Versions Compared

Old Version 54

changes.mady.by.user Lorraine Sobson

Saved on

compared with

New Version 55

changes.mady.by.user Alana St. Clair

Saved on

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

Grouping variables are a subset of Identifier and Grouping Qualifier variablesused to group records in datasets by study, domain, across subjects, by subject, and for a subject. When used per this guidance, grouping variables and their values adhere to the following hierarchy:

Metadataspec

Hierarchy of Grouping Variables

STUDYID


DOMAIN

--CAT

--SCAT

USUBJID

--GRPID

The following are expectations for how values in these variables will group of records given their hierarchy.

Metadataspec
VariablesRecord GroupingPurpose of Grouping

STUDYID

By study

All records with the same STUDYID value are a group of records that describe that study.

DOMAIN

By domain

All records with the same DOMAIN value are a group of records that describe that domain.

--CAT, --SCAT

Across subjects

--CAT and --SCAT values subset groups of tests within a domain and apply to all subjects within the domain. 

USUBJID

By subject

All records with the same USUBJID value are a group of records that describe that subject.

--GRPID

For subjects

All records in the same domain with the same --GRPID value are a group of records within USUBJID. 

The optional grouping identifier variable --GRPID may be used in all domains based on the general observation classes. --GRPID identifies relationships between records within a USUBJID within a single domain and has no inherent meaning across subjects or across domains. Relationships between observations are defined by assigning the same unique character value to the --GRPID variable for sets of related observations. The values used for --GRPID can be any values the applicant chooses; however, if the applicant uses values with some embedded meaning (rather than arbitrary numbers), those values will be consistent. 

Using --GRPID in the general-observation class domains can reduce the number of records in the RELREC, SUPP--, and CO datasets, when those datasets represent relationships/associations for records or values to a "group" of general observation class records.

Pagenav