don't
procrastinate.
promptcrastinate.

Plan a Home Experiment to Settle an Argument With Actual Data

<context>
You are a practical experiment designer who turns household disagreements into small, genuinely testable experiments rather than debates. {DISPUTE} is the disagreement the user wants settled, such as whether a plant does better in one spot than another, whether a particular cleaning method actually works, or whether one route really is faster.
</context>

<task>
**Plan a simple home experiment to settle {DISPUTE} with actual data.**
1. Turn {DISPUTE} into a specific, testable question, defining clearly what would count as one side being right.
2. Design the simplest experiment that could answer it, specifying what to measure, how often, and for how long.
3. Identify the most obvious way the experiment could be accidentally biased, an unfair comparison, too small a sample, an outside factor being ignored, and how to control for it cheaply.
4. Specify what result would actually count as a clear answer, versus a result too close to call, so the argument cannot just restart over ambiguous data.
</task>

<output_format>
- **The testable question**: one sentence, rephrasing {DISPUTE} as something measurable
- **The experiment**: what to measure, how often, over what period
- **Avoiding bias**: the main risk and how to control for it
- **What counts as an answer**: a short paragraph defining a clear win versus an inconclusive result
- Length: around 250 to 350 words
</output_format>