OCR A Level Biology: Development of Practical Skills — Revision Notes

Condensed for the final weeks. For the full explanation, use the Development of Practical Skills study guide.

Variables

Type	Meaning
Independent	The one you change
Dependent	The one you measure
Control	Kept constant so they do not affect the result

A control experiment is different from a control variable: it is a parallel run with the independent variable absent, showing that the effect observed is caused by the factor being tested and nothing else. Boiled enzyme in place of active enzyme is the standard example.

The five words that are constantly confused

Term	Meaning
Accurate	Close to the true value
Precise	Repeat readings are close to each other
Repeatable	The same person, using the same method and apparatus, gets the same result on repeating the measurement
Reproducible	The same result is obtained when the operator, apparatus or method changes
Resolution	The smallest change a measuring instrument can detect — a property of the instrument, not of the readings taken with it

Separately, valid means measuring what it is supposed to measure — it requires a fair test with controlled variables, and is a different kind of quality from the five above.

A set of results can be precise but not accurate — a wrongly calibrated balance gives consistent readings that are consistently wrong. That distinction is the most commonly examined idea in the whole topic.

Increasing repeats improves reliability and lets you spot anomalies, and it improves the accuracy of the mean by reducing the effect of random error (repeating and averaging cancels out random scatter around the true value). It does nothing for a systematic error — a consistent bias caused by, for example, a wrongly calibrated instrument — which needs recalibration, not more repeats.

Uncertainty

uncertainty of a single reading = half the smallest scale division
percentage uncertainty = (uncertainty / measured value) x 100

For a measurement requiring two readings — a burette, a thermometer measuring a temperature change, a ruler measuring a length between two marks — the uncertainty is doubled, because both readings carry uncertainty.

To reduce percentage uncertainty, measure a larger quantity. The absolute uncertainty stays the same, so it becomes a smaller proportion. That is why you time ten swings rather than one, or weigh a larger mass.

Errors

• Random error — scatter around the true value, from unpredictable variation. Reduced by repeating and averaging.
• Systematic error — every reading wrong by the same amount or proportion, from faulty apparatus or method. Repeating does not help; the instrument must be recalibrated or zeroed.
• Zero error — a specific systematic error where the instrument does not read zero when it should.

If a graph line should pass through the origin but has an intercept, that indicates a systematic error.

Graphs

• Independent variable on the x-axis, dependent on the y.
• Scales must be linear and use more than half the available grid.
• Axes labelled with quantity and unit.
• Draw a line of best fit, not dot-to-dot; identify anomalies and exclude them from the line.
• Draw a curve if the data curve — forcing a straight line loses marks.

Error bars show the range or uncertainty of the data. Where error bars for two conditions overlap, the difference between them may not be significant.

Improving an experiment — what actually scores

Generic answers such as “be more careful” or “repeat it” score nothing on their own. Marks come from naming a specific improvement and its reason:

• Use a water bath to keep temperature constant, because enzyme activity changes with temperature.
• Use a colorimeter rather than judging colour by eye, because visual judgement is subjective.
• Use a buffer solution, because pH changes would alter the enzyme’s active site.
• Use apparatus with a smaller scale division, because it reduces percentage uncertainty.
• Randomise sample positions, to remove bias.

Safety and ethics

Risk assessments must state the hazard, the risk, and the control measure — three separate things. “Wear goggles” alone is not a risk assessment.

Microscopy and magnification

Calibrating an eyepiece graticule against a stage micrometer converts arbitrary graticule divisions into real lengths at each objective magnification used.

magnification = image size / actual size

Biological drawings should use clear, continuous lines with no shading, be drawn to a stated magnification or scale, use label lines that do not cross, and show only what is actually visible in the specimen — not what is expected to be there.

Quantitative techniques

Serial dilution produces a known concentration range from a single stock solution — a tenfold series is made by transferring 1 cm³ into 9 cm³ of solvent at each step.

Colorimetry measures absorbance or transmission; a calibration curve made from solutions of known concentration allows an unknown concentration to be read off.

Rate is usually calculated as 1/time, where time is measured to a fixed end point, or from the initial gradient of a tangent to a curve when the reaction slows over the course of the experiment — the initial rate is used because substrate concentration is highest, and least changed, at the very start.

Choosing a statistical test

Choosing the right test is examined as often as performing it:

Purpose	Test
Difference between two means, normally distributed data	t-test
Association between two variables	correlation coefficient
Observed versus expected frequencies	chi-squared

Compare the calculated value with the critical value at p = 0.05. If the calculated value exceeds the critical value, the result is significant and the null hypothesis is rejected — meaning there is less than a 5% probability the difference or association is due to chance.

Exam traps

• Swapping accuracy and precision.
• Claiming more repeats improve accuracy when the error is systematic — repeating only helps with random error; a systematic error needs recalibration.
• Forgetting to double the uncertainty for a two-reading measurement.
• Saying repeating removes systematic error.
• Suggesting improvements without giving reasons.
• Drawing a straight line through curved data.

Self-test

1. Distinguish accuracy from precision, with an example of precise but inaccurate results.
2. How is the uncertainty of a burette reading calculated, and why?
3. What kind of error does repeating and averaging reduce, and what kind does it not?
4. How do you reduce percentage uncertainty?
5. What does it mean when error bars for two treatments overlap?

Answers: 1. Accuracy is closeness to the true value; precision is how close repeat readings are to each other. A balance with a zero error gives consistent readings that are all consistently wrong — precise but inaccurate. 2. Half the smallest scale division, doubled, because a burette measurement requires an initial and a final reading, each carrying uncertainty. 3. It reduces random error; it does not reduce systematic error, which requires recalibration. 4. Measure a larger quantity, so the fixed absolute uncertainty forms a smaller proportion of the reading. 5. The difference between the two treatments may not be statistically significant.