Reproducibility is the ability to obtain consistent results by repeating an experiment under identical conditions (by the same operator) or similar conditions (by other operators, with other instruments). It is a fundamental criterion of scientific validity.
Discover FizziQ
How to measure it in class
With the FizziQ app, reproducibility can be tested and quantified.
Steps:
- Perform an experiment and note the result
- Repeat the experiment several times under the same conditions
- Calculate the mean and the standard deviation of the results
- Have another group repeat the experiment with the same protocol
- Compare the results of the different groups
- Discuss the sources of variability
Scientific activities on this topic
Learn more
Repeatability vs reproducibility:
| Concept | Conditions | Operator |
|---|---|---|
| Repeatability | Identical | The same |
| Reproducibility | Similar | Different |
Factors affecting reproducibility:
- Precision of the protocol
- Quality of the instruments
- Competence of the operators
- Experimental conditions (temperature, pressure…)
- Natural variability of the phenomenon
Scientific importance:
A non-reproducible result is not considered established. The “reproducibility crisis” in certain scientific fields underscores the importance of this criterion.
Reproducibility indicators:
- Standard deviation between repetitions (repeatability)
- Deviation between laboratories (reproducibility)
- Coefficient of variation (CV = σ/μ × 100%)
Formula
Coefficient of variation: CV = (σ/μ) × 100%
where:
- σ: standard deviation of the measurements
- μ: mean of the measurements
Interpretation:
- CV < 5%: excellent reproducibility
- CV < 10%: good reproducibility
- CV > 20%: poor reproducibility
Compatibility test (two results): |x₁ - x₂| < 2 × √(u₁² + u₂²) → compatible results
Application examples
- Two groups measure g = 9.78 ± 0.10 and g = 9.83 ± 0.12 m/s²: reproducible results
- An isolated “extraordinary” measurement must be reproduced before being accepted
- Lab work protocols are tested for their reproducibility
- ISO standards require reproducibility studies
FAQ
Q: How can reproducibility be improved? A: Specify the protocol, standardize the conditions, train the operators, use calibrated instruments, repeat the measurements.
Q: Is a non-reproducible result wrong? A: Not necessarily, but it is not reliable. There may be an uncontrolled parameter influencing the result.
Q: Does FizziQ improve reproducibility? A: Yes, because the measurements are automated (fewer reading errors), timestamped and exportable. The sensor is standardized.
Q: Should reproducibility always be verified? A: In research, yes. In educational lab work, it is an excellent inter-group comparison exercise.
Related concepts
Measurement Uncertainty - Experimental Protocol - Standard Deviation