Skip to main content
ScotlandGeographySyllabus dot point

How do you use chi-squared, linear regression and nearest neighbour analysis in Advanced Higher Geography?

Inferential techniques: chi-squared analysis for association, linear regression for the relationship between two variables, and nearest neighbour analysis for settlement or point patterns.

How to use three inferential techniques in SQA Advanced Higher Geography: chi-squared analysis to test association between categories, linear regression to model the relationship between two variables, and nearest neighbour analysis to measure how clustered or dispersed a point pattern is.

Generated by Claude Opus 4.814 min answer

Reviewed by: AI editorial process; not yet individually human-reviewed

Have a quick question? Jump to the Q&A page

Jump to a section
  1. What this key area is asking
  2. The three techniques
  3. Interpreting nearest neighbour
  4. A routine for an inferential test
  5. Examples in context
  6. Try this

What this key area is asking

Beyond correlation, the spec names three more inferential techniques: chi-squared analysis (association between categories), linear regression analysis (modelling the relationship between two variables), and nearest neighbour analysis (the clustering of a point pattern). For each you should know what it tests, the data it requires, how to interpret the result, and the importance of significance.

The three techniques

These techniques answer different questions: is there an association, what is the relationship, and how is the pattern arranged in space. Choosing the right one depends on the data and the question.

  • Chi-squared. Frequencies (counts); tests significant association between categories.
  • Linear regression. Two interval variables; best-fit line to model and estimate.
  • Nearest neighbour. Point locations; index of clustering from 0 to about 2.15.

Interpreting nearest neighbour

The index turns a settlement or point pattern into a single number describing how clustered or spread it is. A value near 0 means clustering (perhaps around a resource); near 1 means randomness; near 2.15 means even spacing. The result depends on the area boundary chosen, which is a key limitation to note.

A routine for an inferential test

  1. Match technique to question. Association points to chi-squared; relationship to regression; spatial pattern to nearest neighbour.
  2. Check the data. Frequencies for chi-squared; interval data for regression; point locations for nearest neighbour.
  3. Calculate and interpret. Compute the statistic and read what it means.
  4. Test significance. Compare with the critical value before concluding, and avoid claiming cause.

Examples in context

Try this

Q1. What does a nearest neighbour index (Rn) of about 1 indicate? [1 mark]

  • Cue. A random point pattern.

Q2. What kind of data does chi-squared analysis require? [1 mark]

  • Cue. Frequency (count) data, not percentages.

Exam-style practice questions

Practice questions written in the style of SQA exam questions on this dot point, with worked answer explainers. The year tag is the paper they imitate, not the source.

SQA AH data5 marksExplain what nearest neighbour analysis measures and how the result is interpreted.
Show worked answer →

Nearest neighbour analysis measures how clustered, random or regular a pattern of points (such as settlements) is. It compares the observed mean distance between each point and its nearest neighbour with the distance expected in a random pattern, producing an index (Rn) from 0 to about 2.15.

A full answer states what it measures, explains the index: Rn near 0 means strongly clustered, Rn near 1 means random, and Rn near 2.15 means regularly (evenly) spaced. The strongest answers note the limits (the result depends on the boundary chosen for the area), check significance, and interpret the value in context, for example settlements clustered near a river.

SQA AH data4 marksExplain when chi-squared analysis is used and what it tests.
Show worked answer →

Chi-squared analysis tests whether there is a significant association between categories, using frequency (count) data. It compares observed frequencies with the frequencies expected if there were no association, and tests whether the difference is statistically significant.

Strong answers state the data requirement (frequencies or counts, not percentages), explain that it compares observed with expected frequencies, and stress checking the chi-squared value against the critical value for the degrees of freedom and probability level before concluding. They give a geographical example, such as testing whether land use is associated with distance from the centre, and note it shows association, not cause.

Related dot points

Sources & how we know this