Complete AI Training
Sign inGet started

AI agent for research scientists

Statistical Power and Sample Size Replanning Agent

Choose a sample size and design that give adequate power within the available limits

DecisionsPlanningCreatingScience and research
See how it worksGet access
Statistical Power and Sample Size Replanning Agent: what goes in, what the agent does and what you get

What it does

A researcher plans 6 animals per group because that is what the lab always uses, and after months the study cannot detect the effect. Or they plan 20 when 10 would do and spend animals and reagent money. This agent plans the sample. It takes the pilot variance and the effect size of interest, calculates the sample needed, tests how sensitive the answer is to assumptions and checks against animal, reagent and cost limits. If the limits are exceeded, it proposes design changes such as a paired design or fewer groups, and recomputes the power. The PI approves the design. Edge case: the pilot has only 3 replicates. The agent warns that the variance estimate is uncertain and uses a conservative value.

How it works

Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.

Start and resultWhat it doesA check on its own workWaits for your OKGoes back and retries
Yes, continueYes, continueApprovedNoNo 1 STARTS WHEN Experiment design drafted 2 USES A TOOL Load pilot data, effect size and the design 3 CHECKS THE RESULT Is the variance estimate based on enough replicates? If not: Use a conservative variance estimate and flagthe uncertainty. Back to step 2. 4 DOES Calculate the sample size for the target power 5 DOES Test sensitivity: higher variance and smaller effect 6 CHECKS THE RESULT Does the needed sample fit within the animal,reagent and cost limits? If not: Propose design changes such as paired design orfewer groups and recompute. Back to step 4. 7 DOES Write the design summary with assumptions and thepower curve 8 YOU APPROVE PI approves the design 9 RESULT Sample size plan and power analysis
Read the steps as a list
  1. Experiment design drafted
  2. Load pilot data, effect size and the design
  3. Is the variance estimate based on enough replicates?If not: Use a conservative variance estimate and flag the uncertainty. Back to step 2.
  4. Calculate the sample size for the target power
  5. Test sensitivity: higher variance and smaller effect
  6. Does the needed sample fit within the animal, reagent and cost limits?If not: Propose design changes such as paired design or fewer groups and recompute. Back to step 4.
  7. Write the design summary with assumptions and the power curve
  8. PI approves the designThe agent waits here for your OK.
  9. Sample size plan and power analysis

How it decides

It computes the sample size for the target power, and repeats with a higher variance and smaller effect. If the needed size exceeds limits, it tests design changes that raise power.

  • Target power is 80 percent at 0.05 significance unless the PI sets another
  • Pilot variance from fewer than 5 replicates is inflated by 20 percent
  • A plan that falls below 70 percent power under pessimistic assumptions is flagged
  • Limits are checked for animals, reagents and cost

Make it yours

Every agent is a starting point. You choose these settings for your own situation.

  • Target power and significance level
  • Variance inflation for small pilots (default 20 percent)
  • Resource limits
  • Designs to consider
  • Report format

What keeps you in control

It always asks you first

  • PI approves the final design and sample size

Hard limits

  • Never changes the research question
  • States all assumptions in the plan

It stops when

  • Done: design approved with power analysis
  • Stop: no feasible design reaches the target power, discuss feasibility with the PI

Set it up

We guide you through the set-up, step by step

Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.

10 minto set it up in your AI
5 AIsChatGPT, Claude, Copilot, Gemini, Grok
  • One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
  • The agent then walks you through connecting your own data, one source at a time
  • A downloadable copy with the flow chart, the rules and the full guide
Get access to this agent

An example run

What happensA pilot with 4 replicates shows a standard deviation of 12 units and the effect of interest is 10. The agent inflates the variance by 20 percent and finds 16 per group for 80 percent power. The lab limit is 12, so it proposes a paired design, which needs 9 per group. Sensitivity shows 74 percent power at the pessimistic case. The PI approves 10.

More agents for research scientists

Research · Document handling

Research Literature Update Agent

An up-to-date evidence map with every new study screened.

7 steps1 check11 jobs Quality control · Science and research

Scientific Unit Consistency Agent

An analysis working copy in which every calculation uses consistent, documented units.

7 steps2 checks13 jobs Research · Analysis

Research Method Comparison Agent

A fair, source-linked methods comparison.

7 steps1 check8 jobs Quality control · Decisions

Research Analysis Assumption Agent

Every analysis has its assumptions checked, with violations decided by the investigator.

8 steps1 check11 jobs