Geographical Skills, Fieldwork, and Enquiry

You are given data from an unfamiliar fieldwork investigation. The data shows that environmental quality scores decrease with distance from a park. Evaluate the methodology used and suggest how you would investigate this relationship more thoroughly.

Model response: The methodology is a radial transect from a park, using a bipolar environmental quality survey at regular intervals. Several methodological questions need addressing. First, the environmental quality index (EQI) is subjective — different surveyors might score the same location differently. To improve reliability, the survey should be conducted by multiple surveyors at each point and scores averaged, or a more objective scoring rubric should be used. Second, the transect represents only one direction from the park; the pattern might differ in different directions depending on land use, road proximity and socioeconomic factors. Multiple transects in different directions would test whether the pattern is consistent. Third, the relationship might be correlational rather than causal: the park might be located in an area of high environmental quality for reasons unrelated to the park itself (e.g. a wealthy neighbourhood), meaning the park is not causing the pattern but is associated with it. To investigate causality, I would compare the environmental quality gradient around multiple parks in different contexts (wealthy vs deprived areas; large vs small parks; urban vs suburban parks) to test whether the pattern holds regardless of context. I would also collect secondary data (house prices, deprivation indices, population density) to control for socioeconomic factors. To test whether the park itself causes the gradient, I would compare the gradient before and after a new park was created to see if environmental quality improved following park development. This multi-method approach would provide stronger evidence for or against the causal claim.

How can maps be misleading, even when they use accurate data? Give examples of how cartographic choices affect the message a map communicates.

Model response: Maps are powerful tools for geographical communication, but every map involves choices that shape the message it conveys. Projection choice distorts spatial relationships: the Mercator projection (used by Google Maps) exaggerates the size of high-latitude countries (Greenland appears the same size as Africa despite being 14 times smaller), which some critics argue reinforces a Eurocentric worldview by making Europe and North America appear larger and more central than they are. Colour choice affects emotional response: red typically signals danger or intensity, so a map showing immigration flows in red may create a more alarming impression than the same data shown in blue. Class interval choice in choropleth maps can dramatically change the visual impression: a map of poverty rates using quartile intervals will show a balanced distribution, while one using equal intervals may show most areas as 'low poverty' with a few extreme outliers, depending on the data distribution. Scale affects what is visible: a national-scale map of deprivation hides local variation, while a neighbourhood-scale map may overemphasise local differences. Maps showing data by administrative area (countries, wards) impose artificial boundaries on continuous phenomena: poverty does not stop at a national border. GIS has made it easier to create professional-looking maps, which can increase the authority of maps that may be based on poor data or misleading choices. The most important geographical skill is the ability to read maps critically: understanding what choices the mapmaker has made, what alternative presentations would show, and what the map conceals as well as what it reveals.

A student claims that their fieldwork data 'proves' that river velocity increases downstream because their Spearman's rank result was significant. Evaluate this claim.

Model response: The claim overstates what the data can demonstrate. While a significant Spearman's rank correlation provides evidence that supports the hypothesis, it does not 'prove' it, for several reasons. First, statistical significance at the 0.05 level means there is still a 5% probability that the correlation occurred by chance — significant is not the same as certain. Second, the result applies only to the specific river section and the specific dates on which measurements were taken; discharge, gradient and channel conditions vary between rivers and between seasons, so the pattern may not generalise. Third, with a small sample size (common in fieldwork — often 5-10 sites), a single anomalous measurement can significantly affect the correlation coefficient. Fourth, correlation does not establish causation: velocity might increase downstream because of increasing discharge, decreasing bed roughness, or changing channel efficiency, and the statistical test does not tell us which mechanism is responsible. Fifth, Spearman's rank tests whether the data shows a monotonic trend (consistently increasing or decreasing) but cannot detect non-linear relationships — velocity might increase downstream overall but decrease at specific points (behind obstacles, at confluences) in ways that Spearman's rank does not capture. The appropriate claim is: 'Our data provides statistically significant evidence that supports the hypothesis that velocity increases downstream at this river on this date, but the small sample size and potential confounding variables mean the result should be interpreted cautiously.' This precision is what distinguishes rigorous geographical analysis from over-confident assertion.