Scientific Skills and Processes: Investigation, Measurement and Data Handling
Planning a fair test
Every investigation begins with a hypothesis - a prediction that can be tested and shown to be wrong. Identify the independent variable (what you deliberately change), the dependent variable (what you measure) and the control variables (everything kept the same). If control variables are not held constant, the test is not fair and the conclusion is unsafe. A control group is a separate comparison group that receives no treatment; it is not the same thing as a control variable.
Measuring properly
Use SI units - metre, kilogram, second, ampere, kelvin, mole - with prefixes such as milli-, centi- and kilo-. Read scales at eye level, taking the bottom of the meniscus for water. Record to the smallest division plus one estimated digit, and attach units to every value. Allow a thermometer to stabilise before reading it.
Recording and processing
Place the independent variable in the first table column with headings and units, and plot it on the x-axis. Draw a line of best fit rather than joining every dot. Repeat readings at least three times, identify anomalous results, exclude them, and average the remainder.
Accuracy, precision and error
Accuracy is closeness to the true value; precision is how closely repeats agree. Systematic error shifts every reading the same way (a zero error); random error scatters readings (reaction time). Precise readings can still be inaccurate.
Traps examiners set
Confusing accuracy with precision; calling a control variable a control group; averaging in an anomaly; forgetting unit conversion; claiming causation from a correlation; choosing a graph scale that hides the pattern.
Worked example
Four titres read 24, 25, 31 and 26 cm³. The 31 cm³ value is anomalous, so the mean is (24 + 25 + 26) ÷ 3 = 25 cm³.