the cohorts a study compares came about. Open vocabulary: write
  the source''s own wording in the field when nothing here fits.

  **The first question is whether there are arms at all.** Two cohorts are in parallel
  in the ordinary sense of the word, so a study that assigned nobody to anything reads
  as `parallel` on the plain meaning and is wrong. `observational_cohorts` is the
  value for it, and it has no arms.'
from_schema: https://neurostuff.org/schemas/neuroimaging-study-extraction
rank: 1000
permissible_values:
  observational_cohorts:
    text: observational_cohorts
    description: 'Two or more cohorts that were **not assigned to anything**: they
      differ by a characteristic the participants already had -- a diagnosis, an exposure,
      an age band, a trait score -- and nothing was administered. The commonest design
      in this literature, and it has **no arms**: leave `arms` empty and leave `Group.arm`
      empty, and let the cohorts reach an analysis through `FactorLevel.groups`, which
      is the slot that records which cohorts a contrast compared.

      Patients against healthy controls is this. So is smokers against non-smokers,
      or carriers against non-carriers. `parallel` is not: nobody allocated anyone
      to having bvFTD.

      One cohort rather than two is `single_group`. A cohort that was observed *and*
      given something -- an untreated patient group against a treated one -- has arms,
      so it is `parallel`.

      **A within-subject structure takes priority over this value.** Where the same
      participants were also measured under several arms or on several occasions,
      answer `crossover` or `within_subject` even though the cohorts themselves were
      not assigned: the cohort comparison is already recorded, in `FactorLevel.groups`,
      and the repeated-measures structure has no other slot to go in. A study comparing
      two cohorts under two portion sizes, or men against women across a stress manipulation,
      is a mixed design and this field records the within-subject half.'
  parallel:
    text: parallel
    description: 'Each arm is a separate cohort, and no participant is in more than
      one, so the allocation is a property of a Group: set `Group.arm`.

      Requires that something was administered or assigned, and that the study has
      arms to put in `arms`. A between-cohort comparison with nothing administered
      is `observational_cohorts`; this value asserts an allocation, and asserting
      one that never happened is the error it is easiest to make here.'
  crossover:
    text: crossover
    description: Every participant passes through several arms, in an order that was
      randomized or counterbalanced. No cohort was allocated to one arm, so `Group.arm`
      stays empty and the arms reach an analysis as levels of a model term, named
      in `FactorLevel.arms`.
  within_subject:
    text: within_subject
    description: Every participant is measured on several occasions around one intervention,
      with no randomized order of arms to cross over. A single-arm pre-post design
      is this rather than a crossover.
  single_group:
    text: single_group
    description: 'One cohort measured **once**, with no arms and no repeated occasions
      to structure. Both halves are required: two or more cohorts that were never
      assigned to anything is `observational_cohorts`, and one cohort measured on
      several occasions is `within_subject`.'