We recommend a research design consult for all clients with research studies.
For Longitudinal Studies:
- Biostatisticians may have a preference for output data that will impact what instruments users design in REDCap and whether or not a longitudinal design is recommended.
- In general, they may make recommendations and specifications for variable structure, and they are unlikely to work with open-ended formats. They do require that the data collected is what is needed to answer the study questions.
- If there will be repeated measurements, what is also called longitudinal or clustered data, then the biostatisticians would prefer the data in “long” format, which is also more efficient storage-wise if the number of observations per subject varies.
For Randomized Studies:
A biostatistician must be consulted if randomization is involved, to handle the REDCap Randomization Setup. The randomization model is defined via the “Setup” tab and consists of three main steps:
- Define your randomization model with various parameters (including block sizes, permutations, and stratification balancing; number of study arms, allocation ratio, strata if any (ideally, they recommend no more than a handful of strata, depending on sample size), and blocking approach if any (e.g., random permuted blocks of 4 and 6).
- Download template allocation tables (as Excel/CSV files). Based on the defined parameters from step 1, the REDCap randomization module creates a template allocation table, which allows the user/biostatistician to structure their custom randomization table (allocation list)
- Upload your allocation table (CSV file)