HookTwo numbers that find you on a mountain
When a walker calls mountain rescue from a fog-bound ridge, 'we're somewhere near the summit' is almost useless. A six-figure grid reference — six digits, read off an Ordnance Survey map or a phone's GPS — pins them to a square just 100 metres across anywhere in Great Britain. The whole of the country sits under a grid of numbered lines, and the reference is simply how far east and then how far north the point lies. Rescue teams live or die by getting those two numbers the right way round.
Geographical skills are the mountain-rescue part of your GCSE: unglamorous, precise, and worth a great deal when you get them right. They are not a topic you sit once — they are assessed across all three papers, woven into questions about hazards, cities, rivers and development, and they make up close to a third of the marks on offer. They are also the marks students throw away most cheaply: a grid reference written back-to-front, a percentage change divided by the wrong number, a bar chart drawn where a line graph was needed. Every one of those is a mark you already knew how to earn. This section drills the five skill families — cartographic, graphical, numerical, statistical, and handling qualitative and quantitative data — until they become reflex.
ModelA third of your marks live here
Geographical skills are not confined to one paper. AQA assesses them everywhere, mostly under AO4 — selecting, adapting and using a variety of skills and techniques to investigate questions and communicate findings — with a good share under AO3, where you interpret and analyse geographical information. Across the whole GCSE these objectives are worth roughly half the marks between them, and skills questions are the most reliable of the lot because they are usually point-marked: a grid reference is either correct or it is not, so there is no partial-credit fog to hide in.
That cuts both ways. Point marking means precision is everything and there is no reward for 'nearly'. But it also means these are the highest marks-per-minute in the exam if you have drilled the method. Learn the five families as a toolkit — know which tool a question is reaching for, then apply it exactly. The families are cartographic (maps), graphical (charts), numerical (arithmetic and averages), statistical (spread and correlation), and the handling of qualitative versus quantitative evidence.
MechanismCartographic skills — reading the OS map
Two Ordnance Survey scales dominate. On a 1:50,000 map (the pink Landranger series) 2 cm on the paper is 1 km on the ground, so 1 cm is 500 m. On a 1:25,000 map (the orange Explorer series) 4 cm is 1 km, so 1 cm is 250 m and far more detail is shown. A larger-scale map covers a smaller area in more detail — a point examiners love to test.
Grid references locate points on the numbered national grid. A four-figure reference names a whole 1 km square; a six-figure reference pins a point to about 100 m. The unbreakable rule is eastings before northings — read along the bottom first, then up the side. Generations have remembered it as 'along the corridor, then up the stairs'. For six figures, imagine each grid square divided into tenths and estimate the extra digit in each direction.
Beyond grid references you must handle direction (the eight-point compass and bearings measured clockwise from north), distance (measure with a ruler or piece of string, then convert using the scale), relief and gradient (read the contours — lines joining points of equal height; contours close together mean steep ground), cross-sections, and sketch maps. Then there are thematic maps: choropleth maps shade areas by value, isoline maps join points of equal value (like contours or isobars), dot maps place one dot per quantity, proportional symbol maps size the symbol to the value, and flow-line maps show movement with arrows of varying width. When asked to describe a distribution, give the general pattern, quote the highest and lowest values, and name any anomaly.
Two quick map calculations. First, a six-figure grid reference. A church lies four-tenths of the way across grid square 63–64 (easting) and seven-tenths up square 42–43 (northing). Read eastings first: 63, then the tenth, gives 634. Then northings: 42, then the tenth, gives 427. The reference is 634427 — never 427634. Second, a gradient. A footpath climbs from a car park at 120 m to a summit at 220 m over a horizontal distance measured on the map as 500 m. Gradient = vertical rise ÷ horizontal distance = (220 − 120) ÷ 500 = 100 ÷ 500 = 0.2, usually written as 1 in 5 (for every 5 m along, you climb 1 m). Both marks turn entirely on order and units — eastings before northings, and rise and run in the same unit before you divide.
MechanismGraphical skills — the right chart for the data
Half of a graph mark is choosing the correct type, so learn what each is for. A line graph shows continuous change, especially over time — the temperature line on a climate graph. A bar chart compares separate categories. A divided (compound) bar or a pie chart shows how a whole splits into parts — the composition of a country's energy mix, say. A histogram shows continuous data grouped into equal class intervals, with no gaps between the bars. A scattergraph tests the relationship between two variables and takes a line of best fit. A population pyramid shows age and sex structure. A climate graph combines two: rainfall as bars, temperature as a line, on the same axes.
You need to complete and label part-drawn graphs, plot points accurately, and read values off precisely. When a question says 'describe the graph', do not just state the obvious — quote figures, give the trend, and flag anomalies. Accuracy in plotting is point-marked, so a point placed even slightly off the gridline can cost the mark.
Reading a population pyramid tells you a country's development story at a glance. A pyramid with a very wide base — many young children — and a narrow top signals a high birth rate and low life expectancy, typical of a lower-income country such as The Gambia or Niger; it also means a high youth dependency, with many children supported by relatively few working adults. A pyramid that is narrow at the base and bulges at the top shows an ageing population with a low birth rate, typical of Japan or the UK, where the pressure is old-age dependency — pensions and healthcare for a growing elderly share. If asked to compare two pyramids, name the shape, quote the widest age band on each, and translate it into a dependency point rather than just saying one has 'more old people'.
DataNumerical skills — percentages without slips
The numerical family covers understanding of number and place value, area, scale, proportion and ratio, magnitude and frequency, percentages, and the measures of central tendency: the mean (add them up, divide by how many), the median (the middle value when they are ordered) and the mode (the most common value). None of it is hard maths; all of it is easy to fumble under pressure.
Percentage change is the single most-tested calculation and the most misfired. The formula is (new value − original value) ÷ original value × 100. The mark is almost always lost by dividing by the new value instead of the original, or by dropping the minus sign on a fall. Ratio and proportion come up too — expressing, say, a 3:1 split of urban to rural population, or the proportion of a budget spent on flood defence.
Percentage change, both directions, one worked example. A Cornish fishing village sees its population fall from 2,500 to 1,900 over a decade. Percentage change = (1,900 − 2,500) ÷ 2,500 × 100 = (−600 ÷ 2,500) × 100 = −24%. The value fell, so the answer is negative — dropping that minus sign changes the meaning entirely. Now a rise: a city's solar capacity grows from 40 MW to 70 MW. Percentage change = (70 − 40) ÷ 40 × 100 = (30 ÷ 40) × 100 = +75%. Notice the denominator is the ORIGINAL figure in both cases. Divide 30 by the new value of 70 by mistake and you get about 43% — a wrong answer that looks plausible, which is exactly why it is the classic trap.
DataStatistical skills — spread and correlation
Statistics move beyond a single average to describe the spread of data and the relationship between two sets. Alongside mean, median, mode and modal class you need the range (highest − lowest), quartiles (the values a quarter and three-quarters of the way through ordered data), and the interquartile range (IQR = upper quartile − lower quartile), which measures the spread of the middle half and, unlike the range, ignores freak extremes. Quantiles such as quartiles and percentiles split ranked data into equal-sized groups.
For two variables you describe correlation from a scattergraph: a positive correlation slopes up (as one rises, so does the other), a negative correlation slopes down, and scattered points with no trend show no correlation. You summarise the trend with a single straight line of best fit drawn through the middle of the cloud — you do not join the dots, because joining them hides the very relationship you are trying to show.
Interquartile range from a pedestrian survey. Counts at eleven high-street sites, put in order, are: 4, 8, 11, 14, 18, 22, 27, 33, 40, 52, 61. With eleven values the median is the sixth: 22. Split the data either side of it. The lower half is 4, 8, 11, 14, 18, whose middle value (the lower quartile) is 11; the upper half is 27, 33, 40, 52, 61, whose middle value (the upper quartile) is 40. So IQR = 40 − 11 = 29 people. Compare that with the full range, 61 − 4 = 57: the IQR is much smaller because it strips out the quiet site (4) and the crowded one (61), giving a fairer picture of a typical site. If you then plotted footfall against distance from the centre and the points sloped downward, you would describe it as a negative correlation and draw one straight line of best fit through them — not a zig-zag joining every dot.
MechanismQualitative and quantitative data — different truths
Geographers weigh two kinds of evidence. Quantitative data is numerical and measured — river velocity, rainfall totals, census populations, pedestrian counts. Its strength is that it is objective, comparable and easy to graph; its weakness is that numbers alone cannot capture how a place feels or why people act. Qualitative data is descriptive — interview transcripts, questionnaire opinions, field sketches, photographs, and written sources. Its strength is depth and meaning; its weakness is that it is subjective and can carry bias, so it is harder to compare fairly.
The skill the exam rewards is matching the data to the question and combining the two. To judge whether a regeneration scheme 'worked', hard numbers (jobs created, house prices, footfall) tell you the scale of change, while residents' interviews tell you whether the change improved daily life. Using both — triangulation — gives a fuller, more trustworthy answer than either alone. You should also be able to interpret and critique a given source: a photograph shows one moment from one angle and may be chosen to persuade; a questionnaire's results depend on who was asked and how the question was framed. Reading evidence sceptically is itself a marked skill.
VocabularyKey terms the mark scheme pays for
TrapsMisconceptions that cost marks
ExamWhat examiners want
Skills questions are assessed mainly under AO4, with interpretation under AO3, and most are point-marked — so precision earns the mark and 'nearly' earns nothing. Write grid references eastings-first with no spaces, keep rise and run in the same unit before finding a gradient, and label graph axes with units. On any calculation, show the working: an answer with method visible can still pick up marks if the final figure slips, whereas a bare wrong number scores zero.
When a command word is 'describe' — a distribution on a map, a trend on a graph — quote specific figures, state the overall pattern, and name any anomaly; a vague 'it increases' sits in the lowest band. When it is 'calculate', give the units and a sensible number of significant figures. When it is 'suggest which technique', justify the choice by matching the data type to the graph (continuous change to a line, composition to a divided bar, a relationship to a scattergraph).
Finally, treat qualitative and quantitative evidence critically: if handed a photograph or a questionnaire result, note that it captures one viewpoint and may be biased, and say how a second source would test it. Examiners reward the student who reads a source sceptically over the one who takes it at face value.