1.1.4(a)-(e) - Evaluation, uncertainty and method improvement
This lesson is about judging the quality of practical evidence. You will learn how to move from results to a supported conclusion, how to spot anomalies and limitations, how precision, accuracy and uncertainty affect confidence, and how to suggest method improvements that are specific enough to earn credit. These skills apply across practical contexts, so the examples use simple biological investigations without turning this into a lesson on one named PAG.
Evidence And Conclusions
A result is a measurement or observation. A conclusion is the answer you draw from the results. Evaluation is the judgement of how strong that conclusion is, based on the quality of the method, the pattern in the data and the uncertainty in the measurements.
Conclusion
A statement that answers the experimental question using the results as evidence.
In a practical answer, keep three ideas separate:
| Skill | What it sounds like | What it does |
|---|---|---|
| Describe the results | "The mean rate increased from 12 to 19 arbitrary units." | Reports the evidence. |
| Draw a conclusion | "Increasing the independent variable increased the rate." | Answers the question. |
| Evaluate the conclusion | "This conclusion is limited because only two values were tested and the ranges overlap." | Judges confidence in the conclusion. |
Suppose a student investigates whether a higher substrate concentration increases the rate of a reaction. The mean rates are:
| Substrate concentration / mol dm-3 | Mean rate / arbitrary units |
|---|---|
| 0.2 | 12 |
| 0.4 | 19 |
A weak conclusion is: "The result was higher at 0.4 mol dm-3." That only describes the data. A stronger conclusion is: "Increasing substrate concentration from 0.2 to 0.4 mol dm-3 increased the mean reaction rate from 12 to 19 arbitrary units, so these results support the conclusion that substrate concentration affected rate over this range." It uses the data and keeps the conclusion inside the evidence.
The final phrase matters. The investigation did not test every possible concentration, every enzyme, every temperature or every pH. In evaluation, you should avoid words such as "proves" unless the evidence really removes the main alternatives. In biology practicals, conclusions are usually supported to a degree, not proven absolutely.
For Evaluate or Use the data questions, pair a conclusion with evidence and a limitation. A conclusion without data is often too vague; data without a judgement is only description.
Now apply that structure to a small data set where the conclusion is possible but not unlimited.
Anomalies And Limitations
An anomaly is a result that does not fit the pattern of the other results. In experimental measurements, an anomaly may be caused by a one-off procedural error, a recording error, a damaged sample or genuine biological variation. The skill is not just to name the odd value; it is to decide what to do with it.
Anomaly
A measurement that is inconsistent with the pattern or with the other repeat measurements, so it needs checking before it is used in a conclusion.
For example, a student measures the time for a colour change in four repeats:
| Repeat | Time / s |
|---|---|
| 1 | 26 |
| 2 | 25 |
| 3 | 52 |
| 4 | 24 |
The value 52 s is anomalous because it is much higher than the other three values, which cluster between 24 s and 26 s. A good evaluation would say that the repeat should be checked or repeated. It should not simply be deleted without a reason. If the student knows that the timer was stopped late for repeat 3, excluding it from the mean is justified. If there is no known reason, the safer improvement is to repeat the measurement and compare the new result with the pattern.
A limitation is a weakness in the procedure that affects the quality of the evidence. It may affect validity, precision or accuracy.
Limitation
A feature of an experimental procedure that reduces confidence in the results or the conclusion.
Common limitations in biology practicals include:
| Limitation | Likely effect on results | Better evaluation language |
|---|---|---|
| Temperature not controlled | Enzyme or membrane behaviour may change for reasons other than the independent variable | "This reduces validity because temperature is a confounding variable." |
| Endpoint judged by eye | Different observers may stop timing at different points | "This reduces precision because repeated timings may vary." |
| Small change measured with coarse apparatus | Percentage uncertainty is large | "This reduces confidence because the uncertainty is large compared with the measured change." |
| Too few repeats | Anomalies and random variation are harder to identify | "This reduces confidence in the mean because anomalies and random variation are harder to detect." |
Precision, Accuracy And Uncertainty
Precision and accuracy are not the same thing.
Precision
The closeness of repeated measurements to each other.
Precise results have a small spread. For example, lengths of 31.0 mm, 31.1 mm and 31.0 mm are precise because the repeats are close together. Precision is mainly affected by random error, biological variation and the resolution of the apparatus.
Accuracy
The closeness of a measurement to the true or accepted value.
Accurate results are close to the true value. A set of results can be precise but inaccurate. For example, if a balance is not zeroed, every mass measurement may be close to the others but all may be too high. That is why repeated close results do not guarantee a valid conclusion.
Uncertainty is the doubt attached to a measurement. It is often written as a margin of error, such as 25.0 +/- 0.5 mm. This means the measured value is treated as lying in a range around 25.0 mm, based on the measuring apparatus or method.
Uncertainty
An estimate of the range within which the true value of a measurement is likely to lie.
In this course, apparatus uncertainty is usually based on the measuring instrument:
| Apparatus situation | Sensible uncertainty rule |
|---|---|
| Analogue scale, such as a ruler or measuring cylinder | Often half the smallest division, unless the question states otherwise |
| Digital apparatus, such as a balance | Usually +/- the resolution shown by the display |
| Timing by human reaction with a stopwatch | Human reaction time may be a bigger uncertainty than the stopwatch resolution |
| Measurement by difference, such as mass lost | Include uncertainty from both readings |
Different sources use different conventions for the absolute uncertainty of measuring instruments. If a question states the absolute uncertainty, use the value given. If it does not, state your assumption clearly in your working.
Percentage Uncertainty
Absolute uncertainty tells you the size of the margin of error in the original unit. Percentage uncertainty tells you how large that uncertainty is compared with the quantity measured. This is useful because the same absolute uncertainty can be small for a large measurement but large for a small measurement.
You may also see this idea described as percentage error where the error comes from uncertainty in measurement. In this lesson, percentage uncertainty is the calculation that shows the percentage size of that measurement uncertainty.
Percentage uncertainty for one measured quantity
For example, measuring 20.0 cm3 with an uncertainty of +/- 0.5 cm3 gives:
0.5 / 20.0 x 100 = 2.5%
The same apparatus used to measure only 2.0 cm3 gives:
0.5 / 2.0 x 100 = 25%
The apparatus has the same absolute uncertainty, but the smaller measured volume has a much larger percentage uncertainty. That is why apparatus choice must match the size of the measurement.
When a quantity is measured by difference, there is uncertainty in both readings. A mass lost, temperature change or change in volume often works like this.
Percentage uncertainty for a quantity measured by difference
This worked example shows the extra step that is needed when the measured change comes from two readings.
Mass Loss Measured By Difference
A potometer has a mass of 23.45 g at the start and 23.21 g at the end. The balance has an uncertainty of +/- 0.01 g for each reading.
- Calculate the mass lost:
23.45 - 23.21 = 0.24 g
- Combine the uncertainty from both readings:
2 x 0.01 = 0.02 g
- Calculate the percentage uncertainty:
0.02 / 0.24 x 100 = 8.3%
The result should be reported as a mass loss of 0.24 g with an approximate percentage uncertainty of 8.3%.
The biological interpretation is the important final step. A percentage uncertainty of 8.3% may be acceptable if the measured difference between treatments is much larger than this, but it would weaken the conclusion if the difference between treatments is only a few percent.
Method Improvement
A strong method improvement is targeted. It names the change, links it to a limitation and explains how it improves the quality of the evidence. Vague statements such as "be more careful" or "use better equipment" are usually not enough.
Use this chain:
limitation -> effect on evidence -> improvement -> why the improvement helps
| Limitation | Weak improvement | Strong improvement |
|---|---|---|
| Endpoint judged by eye | "Measure it better." | "Use a colorimeter to measure absorbance at fixed time intervals, reducing observer judgement and improving precision." |
| Temperature not controlled | "Keep it fair." | "Use a thermostatically controlled water bath so temperature is controlled and the independent variable is the main factor affecting the result." |
| Volume measured with a large measuring cylinder | "Use better apparatus." | "Use a graduated pipette or syringe with a smaller absolute uncertainty for the required volume, reducing percentage uncertainty." |
| One anomalous repeat | "Ignore it." | "Repeat that measurement and compare it with the other repeats before deciding whether to exclude it from the mean." |
| Small measured change | "Do it again." | "Increase the duration or scale of the measurement, if biologically appropriate, so the measured change is larger compared with the apparatus uncertainty." |
| Too few samples or repeats | "Use more." | "Use more samples or repeats and calculate a mean, reducing the effect of random variation and making anomalies easier to identify." |
Notice that an improvement can target different parts of evidence quality:
- Improve precision by reducing random variation, taking more repeats, using a clearer endpoint or using apparatus with smaller uncertainty.
- Improve accuracy by calibrating or zeroing apparatus and reducing systematic error.
- Improve validity by controlling a confounding variable or making sure the method measures the intended dependent variable.
- Improve confidence in the conclusion by checking anomalies and reporting uncertainty rather than hiding inconvenient data.
In method-improvement questions, write an improvement and an explanation. "Use a water bath" is weaker than "Use a water bath to keep temperature constant, so temperature does not affect the dependent variable."
The next check rewards the change and the reason for the change separately.
Method Improvement Continued
That same improvement logic applies whether the context is enzymes, membranes, transport, photosynthesis or any other practical setting.
Evaluation is not a complaint list. It is a judgement about how far the evidence supports the conclusion, using anomalies, limitations, precision, accuracy, uncertainty and specific method improvements.