How does reproducing differ from replicating?

• Reproducing repeats the experiment fresh
• Replicating re-runs your script only
• Reproducing re-runs analysis; replicating repeats the experiment

What does reproducibility require of a stranger?

• Rebuilding every figure without asking you
• Believing every claim without checking
• Rewriting your code from scratch

A reproducible package pins data, code, and tool versions plus the run command.

Circle one:   True   False

Rerunning your script on your data confirms what?

• The phenomenon in nature
• Your analysis and claims
• Every future experiment

A stranger reruns your package and the figure differs. What do you suspect first?

• Fraud by the stranger
• A law of nature changed
• An unpinned version or a missing step

Where do raw DNA sequences belong?

• A sequence archive with metadata
• A screenshot in a slide deck
• An inbox attachment alone

What must your README contain so figures rebuild years later?

• The exact command plus pinned versions
• Your biography and funding history
• Opinions about rival labs

A lab re-runs your script on your data and declares the phenomenon confirmed. What is wrong?

• Scripts can never be rerun
• Same-data reruns check analysis, not nature
• Phenomena need no confirmation