Experimental Design Planner

Design a rigorous, statistically sound experiment with controls, sampling, analysis plan, and ethics safeguards.

// prompt
You are a senior research methodologist with deep expertise in experimental design, statistics, and research ethics. Design a rigorous, reproducible study to investigate **{{research_question}}** in the field of **{{field_or_domain}}**. Work through the design systematically and challenge weak assumptions as you go. ## 1. Frame the Study - Restate the **{{research_question}}** as a testable hypothesis (H1) and a null hypothesis (H0). - List any secondary questions worth exploring: **{{secondary_questions}}**. - State the expected direction of effect and why it matters. ## 2. Define Variables - **Independent variable(s)** to manipulate: {{independent_variable}}. - **Dependent variable(s)** to measure, with units and instruments: {{dependent_variable}}. - **Control variables** to hold constant: {{control_variables}}. - **Likely confounders** and how the design neutralizes each: {{potential_confounders}}. ## 3. Conditions and Sampling - Define the **control group** ({{control_condition}}) and **treatment group(s)** ({{treatment_conditions}}). - Specify the design type (between-subjects, within-subjects, factorial, crossover, etc.) and justify it. - Population and inclusion/exclusion criteria: {{target_population}}. - Recommend a **sample size** via power analysis, stating the assumed effect size, alpha, and power (default α = 0.05, power = 0.80). - Describe **randomization and blinding**: {{randomization_method}}. ## 4. Procedure and Measurement - Step-by-step protocol from recruitment to final measurement. - Data collection methods, timing, and instrument reliability checks. ## 5. Statistical Analysis Plan - Primary test matched to the design and data type; state assumptions and how to verify them. - Secondary analyses, effect-size reporting, and how multiple comparisons are handled. ## 6. Ethics and Threats to Validity - Informed consent, risk assessment, and approvals needed: {{ethical_constraints}}. - List the main threats to internal and external validity and your mitigations. **Deliverable:** A structured design document with the sections above, a one-paragraph rationale for each major choice, and a checklist of decisions I still need to confirm. Flag any assumption that could invalidate the results.
Fill in the variables
Example response

Experimental Design Framework

Research Question

Primary: Does AI-assisted learning improve medical diagnostic skills compared to traditional methods?

Experimental Variables

Independent Variable: Training intervention (AI-assisted vs. control)

Dependent Variable: Diagnostic accuracy score (0-100%)

Control Variables: Student GPA, clinical rotation completed, study time

Experimental Conditions

Control Group (n=50): Traditional case-based learning with textbook resources

Experimental Group (n=50): AI-powered diagnostic assistant with explanation features

Sample Size Justification: Power analysis: α=0.05, β=0.80, effect size d=0.6

Methodology

Participants: 3rd-year medical students from partner university

Randomization: Block randomization stratified by GPA quartiles

Duration: 8-week intervention with pre/post assessment

Statistical Analysis Plan

Primary Analysis: Independent t-test comparing post-intervention scores

Secondary: ANCOVA controlling for baseline knowledge

Effect Size: Cohen's d with 95% confidence intervals

Related prompts

Science & Research

Data Analysis Interpreter

Acts as a statistician to choose, run conceptually, and interpret the right analyses for your research data.

Science & Research

Hypothesis Formation Assistant

Turns observations and theory into clear, testable, falsifiable research hypotheses with defined variables and predictions.

Science & Research

Research Ethics Advisor

Expert guidance on research ethics, IRB approval, and responsible conduct for your study.

Science & Research

Scientific Manuscript Reviewer

Acts as a senior peer reviewer giving structured, criteria-based, constructive feedback and a recommendation on a scientific manuscript.