Curation Procedures: Naming Conventions🔗

Variable naming conventions for the release data are described on the HBCD Release Data Docs site on the Metadata & Naming Conventions under Naming Conventions. The naming convention is applied when data is transferred from LORIS to Lasso for release staging.

Source Data Naming Schemes🔗

Prior to staging in Lasso, instrument and fieldname conventions generally follow the REDCap naming standardization scheme, but may differ based on the source of the data, including:

Data Source	Naming Convention Notes
REDCap surveys	Variables are named according to the conventions defined in REDCap and imported into LORIS via the REDCap Data Dictionary (DD). These variable names generally match those in the REDCap DD, as LORIS instruments are created directly from that dictionary.
LORIS native forms	Initially coded using a flexible scheme, these instruments were later adapted to align with the REDCap naming convention. Field and instrument names now generally follow the REDCap standard, with only minor deviations in some cases.
LORIS Core Fields	These fields are not associated with instruments and have their own pre-established names that may or may not be be adapted for the Data Release.
Third-party instruments	Collected externally and imported into LORIS via parsing scripts, these instruments function like LORIS native forms and follow the same naming convention, typically adhering to the standardized REDCap scheme.
Ripple / ETL fields	Fields related to screening, demographics, or other study metadata, such as Transition in Care (TIC), Alternate Caregiver (ACG), Geocoding, or Study Navigator (SN), are collected or calculated in Ripple or via ETL then transferred to LORIS. These fields may not fully follow the REDCap naming convention and are usually stored at the participant or session level.

Curation Procedures🔗

When data are transferred to Lasso for release staging, the variable names are updated to follow the naming conventions, mentioned above, described on the central HBCD Data Release Docs site here. Here we describe certain aspects of how the naming conventions are instituted.

REDCap Naming Convention: _i_🔗

Table names in Lasso using single ( _ ) and double ( __ ) underscores, as explained here. In REDCap and subsequently LORIS, _i_ is used in place of double underscores for instruments and fieldnames to denote hierarchies for scales and sub-scales in the instrument name and field counters. For the Data Release, the i is removed, resulting in __ instead of _i_. This naming convention conversion occurs when the data is transferred from LORIS to Lasso.

Correspondence to JSON Metadata🔗

The domain and source are included in the JSON metadata and are typically derived from the corresponding sections of the instrument name. However, in some cases, data are collected directly into fields or tables that do not follow the standard naming convention. In those instances, the domain and source values are added later during the Data Release process.

This applies to:

• BioSpecimens
• Imaging file based data & derivatives
• Some session-level elements (e.g. informantID)
• Participant-level data

Known Issues🔗

Below is a running list of variables that need to be updated to conform to naming conventions aligned with ABCD (see outline of current naming standards here). Note that variable names are fairly internally consistent within tables (as they are often dictated by the conventions of the source platform), so the issues and examples below will generally apply to all of the variable names within a given instrument (which should make implementing changes towards standardization a more straightforward process).

See this spreadsheet for additional notes and details.

Issue 1: Scale should be a subcomponent of table🔗

As described here, scale is currently included in the Release 1.0 variables as a main naming component (separated by single underscores):

domain_source_table_{scale}_item

This perhaps follows this REDCap standardization documentation. However, it makes the number of naming components across variables inconsistent (i.e. a mixture of 4 or 5), because not all variables have scales. It is also inconsistent with ABCD naming conventions, which uses double underscores to make scale a subcomponent of table, resulting in 4 consistent main components across variables:

domain_source_table_item

A rough estimate from a simple parsing script shows the majority of variables in R1.0 currently follow the convention that includes scale as a main component (3531 vs. 965). Below are impacted instruments with examples variables and fixes:

Domain	Table	Table Name	Example Variable	Example Fix
BCGI	IBQ-R (VSF)+BI	mh_cg_ibqr	mh_cg_ibqr_efrt_011	__efrt
NCL	SPM-2	ncl_cg_spm2__inf	ncl_cg_spm2__inf_soc_001	__soc
PH	Growth	ph_ch_anthro	ph_ch_anthro_head_001__01	__head
PEX	ALL preg health tables, eg Chronic conditions:	pex_bm_health_preg__chroncond	pex_bm_health_preg__chroncond_001___1	__preg
PEX	FAM MH	pex_bm_psych	pex_bm_psych_bf_001	__bf
PEX	ASSISTV1/2/3	pex_bm_assistv1	pex_bm_assistv1_lt__use_001	__lt
SED	BFY	sed_bm_bfy	sed_bm_bfy_econstr_008	__econstr
SED	Demographics	sed_bm_demo	sed_bm_demo_herit_002__06___2	__herit
SED	PROMIS	sed_bm_strsup	sed_bm_strsup_socspprt_001	__socspprt

Issue 2: Scale should be a subcomponent of table (Issue 1) + Add double underscores for item components🔗

The following are table variables that have scale as a separate main component and also require additional double underscores in order to nest the item subcomponents:

Domain	Table	Table Name	Example Variable	Example Fix
NT	Infant Questionnaire	nt_ch_sens__qtn_1	nt_ch_sens__qtn_1_beh_002	__1__
PEX	APA 1/2	pex_bm_apa	pex_bm_apa_1_depr_001	__1__
PEX	TLFB	pex_ch_tlfb	pex_ch_tlfb_alc_wk_01	__alc_wk__
Biospec	Nails	bio_bm_biosample_nails_results
Biospec	Nails	bio_bm_biosample_nails_type
Biospec	Urine	bio_bm_biosample_urine	bio_bm_biosample_urine_bio_c_pcp_u

Issue 3: Admin and summary score variables - Replace frequent single underscores🔗

Admin and summary score variables often have single underscores that should be replaced by double underscores, e.g., date_taken, candidate_age, gestational_age, adjusted_age, summary_score, total_score, etc. See HBCD Docs for additional details.