CourseCourseQ BankQuestion BankDocsDocuments

Paper 5 Planning, Analysis and Evaluation

Syllabus
9700–2028–2029
Topic
—
Level
A2

Learning objectives

P5.1Defining the problem and predictionUse the context to state a relevant prediction in words or as a sketch graph, link it to a hypothesis, identify independent and dependent variables and identify key variables that must be standardised.P5.2Planning experimental methodsDescribe how to vary the independent variable; measure independent and dependent variables accurately and to appropriate precision; standardise key variables; choose suitable volumes and concentrations; prepare different concentrations by serial dilution or proportional dilution; include controls; sequence the method logically; plan how to collect results.P5.3Assessing result quality, validity and riskDescribe how to judge result quality using anomalies, spread, standard deviation, standard error or 95% confidence intervals; assess validity using accuracy and repeatability; prepare a simple risk assessment considering hazard severity and probability; state precautions to reduce risk.P5.4Dealing with experimental dataUse tables and graphs to show key quantitative data; draw suitable graphs with independent variable on the x-axis and dependent variable on the y-axis; add confidence-limit error bars where required; choose calculations including means, percentages, percentage gain or loss and rates of change.P5.5Statistical tests and data types in practical contextsUse standard deviation, standard error or error bars to judge likely significance; choose and carry out appropriate statistical tests from the mathematical requirements; justify the test; state a null hypothesis; recognise categoric, ordinal and continuous data.P5.6Conclusions from raw and processed dataSummarise main conclusions; identify key points in raw data, processed data, graphs and statistical results; discuss how far a hypothesis is supported; give scientific explanations; make further predictions and hypotheses.P5.7Evaluation of investigationsIdentify anomalous values and possible explanations; assess replication, independent-variable range and intervals, measurement method and control of variables; judge validity, how well data test the hypothesis and confidence in conclusions; suggest improvements that increase confidence.

A prediction links the IV, DV and biological hypothesis

A Paper 5 plan begins with a testable prediction: the expected relationship between the independent variable (IV) and dependent variable (DV). The prediction must follow from an underlying biological hypothesis, not merely restate the investigation topic.

  1. Identify the factor deliberately varied as the IV and the measurable response as the DV.
  2. Predict direction and, where justified, shape or limit: increase, decrease, optimum, plateau or no change.
  3. Explain the biological mechanism that makes this pattern plausible.
  4. State the key variables that could also affect the DV and therefore must be standardised.
  5. If using a sketch graph, put IV on x, DV on y, label both, and draw only the expected relationship.

Prediction: increasing light intensity will increase photosynthetic rate until a plateau. Hypothesis: more light initially supplies more energy for the light-dependent reactions; at high intensity another factor becomes limiting. IV = light intensity, DV = oxygen production per unit time; temperature, carbon dioxide availability and plant material must be standardised.

'Investigate the effect of light on photosynthesis' is an aim, not a prediction. Do not standardise the IV or DV, and do not list trivial variation that is unlikely to affect the result.

A method must be reproducible, controlled and measurable

  1. State how the IV will be varied across a suitable range and intervals, with values, units and enough levels to reveal a pattern.
  2. State how IV and DV will be measured accurately and to appropriate precision: apparatus, volume/concentration, timing, endpoint and unit.
  3. Standardise each influential variable with a named method and value.
  4. Include a control that removes the tested factor or supplies a baseline.
  5. Repeat each treatment, identify anomalies and calculate a mean where appropriate.
  6. Write a logical numbered sequence showing quantities, order, apparatus and how results are recorded.
Dilution Method Best interpretation
Proportional calculate stock volume with C1V1=C2V2C_1V_1=C_2V_2, then add diluent to the chosen final volume each concentration is made independently from stock
Serial transfer a fixed volume into diluent, mix, then use that dilution to make the next each step changes concentration by the same dilution factor

To make 20 cm320\text{ cm}^3 of 0.40 mol dm−3^{-3} solution from 1.00 mol dm−3^{-3} stock: V1=(0.40×20)/1.00=8.0 cm3V_1=(0.40\times20)/1.00=8.0\text{ cm}^3 stock; add diluent to a final volume of 20 cm320\text{ cm}^3. Keep total volume constant across treatments.

A method is not reproducible if it says only 'measure accurately' or 'keep conditions constant'. Name the apparatus, precision, quantity, condition and action. A control is a purposeful comparison, not an unlabelled extra repeat.

Quality, validity and risk answer different questions

Judgement Evidence to inspect Meaning
Result quality anomalies, replicate spread, standard deviation, standard error, 95% confidence intervals how variable/precise the estimates are and whether repeats agree
Accuracy calibration, resolution, method bias and closeness to the true/accepted value where available whether measurements are centred correctly
Repeatability agreement when the same method/equipment/operator repeats measurements one aspect of precision, not proof of accuracy
Validity suitable DV, effective control/standardisation, appropriate range and absence of confounding whether the data genuinely test the hypothesis
Risk severity of harm × probability in the planned procedure whether precautions reduce risk to an acceptable level

A larger sample and repeat measurements reveal variation. SD describes spread of observations; SE and 95% CI describe uncertainty in an estimated mean. Narrower intervals support a more precise mean, but only if sampling and design are appropriate.

For each relevant hazard: name it → state possible harm → judge severity and probability → give a specific precaution → state safe disposal or emergency action where needed. Use concentration, quantity, temperature, exposure route and organism when judging risk.

Closely grouped wrong readings are repeatable but inaccurate. A precise method can still be invalid if it measures the wrong response or leaves a confounding variable uncontrolled; low risk says nothing about scientific validity.

Process data only when it clarifies the biological comparison

Data processing should expose the pattern or comparison needed for a conclusion. Preserve key raw data, then choose tables, graphs and calculations that answer the stated question rather than calculating everything available.

Question Useful processing
typical response at each IV value mean of valid replicates, with variation shown
part of a total percentage =(part/total)×100=(\text{part}/\text{total})\times100
change relative to a baseline percentage gain/loss =((final−initial)/initial)×100=((\text{final}-\text{initial})/\text{initial})\times100; retain sign or label gain/loss
speed of change rate =Δquantity/Δtime=\Delta\text{quantity}/\Delta\text{time} with units
relationship across IV values suitable graph with IV on x and DV on y
uncertainty in means SD/SE or 95% confidence-limit error bars as requested

Use headings and units, justified precision and all key values. Choose scales that show the data clearly. Error bars must be identified (for example SE or 95% CI); their overlap is evidence about likely differences, not a universal statistical decision by itself.

Percentage change needs a non-zero stated baseline, and rate needs a stated interval. A graph cannot repair missing units, omitted data or an inappropriate calculation; do not remove points merely to create a smoother trend.

Choose a statistical test from data type and question

Data type Meaning Example
nominal/categorical names or groups with no order flower colour
ordinal ranked categories; gaps need not be equal order in which tubes lose colour
continuous any measured value within a range mass or leaf length
Question and suitable data Test
observed categorical frequencies versus expected frequencies chi-squared
difference between two means of continuous data under suitable assumptions t-test
monotonic association using ranks or data not meeting Pearson assumptions Spearman's rank correlation
linear association between two continuous variables under suitable assumptions Pearson's linear correlation
  1. State H0H_0: no difference/association beyond chance.
  2. Justify the test from data type, question, pairing/independence, distribution and relationship shape.
  3. Calculate using the supplied formula and correct sample information.
  4. Compare the test statistic or probability with the stated critical value/significance level.
  5. Reject or do not reject H0H_0, then state the conclusion in biological context.

SD, SE and error bars can indicate spread or uncertainty and whether means may differ, but they are not substitutes for a required statistical test. A small probability under H0H_0 supports statistical significance; effect size and biological importance still need interpretation.

Do not 'accept' a biological mechanism because H0H_0 is rejected. The result supports a difference or association under the test assumptions; it does not prove causation or explain why the pattern exists.

A conclusion integrates pattern, uncertainty and biological explanation

  1. State the main relationship or difference and support it with key raw or processed values.
  2. Include relevant graph features, variation/error bars and statistical result.
  3. State how far the experimental evidence supports the hypothesis—not whether it is universally proved.
  4. Identify strengths and weaknesses that change confidence.
  5. Give a detailed biological explanation consistent with the observed pattern.
  6. Form a further prediction or hypothesis that logically follows and could be tested.
Evidence What it supports
raw values and sample size what was actually observed and whether a claim is representative
means/rates/percentages and graph size, direction and shape of the effect
SD/SE/95% CI variation or uncertainty in estimates
statistical test likelihood of the observed difference/association under H0H_0
method evaluation confidence that the pattern tests the intended hypothesis

If treatment mean is higher but 95% confidence intervals overlap substantially and replication is limited, report the observed increase but keep hypothesis support cautious. If a valid test is significant, reject H0H_0 for that comparison; then explain the possible biology separately.

Statistical significance does not establish the mechanism, causation or practical importance. A conclusion must not omit conflicting data, and a further hypothesis must be testable rather than a speculative story.

Evaluate the limitations that change confidence

Check Why it matters Targeted improvement
anomalous values may reflect error or real variation and can distort summaries verify records/method, repeat the measurement; exclude only with justification
replication/sample size too few repeats hide variability and weaken means/statistics increase independent replicates using the same standardised method
IV range and intervals may miss an optimum, threshold, plateau or trend shape extend range safely or add closer intermediate values where change is rapid
DV measurement subjective/low-resolution measure can reduce accuracy and repeatability use calibrated, objective or higher-resolution measurement
controlled variables confounding prevents attribution to the IV standardise the influential variable with a named value/method
controls missing baseline makes treatment effect ambiguous add a negative/positive or no-treatment control appropriate to the hypothesis

Conclude by judging: whether the investigation is valid; whether the data actually test the hypothesis; and how much confidence the conclusion deserves. Prioritise limitations by their likely effect on the main result, not by how easy they are to mention.

Write each evaluation as limitation → effect on data/conclusion → specific change → reason confidence improves. For example, uncontrolled temperature could change enzyme rate independently of substrate concentration; use a monitored thermostatic bath so differences are more plausibly due to the intended IV.

'Use better equipment', 'repeat more' and 'be more careful' are incomplete unless the equipment, number/type of repeats, procedural change and expected improvement are specified. No design becomes perfect; state the most consequential remaining uncertainty.