Evaluation — OCR A-Level Biology

Evaluating results means judging how far the data support a claim, not just restating them. Work through four steps. First, describe the pattern the data show, quoting values with units. Second, judge reliability: were repeats consistent, and how much do readings scatter about the mean? Third, judge validity: did the method measure the intended variable, were control variables held constant, and could systematic error shift every reading the same way? Fourth, draw a conclusion that states what the evidence supports and how confident you can be. For example, if a respirometer gives oxygen uptake of 0.42 cm³ min⁻¹ at 20 °C and 0.61 cm³ min⁻¹ at 30 °C, you can conclude that uptake rose over that range, but you cannot claim a trend beyond the temperatures tested. Link each conclusion to specific evidence and acknowledge limitations.

An anomaly is a measurement that does not fit the pattern shown by the other readings. To identify one, first look at the spread of the repeat readings and the trend across the range tested. A value lying well outside the scatter of the others, or breaking a clear trend, is a candidate anomaly. Decide whether it is a genuine outlier or a recording or measurement error, and justify your decision with the surrounding data. For example, if triplicate readings are 12, 13 and 31 arbitrary units, the value 31 is far from the others and is likely anomalous. Anomalies are normally excluded from a mean, but you should state why. Distinguish an anomaly from natural variation: small scatter is expected, whereas a single value far outside the pattern is not.

Every experimental procedure has limitations: features of the design, apparatus or control that reduce how far the results support a conclusion. To evaluate them, identify the specific limitation, explain its effect on the measured variable, and suggest a realistic improvement. For example, in a catalase-and-hydrogen-peroxide investigation, counting bubbles per minute is limited because bubble size varies and the gas is not measured directly; collecting gas in a syringe or inverted measuring cylinder gives a volume in cm³ and reduces observer subjectivity. Other common limitations include uncontrolled temperature, single readings with no repeats, a narrow range of the independent variable, and sampling bias. Distinguish a limitation (a weakness in the procedure) from an anomaly (an odd data point) and from an error (a deviation in a measurement).

Precision describes how closely repeated measurements agree with each other, while accuracy describes how close a measurement is to the true value. A set of readings can be precise yet inaccurate if a systematic error, such as a zero error on a balance, shifts every value the same way. Uncertainty in apparatus is the doubt in a single reading, often taken as half the smallest scale division for an analogue instrument or the stated tolerance for a digital one. Margin of error is the range within which the true value is expected to lie. Percentage error compares the uncertainty with the measured value: percentage error = (uncertainty ÷ measured value) × 100. For example, a 0.01 cm³ uncertainty on a 2.00 cm³ reading gives (0.01 ÷ 2.00) × 100 = 0.5%, but the same uncertainty on 0.20 cm³ gives 5%, so small readings carry larger percentage errors.

Refining experimental design means proposing specific, justified improvements to procedures and apparatus so that a practical produces more valid and reliable data. Work through the design logically: identify the independent variable and how it is set or changed, the dependent variable and how it is measured, and the control variables and how each is held constant. Then ask where error could enter. For apparatus, suggest a more precise measuring instrument, a better scale, or equipment that reduces heat loss or contamination. For procedure, suggest more repeats, randomising order, standardising timing or temperature, or using a control. Each improvement must be linked to the error it reduces and the effect on the data, not simply listed. For example, replacing a measuring cylinder with a burette reduces reading uncertainty when small volumes are dispensed.