5(f), 5(i); 3i Section 3 - Presenting investigation evidence

Matching representations to evidence

A representation is useful when it exposes a feature of the evidence relevant to the enquiry. A scatter graph can show the association between paired distance and sediment measurements; a distribution plot can show inequality hidden by a mean; a map can locate a pattern. A table remains useful when exact values need to be checked. Choosing three chart types solely to increase variety does not show why any is appropriate.

Start with the data structure and the comparison. Categories suit bars; ordered time may suit a line; continuous frequency classes suit a histogram; paired measurements suit a scatter graph. If time intervals are unequal, spacing on a time graph should reflect that difference. For a histogram with unequal class widths, use frequency density so area represents frequency. These choices affect the argument, not just appearance.

An illustrative access investigation has a mean journey time for each neighbourhood but also the individual observations. Showing only the means might imply similar experiences within each place. A distribution plot can reveal whether one neighbourhood contains both very short and very long journeys. A map can then locate those contrasts. The additional representation has a reason: it exposes variation relevant to social inequality.

State what each figure contains, its units, sample size, dates and relevant definitions. Do not allow a tidy graph to imply more certainty than the method supports. Error bars need an explained meaning—range, standard deviation and confidence intervals are not interchangeable. A plotted uncertainty measure must have been calculated appropriately, not chosen because it makes the graph look scientific.

Innovation is useful when it communicates something geographical more effectively. A clear map or graph can answer a question better than an elaborate display with unexplained scales. Independently select and produce representations that expose relevant patterns.

Worked frequency-density presentation

A frequency class spans 0–3 centimetres and contains six clasts; the next spans 3–9 centimetres and contains 12. Divide each frequency by class width:

The bars have equal height but the second has twice the area, representing twice the frequency. In a separate soil-depth sample, intervals of five and nine centimetres contain 20 and 18 observations:

The taller bar contains more observations here, but its height still represents density rather than count.

Maps and innovative spatial presentation

A spatial representation should help the reader relate evidence to location, scale and relevant geographical processes. A site-location map establishes where observations were made. A thematic map displays their values or categories. A GIS overlay can investigate whether different conditions coincide. Each has a distinct purpose, so a location map alone does not present the observed result.

Select symbols that match the variable. Proportional symbols can compare totals at points; choropleth shading can compare rates or densities over areas; a mapped transect can connect an environmental gradient to observed sites. If symbol circles represent quantities, scale their areas rather than radii to the values. If a shaded map displays an area average, avoid implying that the value applies uniformly throughout the polygon.

An illustrative beach enquiry could map pebble size at sampled points and add a coastline, river mouth and transport direction. These contextual layers allow the reader to relate the pattern to potential sources and pathways. But an arrow indicating inferred transport must be distinguished from directly measured movement. A broad smooth interpolated surface between a few points may imply knowledge of unsampled areas that the evidence cannot support.

GIS and other digital tools can make relationships easier to explore through layers, filtering or linked maps and graphs. Their usefulness depends on accurate inputs and explained processing. State coordinate systems, dates, resolution, classifications and any interpolation assumptions. A visually impressive map cannot repair incompatible boundaries or a missing sample zone.

Make the result accessible within the report. Provide an informative caption, readable legend, scale where relevant and source attribution. If an interactive product is used, explain the finding in continuous prose and reference or evidence the digital material as required, rather than making the reader discover the argument by clicking through it. A stable image or appropriate recorded view can preserve the relevant evidence when a live service changes.

Worked proportional-symbol sizes

A circle with radius four millimetres represents 100 households. To represent 400 households, multiply radius by the square root of the value ratio:

A different legend uses radius three millimetres for 80 people. For 180 people:

Scaling radius directly by the value ratio would exaggerate area and the implied quantity.

ONS's May 2023 Census maps guidance explains a real spatial-presentation limit: a displayed area statistic does not describe the particular building beneath its label. The guidance concerns Census 2021 maps. A learner mapping ward averages should distinguish the represented area from individual households.