Skip to main content

Table 2 Framework for pre-specifying a statistical analysis strategy (Pre-SPEC)

From: How to design a pre-specified statistical analysis approach to limit p-hacking in clinical trials: the Pre-SPEC framework

Pre-specify before recruitment Pre-specify the analysis strategy before recruitment to the trial begins.
Single analysis strategy Specify a single primary analysis strategy.
Plan each aspect Each aspect of the planned analysis should be covered, including analysis population, statistical model, covariates, and handling of missing data.
Enough detail Provide sufficient detail to allow a third party to independently perform the analysis (ideally through statistical code).
Choices made deterministically For adaptive analysis strategies which use the trial data to inform some aspect of the analysis, use deterministic decision rules that prevent analysis choices being driven by results.