Interpreting Scientific Data in Tables, Graphs and Diagrams
This domain asks you to do more than read numbers off a page: you must decode how a data display was built, extract values with their units, and describe the trend the data actually supports rather than the one you expected.
Read the frame before the data
- Identify the independent variable (what the experimenter changed) and the dependent variable (what was measured in response). By convention the independent variable sits on the x-axis and the dependent variable on the y-axis.
- Check axis labels and units, and work out the value of each major and minor graduation. One square might be 2 °C, not 1 °C.
- Look for a broken axis (zigzag) or a non-zero origin; both change how steep the slope appears.
Tables
Column headings carry the unit; the cell carries only the number. Inconsistent decimals or a missing unit in a heading is a legitimate criticism to make.
Graphs
- A line of best fit summarises the trend; joining every dot implies continuous measurement. Either can be acceptable, but you must justify the choice.
- Slope has units of y per x and often a physical meaning (a rate). The y-intercept often represents an initial value.
- Interpolation inside the data range is defensible; extrapolation beyond it is a claim you must flag.
Worked example
A student heats 200 mL of water and plots temperature against time. The line rises from 20.0 °C at t = 0 min to 100.0 °C at t = 8.0 min, then stays horizontal until t = 14.0 min. Rate of heating = (100.0 − 20.0) °C ÷ 8.0 min = 10.0 °C/min. The flat section does not mean heating stopped: the burner is still on, but the energy now changes liquid water into vapour at the boiling point. Stating that interpretation, not merely "the line is flat", is what earns the mark.
Traps
- Reporting a value without its unit, or with more decimals than the instrument supports.
- Confusing the slope with the plotted value ("the temperature was 10 °C").
- Assuming a trend continues past the last data point.
- Ignoring an outlier that actually signals a measurement error, or deleting it without justification.
- Reading a diagram without checking its legend, labels or scale bar.