Reproducibility in Practice: Data, Code and What Another Lab Can Rerun · seed 1 · A4, ink-friendly. The answer key prints on its own page for grown-ups.

Hand over work that reruns

Science · Scientific Inquiry · ages 23-24
Name ______________________   Date ____________

  1. Here, reproduction means rerunning the original workflow on its original data. Replication means collecting new data for the same question. Which statement uses these terms consistently?

    • Reproduction means independent reanalysis; replication means rerunning the unchanged original script
    • Reproduction collects new observations; replication reruns the original data and workflow
    • Reproduction uses the same-data workflow; replication collects new data

  2. Which handover best supports rerunning a figure?

    • Identified inputs, workflow, environment and output checks
    • Final plotted values and a plotting command, without the cleaning steps that produced those values
    • The current script and a package list, without identifying which input-data revision produced the figure

  3. Supporting computational reproduction requires making identifiable participant data publicly downloadable.

    Circle one:   True   False

  4. A researcher regenerates the expected figure with the original data and workflow. What has this demonstrated?

    • The agreement validates the model assumptions used to interpret the scientific finding
    • The agreement establishes that the finding will recur when new observations are collected
    • The workflow reproduces that output under the checked conditions

  5. An analysis uses randomness and its rerun gives slightly different values. What is a useful first check?

    • Edit the output until it matches, without examining the seed or numerical settings
    • Compare seeds, settings and the stated tolerance
    • Require exact file equality solely because the top-level language version was recorded

  6. Raw sequencing reads may be publicly released. Which deposition plan fits their data type and documentation needs?

    • Use a suitable sequence archive and its required metadata
    • Deposit only processed summary tables and describe them as the complete raw-read dataset
    • Deposit the read files without sample metadata because sequence files alone define every sample

  7. A success_rate value of 0.2 means 20%, but another lab reads it as 0.2%. The code runs. Which documentation would address this misunderstanding?

    • A data dictionary defining the column's scale
    • An environment file specifying the language and numerical-library versions
    • A seed file specifying the random state used to generate the final plot

  8. A README identifies the analysis code only as the main branch. The branch has moved since the figure was made. What information is needed?

    • The latest main branch, because using the same branch name identifies the historical code
    • The immutable code version used for that figure
    • Only the figure's creation date, because dates uniquely identify repository snapshots
LightMySky · lightmysky.comW1-mt_QuHuEyasFt-s1

Answer key

For grown-ups. Fold this page away before handing over the rest.

Hand over work that reruns W1-mt_QuHuEyasFt-s1

  1. Reproduction uses the same-data workflow; replication collects new data · Under the convention stated in the question, reproduction reruns the original same-data workflow and replication collects new data for the same question.
  2. Identified inputs, workflow, environment and output checks · The package must identify the inputs and explain the full workflow and its checks.
  3. False · Restrictions may require authorised access rather than public release of identifiable data.
  4. The workflow reproduces that output under the checked conditions · Matching computational output does not validate every assumption or broader scientific claim.
  5. Compare seeds, settings and the stated tolerance · Randomness and numerical conditions can affect output. Evaluate the difference against the documented comparison.
  6. Use a suitable sequence archive and its required metadata · A suitable archive accepts the raw data and specifies metadata needed to interpret the samples and measurements.
  7. A data dictionary defining the column's scale · The misunderstanding concerns input meaning: the column stores fractions, not percentage-point values.
  8. The immutable code version used for that figure · A moving branch name does not identify an old snapshot. Record the immutable version or commit used.
Worksheet · LightMySky