Can I use this for non-experimental data?

While optimized for experimental designs, it provides robust tools for causal diagramming (DAGs) and identification strategy verification applicable to observational studies.

What is a formal estimand in this context?

An estimand is the specific statistical quantity (such as an Average Treatment Effect) that, when calculated, directly tests your theoretical hypothesis.

How does this skill help with research pre-registration?

It ensures your hypotheses are specified at conceptual, operational, and statistical levels, providing the precise detail required for high-quality pre-analysis plans (PAPs).

What is SESOI and why is it required?

SESOI stands for Smallest Effect Size of Interest. Defining it prevents the pursuit of statistically significant but practically or theoretically meaningless results.

Causal Hypothesis Architect

Name: Causal Hypothesis Architect
Author: scdenney

byscdenney

•

Data Science & ML

Transforms theoretical concepts into rigorous, falsifiable hypotheses and formal statistical estimands for experimental and social science research.

This skill provides a structured framework for researchers and data scientists to bridge the gap between abstract theory and empirical testing. It guides users through resolving the Fundamental Problem of Causal Inference (FPCI), mapping causal diagrams (DAGs), defining counterfactual logic, and specifying formal estimands to ensure research designs are pre-registrable and methodologically sound. By enforcing standards like the Smallest Effect Size of Interest (SESOI) and three-level specification (conceptual, operational, and statistical), it helps eliminate ambiguity in experimental design and hypothesis testing logic.

Key Features

01Drafts three-level hypothesis specifications: conceptual, operationalized, and statistical.

02Provides guidance on selecting between NHST, equivalence, and minimum-effect tests.

0315 GitHub stars

04Maps causal diagrams (DAGs) to identify backdoor paths and isolation strategies.

05Verifies experimental prerequisites including SUTVA, exclusion restrictions, and random assignment.

06Specifies formal theoretical and empirical estimands for pre-analysis plans.

Use Cases

01Justifying the Smallest Effect Size of Interest (SESOI) for statistical power analysis.

02Drafting a pre-analysis plan (PAP) for a social science experiment or survey.

03Designing a multi-experiment hypothesis architecture for complex causal chains.

Key Features

01Drafts three-level hypothesis specifications: conceptual, operationalized, and statistical.

02Provides guidance on selecting between NHST, equivalence, and minimum-effect tests.

0315 GitHub stars

04Maps causal diagrams (DAGs) to identify backdoor paths and isolation strategies.

05Verifies experimental prerequisites including SUTVA, exclusion restrictions, and random assignment.

06Specifies formal theoretical and empirical estimands for pre-analysis plans.

Use Cases

01Justifying the Smallest Effect Size of Interest (SESOI) for statistical power analysis.

02Drafting a pre-analysis plan (PAP) for a social science experiment or survey.

03Designing a multi-experiment hypothesis architecture for complex causal chains.