Lesson 8 of 8 / Practical and data skills
Turn a limitation into a useful improvement
What does this result support, and what would make the evidence stronger?
In this lesson: Link specific errors and limitations to their effects and to targeted improvements.
About 6 min
The key ideaA useful evaluation names the limitation, explains its effect and proposes an improvement that addresses that effect.
Work with the evidence
Match the correction to the problem
Effect: All measured masses are biased upward by the same amount. Repeats preserve the offset. Improvement: Check and correct zero/calibration before weighing. This targets the bias.
Explanation
Begin with the result in context. A description states a pattern using data; an explanation connects it to a biological mechanism. A conclusion answers the original question at the strength the design allows. Correlation alone cannot rule out other variables or establish the direction of causation.
Random variation makes repeated values spread out. Additional independent repeats and an appropriate summary can improve the estimate and show the spread. A systematic error shifts results in a consistent way, such as a balance that reads too high. More repeats of the same biased method do not remove that offset.
A method can be precise but fail to measure the intended process. Counting bubbles estimates photosynthetic gas production imperfectly because bubbles vary in volume and some gas dissolves. Measuring collected gas volume over time addresses the unequal-bubble-size limitation, while other limitations still need consideration.
Sampling also limits interpretation. Ten leaves from one plant are not ten independent plants. If the question concerns a population of plants, sample multiple suitable individuals using a defensible selection method and account for the grouping of observations. A large convenience sample can remain biased.
Make the improvement match the problem. If temperature drifts, use and monitor a controlled water bath. If the tested range misses a peak, add values near the suspected peak. If an observation is anomalous, check records and repeat the relevant conditions rather than silently deleting it. Explain which uncertainty each improvement reduces and which claims remain unsupported.
Step by step
- 1
Name a specific problem
Identify the affected measurement, comparison or sample.
- 2
Explain the consequence
State the direction of bias if justified, or the uncertainty it introduces.
- 3
Choose a matching correction
Propose an action that changes the source of the problem, not just a generic request for more data.
Worked example
Work through the evidence
An experiment compares photosynthesis at different light intensities by counting bubbles. Explain one limitation and a targeted improvement.
One way to explain it
Bubbles may have different volumes, so equal bubble counts need not represent equal gas production. Collect gas in suitable calibrated apparatus and measure volume per unit time under otherwise controlled conditions. This addresses bubble-size variation, while dissolved gas and other limitations may remain.
Why this answer works
- Identify the specific proxy measurement.
- Connect the limitation to the inferred rate.
- Explain why the proposed improvement helps without claiming perfection.
Is this true? "Writing "human error" explains why an experiment is unreliable."
Name the action or measurement problem. For example, a delayed start of timing changes the measured interval; a consistent zero offset creates bias. Different problems require different improvements.