Quantitative Place Research

Spec mapping (AQA 7037): Paper 2 (Human), §3.2.2 Changing Places + Geographical Skills — quantitative and statistical skills for place study: sampling; use of census, IMD and geospatial data; measures of central tendency and dispersion; statistical tests including Spearman's rank correlation and the chi-square test; data presentation; significance and confidence. This lesson is deliberately AO3-skills heavy (interpreting, manipulating and statistically testing place data) with supporting AO1 (knowing the datasets and tests) and AO2 (judging the limits of quantitative evidence). The named statistical tests are directly examinable and frequently appear in NEAs (Lesson 10), so this lesson works two of them fully — Spearman's rank and chi-square — with complete arithmetic and significance interpretation.

Where qualitative methods (Lesson 7) capture the meaning and lived experience of place, quantitative methods provide measurable, comparable, testable data — the census tells you what a place is like in numbers, supporting comparison between places and over time, and allowing claims to be tested for statistical significance rather than merely asserted. The signature A-Level skill here is not just to describe a dataset but to manipulate it (calculate change, central tendency, dispersion), test relationships statistically (Spearman, chi-square), and evaluate the limits of the numbers (ecological fallacy, MAUP, false precision). The best place studies are mixed-methods: quantitative data to establish the measurable, qualitative data to interpret the lived — a theme that recurs from Lessons 7–8 and culminates in the NEA (Lesson 10).

---

Census Data Analysis

The UK Census is conducted every ten years (most recently in 2021) and provides the most comprehensive quantitative dataset for analysing places.

Key Variables Available at Small-Area Level

Category	Variables	Use in Place Studies
Population	Total population, age structure, sex ratio	Identifying demographic profiles; ageing/youthful places
Ethnicity	Ethnic group, national identity, country of birth	Analysing diversity and segregation
Housing	Tenure (owned/rented), type (detached/flat), number of rooms	Comparing housing markets; identifying deprivation
Employment	Economic activity, occupation, industry, hours worked	Assessing economic health; identifying dominant sectors
Education	Qualifications, student status	Proxy for social class and human capital
Health	Self-assessed health, disability, unpaid care	Identifying health inequalities
Transport	Mode of travel to work, car ownership	Assessing connectivity and mobility
Deprivation	Derived from multiple indicators (see IMD below)	Identifying and comparing levels of disadvantage

Geographical Units

Census data is available at multiple spatial scales:

• Output Areas (OAs) — smallest unit; approximately 125 households (England and Wales)
• Lower Super Output Areas (LSOAs) — approximately 650 households; the unit used for the Index of Multiple Deprivation
• Middle Super Output Areas (MSOAs) — approximately 3,000 households
• Local Authority Districts — administrative areas
• Regions — e.g., North West, South East

Using Census Data in Place Studies

1. Place profiles — compile data for a specific area to create a quantitative portrait
2. Comparison — compare two or more places on the same variables (e.g., contrasting a deprived urban neighbourhood with an affluent rural village)
3. Change over time — compare census data from different years (2001, 2011, 2021) to identify trends
4. Spatial patterns — map census data to identify clusters, gradients, and anomalies

Accessing Census Data

• NOMIS (Office for National Statistics) — official source for census and labour market data
• Census maps — interactive mapping tools showing spatial distribution of variables
• DataShine — visual tool for exploring census data at neighbourhood level

---

Index of Multiple Deprivation (IMD)

The Index of Multiple Deprivation is the official measure of relative deprivation in England. It ranks every LSOA (32,844 in total) from most deprived (1) to least deprived.

The Seven Domains

Domain	Weight	What It Measures
Income	22.5%	Proportion of people on low incomes (benefits, tax credits)
Employment	22.5%	Proportion of working-age people involuntarily excluded from the labour market
Education, Skills and Training	13.5%	Educational attainment and skills among children and adults
Health Deprivation and Disability	13.5%	Premature death, illness, and disability
Crime	9.3%	Rates of recorded crime (violence, burglary, theft, criminal damage)
Barriers to Housing and Services	9.3%	Affordability, overcrowding, homelessness; distance to services
Living Environment	9.3%	Housing quality (central heating, condition); outdoor environment (air quality, road accidents)

Using the IMD in Place Studies

• Identify the most and least deprived areas within a local authority or region
• Compare domains — an area may rank highly deprived on income but not on crime, revealing a more nuanced picture
• Track change — compare IMD rankings over time (published in 2004, 2007, 2010, 2015, 2019)
• Map deprivation — GIS mapping of IMD data reveals spatial concentrations of disadvantage

Limitations of the IMD

• Measures relative deprivation — it ranks areas against each other rather than measuring absolute poverty
• Ecological fallacy — not everyone in a deprived area is deprived, and not everyone in a wealthy area is wealthy
• LSOA boundaries are arbitrary — they may split coherent communities or combine very different areas
• Does not capture subjective experience — a statistically deprived area may have strong community bonds and high resident satisfaction
• Updated irregularly — data can be several years out of date

---

GIS Mapping

Geographic Information Systems (GIS) are digital tools for storing, analysing, and visualising spatial data. GIS is increasingly central to place research.

Applications in Place Studies

• Choropleth maps — shading areas by data values (e.g., IMD deciles, ethnic diversity, house prices)
• Dot density maps — showing the distribution of specific features (e.g., fast food outlets, listed buildings, crime incidents)
• Buffer analysis — identifying areas within a specified distance of a feature (e.g., households within 1km of a GP surgery)
• Overlay analysis — combining multiple data layers to identify spatial relationships (e.g., overlaying flood risk with deprivation data)
• Change detection — comparing datasets from different time periods to identify spatial change

GIS Software

• ArcGIS — industry-standard software; available in many schools through ESRI licensing
• QGIS — free, open-source alternative
• Google Earth Pro — free; useful for historical imagery comparison
• Digimap — Ordnance Survey data available to UK educational institutions

Evaluation

Strengths: Powerful visualisation; can handle large datasets; reveals spatial patterns invisible in tabular data; increasingly accessible Limitations: Requires technical skills; data quality varies; maps can be misleading (choice of classification, colour scheme, boundaries affects interpretation); may create an illusion of precision

---

Geodemographic Classification: ACORN and MOSAIC

Geodemographic classification systems categorise neighbourhoods into types based on multiple variables, using census data and commercial datasets.

ACORN (CACI)

ACORN (A Classification Of Residential Neighbourhoods) classifies every UK postcode into one of:

• 6 categories (e.g., "Affluent Achievers," "Rising Prosperity," "Comfortable Communities," "Financially Stretched," "Urban Adversity," "Not Private Households")
• 18 groups (e.g., "Lavish Lifestyles," "Executive Wealth," "Starting Out," "Difficult Circumstances")
• 62 types (e.g., "Wealthy Executives," "Thriving Families," "Student Life," "Struggling Estates")

MOSAIC (Experian)

MOSAIC is a similar system that classifies the UK population into:

• 15 groups (e.g., "Prestige Positions," "Country Living," "Senior Security," "Urban Cohesion," "Rental Hubs")
• 66 types with detailed profiles describing typical characteristics, behaviours, and preferences

How They Work

Both systems use cluster analysis — statistical techniques that group postcodes with similar characteristics. Variables include:

• Census data (demographics, housing, employment)
• Consumer data (spending patterns, credit history)
• Lifestyle data (media consumption, leisure activities)
• Property data (house prices, housing type)

Applications in Place Studies

• Quick profiling — rapidly characterise a neighbourhood using its ACORN/MOSAIC classification
• Comparison — compare the geodemographic profiles of contrasting places
• Marketing and planning — used by businesses, local authorities, and public services to target resources

Evaluation

Strengths:

• Provide accessible, intuitive summaries of neighbourhood character
• Useful for rapid comparison and profiling
• Combine multiple data sources into a single classification

Limitations:

• Stereotyping — reduce complex communities to simple labels
• Ecological fallacy — assume all residents of a postcode share the same characteristics
• Commercial bias — designed primarily for marketing purposes, not academic analysis
• Privacy concerns — combining datasets to profile individuals and neighbourhoods
• Static — classifications may not keep pace with rapid neighbourhood change

---

Sampling: getting representative data

Before any statistic can be trusted, the sample must be sound. A sample is a subset of a population studied because the whole cannot be. The choice of sampling strategy directly affects validity and is itself examinable.

Strategy	How it works	Strength / weakness
Random	Every unit has an equal chance (e.g. random-number selection of houses)	Unbiased if a sampling frame exists; may miss small subgroups; can cluster by chance
Systematic	Select at fixed intervals (every 5th house; every 50 m on a transect)	Even spatial coverage; simple; risks bias if the interval coincides with a pattern
Stratified	Divide population into subgroups, sample each in proportion	Ensures subgroups (age, tenure) are represented; needs prior knowledge of strata
Systematic along a transect	Sampling at intervals on a line from CBD outward	Ideal for testing distance-decay (e.g. bid-rent, environmental quality)

Sample size matters too: larger samples reduce the influence of anomalies and increase confidence, but cost more time. A core AO3 evaluation point is that biased or too-small samples invalidate even a correct statistical test — for example, surveying perceptions only on a weekday morning will over-represent the retired and unemployed, distorting the result regardless of how the numbers are then processed.

---

Measures of central tendency and dispersion (worked)

Consider house prices (£000s) for a sample of 7 properties in a gentrifying ward: 180, 210, 215, 230, 260, 290, 615.

Central tendency. The mean is:

$\bar{x} = \frac{180 + 210 + 215 + 230 + 260 + 290 + 615}{7} = \frac{2000}{7} \approx 285.7$xˉ=7180+210+215+230+260+290+615​=72000​≈285.7

The median (the middle value of the 7 ordered figures) is the 4th value, 230. The mean (285.7) is far above the median (230) because the single £615k luxury conversion skews the mean upward.

Explain/evaluate: This is a textbook reason to prefer the median for skewed data such as house prices and incomes: the mean is dragged by outliers and would overstate the "typical" property, whereas the median resists the skew. Reporting the mean alone here would mislead — a direct application of the "false precision" critique below.

Dispersion. The range is 615 − 180 = 435 (huge, again driven by the outlier). The interquartile range (IQR) — the spread of the middle 50% — is more robust: with lower quartile ≈ 210 and upper quartile ≈ 290, the IQR ≈ 80, showing that most properties actually cluster tightly and the apparent huge spread is an outlier effect. Standard deviation would likewise be inflated by the £615k value. The lesson: always pair a measure of central tendency with a measure of dispersion, and prefer outlier-resistant measures (median, IQR) for skewed place data.

---

Comparative analysis: the location quotient (worked)

A location quotient (LQ) measures how concentrated a variable is in an area relative to a larger reference area:

$\text{LQ} = \frac{\text{percentage of variable in the area}}{\text{percentage of variable in the reference area}}$LQ=percentage of variable in the reference areapercentage of variable in the area​

Suppose 36% of a neighbourhood's residents are aged 65+, against 18% nationally:

$\text{LQ} = \frac{36}{18} = 2.0$LQ=1836​=2.0

An LQ of 2.0 means the neighbourhood has twice the national concentration of older residents — a strongly ageing place (consistent with the counter-urbanised retirement villages of Lesson 2). An LQ of 1.0 means exactly the national average; below 1.0 means under-represented. LQs are a quick, powerful way to characterise and compare places (ethnic concentration, social housing, particular industries) — but, like all area statistics, they are vulnerable to the ecological fallacy and the MAUP discussed below.

---

Statistical test 1: Spearman's rank correlation (fully worked)

Spearman's rank correlation coefficient ($r_s$rs​) tests the strength and direction of the relationship between two ranked (ordinal) variables. It is the workhorse statistical test of place and urban fieldwork. The formula is:

$r_s = 1 - \frac{6\sum d^2}{n(n^2 - 1)}$rs​=1−n(n2−1)6∑d2​

where $d$d is the difference between the two ranks for each pair and $n$n is the number of pairs. The result lies between −1 (perfect negative correlation) and +1 (perfect positive correlation), with 0 meaning no correlation.

Worked example

Hypothesis: "There is a negative correlation between distance from the city centre and average house price" — i.e. as you move outward, prices fall (the bid-rent prediction of Lesson 1). We sample 8 sites and record the raw data, then rank each variable separately. We rank distance 1 = nearest and price 1 = highest, so that a negative $r_s$rs​ would confirm "near = expensive, far = cheap".

Site	Distance (km)	Distance rank (1 = nearest)	Price (£000s)	Price rank (1 = highest)	$d$d	$d^2$d2
A	0.4	1	540	1	0	0
B	0.9	2	500	2	0	0
C	1.6	3	470	3	0	0
D	2.5	4	410	5	−1	1
E	3.3	5	430	4	1	1
F	4.8	6	360	6	0	0
G	6.2	7	300	8	−1	1
H	7.5	8	320	7	1	1
					$\sum d^2 =$∑d2=	4

Here $n = 8$n=8 and $\sum d^2 = 4$∑d2=4. Substituting into the formula: