Skills

A good question names the independent variable and its range, the dependent variable and how it is measured, and the organism or system. Compare a vague question, such as “Does temperature affect enzymes?”, with a focused one: “How does temperature, from 10 to 60 °C, affect the rate at which catalase from potato breaks down hydrogen peroxide, measured as volume of oxygen in 30 s?”

• Independent variable: the one you change on purpose.
• Dependent variable: the one you measure.
• Controlled variables: everything else that could affect the result, kept the same. Say how you keep each one constant, and why it matters.
• A control is a separate treatment for comparison, such as boiled enzyme or no light. It is not the same thing as a controlled variable.

Watch out: “keep everything else the same” earns no credit. Name the variables.

A hypothesis is a testable explanation, with a reason based on biology. A prediction says what you expect to measure if it is correct. For a statistical test, the null hypothesis says that there is no difference or no association.

Data can support a hypothesis or fail to support it. They never prove it.

Random sampling avoids bias, because every individual or place has an equal chance of being chosen. Larger samples and more repeats give a better estimate of the true mean and show how much results vary. Aim for at least five values for each independent variable in an experiment, each repeated at least three times, and more if results vary a lot.

Used in: C4.1 sampling.

• Experiments: change one variable and control the others, so a cause can be shown.
• Observations and field studies: record what happens naturally. These can show a correlation, but other factors may be the cause.
• Databases: use data collected by others, such as gene sequences or climate records.
• Models and simulations: test ideas that are too slow, large or unethical to test directly.

Watch out: correlation does not show causation.

• Accurate: close to the true value.
• Precise: repeated values are close to each other.
• Random errors make values scatter. Reduce their effect by repeating and taking a mean.
• Systematic errors shift every value the same way, for example a balance that is not zeroed. Repeating does not help: fix the method or the instrument.

Write a risk assessment: hazards, risks and how to reduce them. Avoid harming animals, and use plants, microbes or simulations where possible. Do not collect personal data without consent. In the field, return organisms and leave the habitat as you found it.

A conclusion describes the trend, uses the data, including numbers, and explains it with biology. An evaluation names the most important weaknesses, says how each affected the results, and suggests a realistic improvement for each one. An extension asks a new, related question.

Place a thin specimen on a slide, add a stain, such as iodine for starch or methylene blue for animal cells, and lower a coverslip at an angle to avoid air bubbles. Focus with the low-power objective first, then move to higher power with the fine focus only.

To measure cells, calibrate an eyepiece graticule against a stage micrometer. Example: 25 eyepiece units line up with 100 µm on the stage micrometer, so 1 eyepiece unit = 4 µm at this magnification. Recalibrate for each objective.

Used in: A2.2 Cell structure.

Choose an instrument with a suitable resolution: a 10 cm³ pipette, not a beaker, for small volumes. Record every value to the resolution of the instrument, with units in the column heading only.

A common convention: the uncertainty of an analogue scale is ± half the smallest division, and of a digital display ± the last digit. A thermometer marked every 1 °C gives ±0.5 °C.

In a serial dilution, each solution is diluted by the same factor. Adding 1 cm³ of solution to 9 cm³ of water gives a tenfold dilution, so 1.0 mol dm⁻³ becomes 0.1, 0.01 and so on. For a range such as 0.2, 0.4 and 0.6 mol dm⁻³, mix stock and water in proportion: 2 cm³ of 1.0 mol dm⁻³ stock + 8 cm³ water gives 0.2 mol dm⁻³.

Used in: D2.3 Water potential.

Spot an extract on a line near the bottom of paper or a thin-layer plate, and stand it in a solvent below the line. Each substance moves at its own rate. R_f = distance moved by the spot ÷ distance moved by the solvent front. Example: 3.6 cm ÷ 8.0 cm = 0.45. Compare R_f values with known pigments.

Used in: C1.3 Photosynthesis.

A colorimeter shines light of one colour through a sample and measures how much is absorbed. Set it to zero with a blank, such as water, choose a filter of the complementary colour, and make a calibration curve from known concentrations. It can follow a reaction, such as the loss of starch–iodine colour as amylase works.

Used in: C1.1 Enzymes.

A simple respirometer holds small organisms, such as germinating seeds, with a chemical that absorbs carbon dioxide. As oxygen is used, the volume of gas falls and a coloured drop moves along a capillary tube. A control tube with glass beads corrects for changes in temperature and pressure.

Used in: C1.2 Cell respiration.

Probes measure oxygen, carbon dioxide, pH, temperature, light or pressure, and a data logger records readings automatically, often every second. They give many precise readings and avoid human error in timing, but they still need calibrating.

PCR copies a chosen section of DNA many times by repeated cycles of heating and cooling. Gel electrophoresis separates DNA fragments by size: DNA is negatively charged, so it moves towards the positive electrode, and small fragments travel furthest. A ladder of known sizes allows sizes to be estimated.

Used in: D1.1 DNA replication.

• Random quadrats: estimate the population of sessile organisms.
• Transects: sample along a line to show how organisms change across a gradient, such as up a shore.
• Capture-mark-release-recapture: estimate the population of motile animals.
• Measuring abiotic factors: light meters, soil thermometers, pH probes.

Used in: C4.1 Populations and communities.

• Line graph: a continuous independent variable, such as temperature.
• Scatter graph: looking for a correlation between two measured variables.
• Bar chart: categories, such as species.
• Histogram: the frequency of values in a continuous range.

Put the independent variable on the x-axis, label both axes with units, and add error bars where you have repeats. Spreadsheets save time on means and standard deviations. Online databases hold gene and protein sequences, and simulations let you vary conditions quickly.

1 mm = 1000 µm and 1 µm = 1000 nm. Very large and small numbers use standard form: 0.0000032 m = 3.2 × 10⁻⁶ m. Give answers to the same number of significant figures as the least precise data used.

Percentage change = (final − initial) ÷ initial × 100.

Used in: D2.3 Water potential.

Magnification = image size ÷ actual size. Convert to the same units first.

Used in: A2.2 Cell structure.

Rate = change ÷ time. On a graph, the rate is the gradient. Where the line curves, draw a tangent and find its gradient. Example: 18 cm³ of oxygen in 60 s gives 0.30 cm³ s⁻¹. A rate can also be 1 ÷ time, for a reaction timed to an end point.

Used in: C1.1 Enzymes.

For 4, 6, 6, 7 and 12: the mean is 35 ÷ 5 = 7, the median (middle value) is 6, the mode (most common) is 6, and the range is 12 − 4 = 8. One very high or low value affects the mean much more than the median. Check outliers for a mistake before deciding whether to leave them out, and say so if you do.

Standard deviation (SD) measures how spread out values are around the mean. Use a calculator or spreadsheet. For data that are normally distributed, about 68% of values lie within one SD of the mean.

Show SD as error bars above and below each mean. If the bars of two means overlap a lot, the difference may not be significant. If they do not overlap, a difference is more likely, but use a statistical test to decide.

Used in: C4.1 quadrats.

A correlation coefficient, r, runs from −1, a perfect negative correlation, through 0, none, to +1, a perfect positive correlation. r² gives the proportion of the variation in one variable that is linked to the other: r = 0.8 gives r² = 0.64, or 64%.

• Chi-squared: compares observed counts with expected counts, for example to test for an association between two species, or whether offspring fit a predicted ratio.
• t-test: compares two means, such as the height of plants in two conditions. It is often used in internal assessments.

Each gives a value to compare with a critical value at p = 0.05. If the calculated value is larger, reject the null hypothesis: there is less than a 5% probability that the result happened by chance.

Used in: C4.1 chi-squared.

Quadrats: population = mean per quadrat × total area ÷ quadrat area. Lincoln index: population = (M × N) ÷ R. Example: 40 marked, 50 caught later, 10 of them marked, so (40 × 50) ÷ 10 = 200.

Used in: C4.1 Populations.

On a log₁₀ scale, each equal step is ten times bigger: 1, 10, 100, 1000. Log scales fit a wide range of values on one graph, and they turn exponential growth into a straight line. pH is a log scale too: pH 5 has ten times the hydrogen ion concentration of pH 6.

Used in: C4.1 Modelling growth.