Pearson Science Double Award Practical 4: Conclusions and Evaluation

Evaluation connects evidence, scientific reasoning, limitations and targeted improvements. It is not a list of generic errors. Across 4SD0, a strong answer distinguishes what the data show from what the method can support.

Write an evidence-based conclusion

Begin with the direction and form of the pattern over the tested range. Cite representative processed values, a gradient, an optimum or a clear comparison. “The dependent variable increased” is weaker than identifying how much it changed and whether the trend was linear, curved, peaked or levelled.

Then explain the pattern with the relevant model. Enzyme activity may rise because collisions become more frequent before falling as active sites change shape. Reaction rate may rise because more frequent successful collisions occur. Current may change according to component resistance. Keep the explanation aligned with the variables actually measured.

State whether the evidence supports the prediction, not whether it “proves” a universal law. A school investigation samples a limited range, apparatus and biological or material set. Conclusions should remain within that boundary.

Reliability, validity and accuracy

Reliability concerns consistency. Close repeats, adequate sample size and reproducible patterns support reliability. Large unexplained scatter weakens it.

Validity concerns whether the method tests the intended relationship. Confounded variables, unsuitable proxies or non-representative sampling reduce validity even when readings are precise.

Accuracy concerns closeness to the accepted or true value. Calibration, zero error, energy loss and systematic endpoint bias can affect accuracy. A reliable set can still be systematically inaccurate.

Avoid using these terms as interchangeable praise. Tie each to evidence from the method or results.

Random and systematic effects

Random effects make repeated readings vary unpredictably. Reaction-time variation, judging a colour endpoint and biological differences can contribute. Repeats, means, longer measurement intervals and objective sensors may reduce their influence.

Systematic effects shift readings consistently. A balance zero error, heat loss, a miscalibrated probe or a ruler viewed from one offset position can bias every result. Repeating alone preserves the bias. Calibration, insulation, blank correction or redesign is needed.

Some limitations affect both scatter and bias. Explain the likely direction where possible. Heat loss in an exothermic reaction makes the measured temperature rise smaller, so calculated released energy is underestimated. Saying only “heat loss affects the result” misses this chain.

Limitation-effect-improvement chains

A useful evaluation contains:

1. a specific limitation rooted in the actual method;
2. its mechanism and likely effect on readings or conclusion;