Designing Variables

Variables form the building blocks of databases.

Each variable should have:

Variable Name

Machine-readable identifier.

Examples:

• age_years
• weight_kg
• temp_c

Avoid:

• age
• age1
• weightnew

Variable Label

Human-readable description.

Example:

Weight of participant at enrollment (kg)

Variable Type

Examples:

• Text
• Numeric
• Date
• Time
• Categorical
• Calculated

Permissible Values

Example:

Sex:

1 = Male

2 = Female

3 = Unknown

Variable Naming Conventions

Consistent naming improves database quality.

Recommended conventions:

Use Lowercase

Good:

weight_kg

Bad:

WeightKg

Use Underscores

Good:

visit_date

Bad:

visitDate

Include Units

Good:

height_cm

weight_kg

temp_c

Avoid Spaces

Good:

blood_pressure

Bad:

blood pressure

Coding Standards

Standardized coding improves interoperability.

Examples include:

Sex

1 = Male

2 = Female

Outcome

1 = Alive

2 = Dead

Yes/No Variables

0 = No

1 = Yes

Using consistent coding simplifies analysis and data sharing.

Visit Schedules and Longitudinal Data

Clinical studies frequently involve repeated observations.

Example:

Data managers must determine:

• Which variables repeat
• Which variables are collected once
• How repeat visits will be structured

In REDCap, repeated observations can be implemented using:

• Repeating instruments
• Longitudinal projects

Repeating events