hypothesis-building

To write a falsifiable hypothesis for preregistration, map theoretical claims to precise estimands like SATE or PATE. This process uses a three-level specification—conceptual, operationalized, and statistical—to reduce ambiguity and explicitly guide your regression model.

SATE (Sample Average Treatment Effect) and PATE (Population Average Treatment Effect) are estimands defining your causal hypothesis scope. Specifying them during preanalysis clarifies whether your statistical inferences target the sampled data or the broader population.

Choosing between NHST and equivalence testing requires defining a SESOI (Smallest Effect Size of Interest). This Skill helps you compare minimum-effect tests with explicit power considerations to establish pre-registered decision rules for your analysis.

Mapping a conceptual hypothesis to a statistical model requires operationalizing the claim into a defined treatment contrast. This Skill formalizes that mapping by specifying backdoor adjustments and linking your theoretical construct directly to the regression specification.

Yes, this Skill supports multi-experiment architectures by designing counter-hypotheses and bridging micro-macro levels of analysis. It applies causal inference frameworks across social science contexts to maintain coherent estimand specifications.

Include backdoor adjustments in your preanalysis plan when confounding variables threaten causal inference. Identifying these adjustments is necessary to formally isolate the treatment contrast and accurately estimate the defined estimand.