Nova Academy · Foundations of Inquiry
Lesson 032 — Research Documentation: Keeping a Record Another Person Can Test
A result is easier to evaluate when its methods, observations, calculations, assumptions and changes are recorded clearly. Documentation does not guarantee that a claim is true; it allows errors and successes to be examined.
Learning objectives
- Distinguish raw observations from interpretation and conclusions.
- Write a dated, reproducible procedure with units, instrument details and conditions.
- Maintain a transparent record of corrections, versions and uncertainty.
- Recalculate a documented result and identify missing information.
What belongs in a research record?
Record a precise question, date and time, location or controlled setting, materials and instruments, calibration and units, protocol, raw measurements, anomalies, calculations, code or spreadsheet versions, uncertainty and limitations. Preserve original readings even if a later correction is needed. Mark corrections with a date and explanation rather than silently replacing data.
Worked example: a pendulum timing record
Question: What is the average time for ten complete swings of a small pendulum? The observer defines one swing as a return to the same position and direction, uses a stopwatch reading to 0.01 s, and records three independent trials under similar conditions. Raw times for ten swings are 20.1 s, 20.3 s and 20.2 s. Stopwatch resolution alone does not describe reaction-time uncertainty.
Mean time for ten swings = (20.1 + 20.3 + 20.2) / 3 = 20.2 s. Estimated mean period per swing = 20.2 / 10 = 2.02 s. The sample range is 20.3 − 20.1 = 0.2 s per ten swings, or 0.02 s per swing. This range describes observed spread, not a formal confidence interval or a complete uncertainty budget.
A useful record includes the pendulum length and how it was measured, amplitude, timing method, raw trials, and whether the same observer started and stopped the stopwatch. Without those details, another student can repeat a similar experiment but not precisely reproduce the procedure.
Activity: create a reproducible notebook entry
- Choose a safe measurement such as the time for a paper pendulum to make ten swings, or the length of a tabletop object.
- Write a question and a step-by-step method another person could follow without asking you for clarification.
- Take three readings; preserve the original numbers and units.
- Calculate the mean. For example, 12.0, 12.2 and 12.1 cm give (12.0 + 12.2 + 12.1) / 3 = 12.1 cm.
- Ask another person to repeat your method, then compare readings and note any ambiguities.
- Add a dated revision describing what you would change and why. Do not erase the first version.
Reflection
Which details seemed obvious to you but were unclear to another observer? Could your measurement change because of technique, instrument or environmental conditions? What evidence would persuade you to revise your conclusion?
Knowledge check
- 1. Why must raw observations be retained when an interpretation changes?
- 2. A notebook lists 4.0, 4.2 and 4.1 cm. What is the mean?
- 3. Is a range of repeated readings the same as a complete uncertainty estimate?
- 4. What should happen when a transcription error is discovered?
Answer key
- 1. So the analysis remains auditable and others can check alternative explanations.
- 2. (4.0 + 4.2 + 4.1) / 3 = 4.1 cm.
- 3. No. It captures observed spread but not necessarily calibration bias, resolution, or other systematic effects.
- 4. Preserve or trace the original entry, record the corrected value, date and reason, and update any affected calculation.
Limitations and ethics
Clear records do not automatically make a sample representative, remove bias, prove causation or validate an instrument. A second observer may still obtain different results. Avoid including private identifying information in shared research notebooks. Record uncertainty honestly; do not invent or discard inconvenient observations.
Sources and further reading
- National Academies, Reproducibility and Replicability in Science (2019) — terminology and research transparency.
- NIST Technical Note 1297, Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results — measurement uncertainty.
- NIST Research Data Management — practices for stewardship and documenting research data.
Previous: Lesson 031 — Reproducibility and Replication · Nova Academy Index
