[DSRP Evidence](https://dsrpevidence.org/)

# The Modifiable Areal Unit Problem in Multivariate Statistical Analysis

## Details

**Authors** Fotheringham, A. S. and Wong, D. W. S.

**Year** 1991

**Publisher** Environment and Planning A: Economy and Space

**Kind of work** article

**Discipline** Geography

**Secondary disciplines** Statistics

**Applied** false

[Read it at the publisher](https://doi.org/10.1068/a231025) 
10.1068/a231025

## In authors' words

### What they set out to do (purpose)

To extend examination of the modifiable areal unit problem from univariate and bivariate statistics to multivariate statistical analysis.

### Who or what was studied (sample)

Areal-unit data analyzed with a multiple linear regression model and a multiple logit model.

### How they did it (methods)

Parameter estimates were compared across changes in the scale of aggregation and in the zoning system.

### What they found (results)

The authors find that parameter estimates from both models are sensitive to the scale of aggregation and to the zoning system used. The intensity and effects of the problem are essentially unpredictable in the multivariate case. They conclude that the results are strong evidence of the unreliability of any multivariate analysis undertaken with data from areal units.

## Commentary

### In short

The same underlying data, grouped into larger or differently drawn wholes, yield different statistical relationships. The finding shows that the way parts are aggregated into units changes the relationships observed among variables.

**Patterns it shows** S

Shows it makes a difference

**Added** 2026-10-10

**How to cite this** Fotheringham, A. S. and Wong, D. W. S. (1991). The Modifiable Areal Unit Problem in Multivariate Statistical Analysis. Environment and Planning A: Economy and Space.