Design a Pre-Registration Study Protocol with Statistical Rules

Create a rigorous, pre-registered study protocol that defines primary endpoints, exclusion criteria, power calculations, and interim analysis stopping rules to prevent bias.

Prompt

Prompt
Act as a Senior Clinical Research Scientist. Your job is to draft a pre-registration study protocol based on the parameters I provide. First, ask me these 4 questions: 1. What is the specific research question or hypothesis? 2. What is the primary outcome measure and how is it quantified? 3. What is the target sample size and the estimated effect size for the power calculation? 4. What are the specific criteria for excluding participants or data points? Once I answer, write the protocol using these rules: 1. Every section must be written in the passive voice common to academic journals. 2. Define the exact statistical stopping rules for interim analysis, including specific p-value thresholds. 3. Include a dedicated section on potential confounding variables and mitigation strategies. 4. Use clear, unambiguous definitions for all variables. Output format: # Study Protocol Title # 1. Research Question and Hypotheses 2. Primary Outcome Measure 3. Statistical Power Analysis 4. Inclusion and Exclusion Criteria 5. Interim Analysis and Stopping Rules 6. Confounding Variable Mitigation. This is not medical or scientific advice; consult with your institutional review board before implementing any study design.

Example output

Example for: a study investigating the effect of a 15-minute mindfulness app intervention on cortisol levels in university students during finals week.

Study Protocol: Impact of Mindfulness Intervention on Cortisol Levels

  1. Research Question and Hypotheses
    The study investigates whether a 15-minute daily mindfulness app intervention reduces salivary cortisol levels in undergraduate students during final examination periods. It is hypothesized that the intervention group will demonstrate a statistically significant decrease in mean cortisol compared to the control group.
  2. Primary Outcome Measure
    The primary outcome is the change in salivary cortisol concentration (nmol/L) measured at baseline (day 1) and post-intervention (day 14). Samples are collected between 08:00 and 09:00 AM to control for circadian rhythm.
  3. Statistical Power Analysis
    Power analysis assumes an effect size of Cohen d = 0.5. To achieve 80% power at an alpha of 0.05, a total sample size of 128 participants (64 per group) is required, accounting for a 15% attrition rate.
  4. Inclusion and Exclusion Criteria
    Participants must be enrolled full-time undergraduates. Exclusion criteria include: current diagnosis of anxiety disorders, use of corticosteroid medication, or habitual meditation practice exceeding 10 minutes per week.
  5. Interim Analysis and Stopping Rules
    An interim analysis will be conducted after 64 participants complete the study. A Haybittle-Peto boundary will be applied; the trial will be stopped for efficacy only if the p-value is less than 0.001 to maintain strict control over type I error.
  6. Confounding Variable Mitigation
    Caffeine intake and sleep duration will be recorded via daily logs. These variables will be included as covariates in an ANCOVA model to adjust for potential baseline imbalances between groups.
Written by AI from this prompt. Your results will vary.

Tips and common edits

Why this works

  • Forces the model to adopt a formal, objective academic tone.
  • Requires explicit statistical boundaries to prevent p-hacking.
  • Structures the protocol to meet high standards for study pre-registration.

Use this prompt to draft rigorous, bias-resistant research protocols for academic or clinical study registration. It works best in ChatGPT, Claude, or Gemini. Have your core hypothesis and data collection plan ready to share when prompted.

Common edits

  • Add a section for data storage and ethical compliance if required by your university.
  • Adjust the alpha threshold if your research requires a more conservative p-value.

Make it repeatable

Keep using this prompt without pasting it in every time.

Keep exploring

More in AI Research