OxfordAQA International A-level Geography (9635): 3.5 Geographical and fieldwork skills – Study Guide

This guide covers section 3.5, Unit 5: Geographical and Fieldwork skills, of the OxfordAQA International AS and A-level Geography (9635) specification, Version 3.2. It works through the geographical skills checklist (3.5.1.1 to 3.5.1.5) and the fieldwork skills (3.5.2.1 to 3.5.2.5). Unit 5 is a written exam of 1 hour 30 minutes with 60 marks, all questions compulsory, worth 20% of the International A-level. The specification also says knowledge of the skills list is needed for Units 1 to 4, so these techniques can appear on every paper.

Use it with the revision notes and the practice questions. The OxfordAQA A-level Geography exam preparation guide explains how the five units fit together. Track your progress on the OxfordAQA A-level Geography hub and checklist.

Syllabus coverage

Syllabus reference	What you must know or be able to do (from the document)	Where assessed
3.5.1.1 Core skills	Use and annotate base maps, sketch maps, diagrams, graphs, field sketches, photographs, geospatial, geo-located and digital imagery; literacy; numeracy; questionnaire and interview techniques	Unit 5; also Units 1-4
3.5.1.2 Cartographic skills	Atlas maps; located proportional symbols; flow lines, desire lines and trip lines; choropleth, isoline and dot maps	Unit 5; also Units 1-4
3.5.1.3 Graphical skills	Line and bar graphs (simple, comparative, compound, divergent); scatter graphs with best fit line; pie charts and proportional divided circles; triangular graphs; logarithmic scales; dispersion diagrams	Unit 5; also Units 1-4
3.5.1.4 Statistical skills	Mean, mode, median; range, inter-quartile range, standard deviation; Spearman’s rank correlation and Chi-square test, and the application of significance tests	Unit 5; also Units 1-4
3.5.1.5 ICT skills	Remotely sensed data; electronic databases; crowd sourcing and ‘big data’	Unit 5; also Units 1-4
3.5.2.1-3.5.2.5 Fieldwork	Planning and preparation; collection of data and sampling; presentation; analysis; drawing conclusions, ethics and evaluating the enquiry	Unit 5 (familiar and unfamiliar fieldwork questions)

How Unit 5 is assessed

• Entry code: GG05.
• 15 of the 60 marks assess your own investigation.
• Fieldwork is tested in two ways. Unfamiliar fieldwork questions ask you to apply fieldwork knowledge to new scenarios. Familiar fieldwork questions ask about your own experiences and findings.
• Approximate weightings of Unit 5 within the A-level: AO1 5.0%, AO2 6.0%, AO3 9.0%. So AO3 (quantitative, qualitative and fieldwork skills) carries the most weight.

The specification does not print a list of command words; your teacher will confirm them.

3.5.1.1 Core skills

Annotation means adding labels that explain, not just name. “Steep slope” is a label. “Steep slope, so faster overland flow and thin soils” is an annotation. Annotate base maps, sketch maps, field sketches, photographs and digital or geo-located imagery the same way.

Literacy covers reading factual and discursive material, such as a planning report or a resident’s letter, and judging its purpose and bias. Numeracy covers using number and measurement: units, percentages, rates of change and sensible rounding.

Questionnaires use fixed questions, often closed, so answers can be counted and compared. Interviews are longer and open, giving depth but taking more time and being harder to compare. Pilot the questions, avoid leading wording, and record when and where each response was collected.

3.5.1.2 Cartographic skills

Map type	Shows	Strength	Limitation
Atlas maps	Small-scale maps of physical and human features, such as relief, climate or population	Puts a place or pattern in its regional or global context	Too generalised to show local detail
Proportional symbols	Circles or bars sized to a value at a point	Shows magnitude and location together	Large symbols overlap
Flow lines	Line width proportional to the volume moving along a route	Shows volume and route	Cluttered where routes converge
Desire lines	Straight lines from origin to destination	Shows the pattern of trips simply	Ignores the actual route
Trip lines	Each line is one individual journey	Shows the raw pattern	Hard to read with many trips
Choropleth	Areas shaded by value per unit area	Clear regional pattern	Implies sharp change at boundaries; hides variation inside areas
Isoline	Lines joining equal values	Shows gradual change	Needs many data points; interpolation between points is an estimate
Dot	Each dot represents a set number	Shows density and clustering	Slow to draw; dots may merge

Choropleth maps should show rates or densities, not raw totals. A large district will often have a large total simply because it is large.

3.5.1.3 Graphical skills

• Line graphs show continuous change. A simple line graph has one line; a comparative line graph plots two or more lines on the same axes. A compound line graph stacks values so the top line is the total. A divergent line graph shows values above and below a central value, such as zero.
• Bar graphs compare categories. Comparative bars sit side by side; compound bars are divided into parts; divergent bars extend either side of a central axis, as in a population pyramid.
• Scatter graphs test a relationship between two variables. Draw a best fit line with roughly equal numbers of points either side. Do not force it through the origin. Name the anomalies (residuals).
• Pie charts show proportions: each segment is the percentage multiplied by 3.6 degrees. Proportional divided circles also make each circle’s area proportional to its total, so the radius is proportional to the square root of the value. A value four times larger needs a radius two times larger.
• Triangular graphs plot data with three components that add to 100%, such as employment in primary, secondary and tertiary sectors, or sand, silt and clay in a soil sample. Read each axis along its own gridlines.
• Logarithmic scales compress a wide range of values (for example 10 to 1,000,000) on to one axis. Each cycle is a tenfold increase. On a graph with a logarithmic vertical axis, equal slopes show equal percentage (proportional) rates of change, not equal absolute changes.
• Dispersion diagrams plot each value as a dot along a vertical scale. They show the spread, clustering and outliers in a data set, and you can mark the median and quartiles on them.

3.5.1.4 Statistical skills

All data below is fictional, from a made-up town, Aldervale.

Central tendency and dispersion

Pedestrian counts at 11 sites, in order: 12, 18, 18, 22, 25, 31, 34, 40, 46, 52, 75.

• Mean = 373 / 11 = 33.9. It uses every value but is pulled up by the high value, 75.
• Median = 6th value = 31. It resists extreme values.
• Mode = 18. It says little here.
• Range = 75 - 12 = 63. Easy, but distorted by one extreme value.
• Inter-quartile range: lower quartile = value at position (n + 1) / 4 = 3rd = 18; upper quartile = position 3(n + 1) / 4 = 9th = 46. IQR = 46 - 18 = 28. It describes the middle half, ignoring extremes.
• Standard deviation = square root of [sum of (x - mean)^2 / n]. The squared deviations sum to 3,414.9, so SD = square root of (3,414.9 / 11) = 17.6. About two-thirds of values in a normal-shaped distribution lie within one SD of the mean. Some textbooks divide by n - 1 instead (giving 18.5 here); your teacher will confirm which version to use.

Spearman’s rank correlation

Charles Spearman published this coefficient in 1904. It tests the strength and direction of a relationship between two sets of paired data.

1. Write a null hypothesis: there is no relationship between distance from the centre and environmental quality score.
2. Rank each variable separately. Tied values get the average of the ranks they share.
3. Find d (difference in ranks) for each pair, square it and add up.
4. rs = 1 - (6 x sum of d^2) / (n^3 - n).

Site	Distance (m)	Rank	EQ score	Rank	d	d^2
A	120	10	11	9	1	1
B	250	9	9	10	-1	1
C	310	8	14	7.5	0.5	0.25
D	400	7	14	7.5	-0.5	0.25
E	480	6	18	4	2	4
F	560	5	16	6	-1	1
G	650	4	21	3	1	1
H	720	3	17	5	-2	4
I	830	2	24	1	1	1
J	900	1	22	2	-1	1

Sum of d^2 = 14.5. rs = 1 - (6 x 14.5) / (1,000 - 10) = 1 - 87 / 990 = 0.912.

rs runs from +1 (perfect positive) to -1 (perfect negative). For n = 10, two-tailed critical values (used with a non-directional null hypothesis like this one) are 0.648 at the 0.05 level and 0.794 at the 0.01 level. One-tailed values for n = 10 are lower: 0.564 (0.05) and 0.745 (0.01). Because 0.912 is above 0.794, reject the null hypothesis: there is a significant positive relationship, with less than a 1% probability that it arose by chance. Tables differ for one-tailed and two-tailed tests, so use the one you are given. With only a few ties, as here, the formula still gives a sound result.

Evaluation: correlation is not causation. Ranks ignore the size of gaps between values, and small samples are unreliable.

Chi-square test

Karl Pearson introduced the chi-square test in 1900. It compares observed frequencies (O) with the expected frequencies (E) you would get if the null hypothesis were true.

Example: 90 litter items counted in three equal-area zones of Hollin Park, Aldervale. Observed 42, 30, 18. Null hypothesis: litter is evenly spread, so E = 30 in each zone.

Chi-square = sum of (O - E)^2 / E = 144/30 + 0/30 + 144/30 = 9.6.

Degrees of freedom = number of categories - 1 = 2. For a table with rows and columns, df = (rows - 1) x (columns - 1). Critical values for df = 2 are 5.99 (0.05) and 9.21 (0.01). Since 9.6 exceeds 9.21, reject the null hypothesis at the 0.01 level: the spread of litter is significantly uneven.

Evaluation: chi-square uses counts, not percentages or means. Expected frequencies should be at least 5 in each category. It shows that a difference exists, not why.

Significance tests

A significance test asks how likely a result is to have arisen by chance. Geographers usually accept the 0.05 level (95% confidence) and treat the 0.01 level as strong evidence. If the result exceeds the critical value, reject the null hypothesis. Then explain it geographically.

3.5.1.5 ICT skills

• Remotely sensed data: images from satellites, aircraft or drones, such as Landsat or Copernicus Sentinel imagery. Useful for land-use change and vegetation cover over large areas. Limitations: cloud cover, resolution, and the need for ground checking.
• Electronic databases: census tables, river flow records, crime or health statistics. Large and consistent, but collected for other purposes.
• Crowd sourcing and ‘big data’: volunteered data such as OpenStreetMap edits, geotagged social media posts or phone location records. Huge and near real-time, but unrepresentative (people without smartphones are missed), with privacy concerns.

3.5.2 Fieldwork: what the specification requires

Every student must carry out a personal fieldwork investigation. Ideally it is practical. Where that is not possible, the specification allows alternatives, such as enquiries in the school grounds, interviews in school or at home, or raw unprocessed data from your teacher that you plan with and sample from. The method must go beyond being handed pre-processed data. There must be planning and selection.

3.5.2.1 Planning and preparation

Define a clear aim and a testable hypothesis or research question, and justify it with geographical theory. Choose a location that is safe, accessible and suitable. Research secondary sources first: maps, census data, previous studies. Carry out a risk assessment: identify hazards (traffic, water, weather, lone working), judge their likelihood and severity, and set control measures.

3.5.2.2 Collection of data

Justify each method in terms of the question. Primary data is collected by you; secondary data by others. The specification names three sampling types:

• Random: every point has an equal chance (random number tables or grid co-ordinates). Avoids bias but can miss parts of the area.
• Systematic: at regular intervals, such as every 50 m along a transect. Good coverage but can miss or repeat a regular pattern.
• Structured/stratified: the sample is split in proportion to known sub-groups, such as land-use zones or age groups. Representative but needs prior knowledge of the groups.

3.5.2.3 Presentation of data

Choose maps, graphs, photographs, field sketches, tables or interview quotations, and justify the choice. A scatter graph suits two continuous variables; a located proportional symbol map suits values at named sites.

3.5.2.4 Analysis of data

Use statistical techniques where appropriate, and other techniques where they are not. Qualitative data, such as interview responses, can be coded into themes and counted, or quoted to show a range of views.

3.5.2.5 Conclusions and evaluation

Link conclusions back to the aim, and accept or reject the hypothesis. Consider fieldwork ethics: consent, anonymity, not damaging the environment, and how presenting data might affect a place or its people. Show how your understanding of the place and the theory grew. Evaluate every stage, suggest improvements, and say how the enquiry could be taken forward.

Worked example: an evaluation paragraph

Question type: evaluate the reliability of a data collection method in your own investigation.

I used systematic sampling, counting pedestrians for 10 minutes at 11 sites every 100 m from the town centre (knowledge). This gave even coverage of the transect, so I could see how footfall changed with distance (application). However, every count was taken on one weekday morning, so the results show one moment, not typical footfall; one site was beside a school at the start of the day, which inflated its count (analysis). The method was reliable in being repeatable, but less valid as a measure of typical footfall. Repeat counts at three times on two days, and a stratified sample that includes each land-use zone, would have made the conclusion more secure (evaluation).

Common errors

• Making a proportional circle’s radius, not its area, proportional to the value.
• Ranking the two variables in opposite directions in Spearman’s rank.
• Using percentages instead of raw counts in chi-square.
• Reading the wrong row (df) of the critical value table.
• Describing a sampling method without justifying it for the question.
• Treating a risk assessment as a list of hazards with no control measures.

Where to go next

Test yourself with the practice questions. Apply the skills to other units with the physical geography practice set and the global systems and governance guide.

Official syllabus

This guide is based on the OxfordAQA International AS and A-level Geography (9635) specification, Version 3.2, section 3.5 (Unit 5: Geographical and Fieldwork skills), the scheme of assessment in section 4 and the entry codes in section 5.1. Check the current version on the OxfordAQA website.