Skip to content
AP Human Geography · Thinking Geographically

MAUP

MAUP appears in 5 reviewed AP Human Geography questions and sits in Thinking Geographically. Below are worked examples showing exactly how the exam tests it, with the answer marked and every distractor explained.

How MAUP is tested

Example 1Medium
A researcher wants to study the spatial association between fast-food restaurants and rates of obesity at the county level in the United States. She finds a strong positive correlation. Which critical limitation should she acknowledge?
  1. ASpatial association is purely descriptive and cannot incorporate quantitative measures such as correlation coefficients or regression
  2. BSpatial association does not establish causation; confounding variables such as income and education may explain the patternCorrect
  3. CBecause the correlation is statistically strong, it confirms that fast-food access is the direct cause of the obesity rates observed
  4. DCounty-level data is too fine-grained and should be aggregated to the state level before any valid spatial pattern emerges
  5. EA positive correlation means the two phenomena are spatially dispersed, so they rarely occur together in the same counties

Why B is right

Spatial association identifies co-occurrence of phenomena but does not establish causation. The observed correlation between fast-food density and obesity could be driven by confounding variables like poverty, education levels, or lack of healthcare access that are associated with both fast-food prevalence and obesity rates.
Why the other 4 choices miss
A
Incorrect because spatial association routinely relies on quantitative methods, including correlation coefficients, regression analysis, and geographic information systems.
C
Incorrect because a strong correlation does not confirm causation; the spatial co-occurrence of fast-food restaurants and obesity could reflect confounding factors rather than a direct causal link.
D
Incorrect because aggregating to a coarser scale reduces analytical precision and can introduce the modifiable areal unit problem; county-level data is an appropriate scale for this analysis.
E
Incorrect because it reverses the meaning of a positive spatial association: a strong positive correlation means counties with more fast-food restaurants also tend to have higher obesity rates, so the two co-occur rather than being dispersed. Students who confuse correlation direction with spatial distribution patterns (clustered vs. dispersed) may choose this.
Example 2Medium
The table shows hypothetical data on the same set of votes cast in one city, grouped into spatial units in four different ways by a political analyst. | Grouping scheme | Number of units | Units with a Republican majority | Units with a Democratic majority | Citywide Republican share of the vote | |---|---|---|---|---| | Large districts | 4 | 4 | 0 | 54% | | Medium districts | 8 | 5 | 3 | 54% | | Census tracts | 16 | 8 | 8 | 54% | | Precincts | 32 | 15 | 17 | 54% | Which of the following conclusions is best supported by the data in the table?
  1. ARepublican support in the city fell from the 4-district scheme to the 32-precinct scheme, as fewer units recorded Republican majorities
  2. BSince all 4 large districts had Republican majorities, a majority of voters in each smaller unit also voted Republican
  3. CThe citywide Republican share is the same in every scheme, yet the count of Republican-majority units changes as the same votes are split into smaller unitsCorrect
  4. DUnder the 32-precinct scheme, Democrats won a majority of the citywide vote because they carried 17 of the 32 precincts
  5. EThe 16-tract scheme gives the most accurate result, because it is the only scheme that splits the units evenly between parties

Why C is right

In the table, the citywide Republican share is 54% in all four rows, so the underlying votes never change. What changes is the unit-level result: Republicans carry 4 of 4 large districts but only 15 of 32 precincts. When the analysis depends on how boundaries are drawn and how large the units are, rather than on the data themselves, that is the modifiable areal unit problem (MAUP).
Why the other 4 choices miss
A
This misreads the table. The number of Republican-majority units falls (4 to 15 of 32 as a proportion), but the citywide Republican share stays at 54% in every row. Support did not change, only the way the votes were grouped.
B
This is the ecological fallacy, and the table contradicts it: under the precinct scheme, 17 of the 32 smaller units had Democratic majorities.
D
This confuses units won with votes won. The Precincts row still shows a 54% citywide Republican share, so carrying more precincts did not give Democrats a citywide majority.
E
An even 8-8 split is not evidence of accuracy. The table gives no basis for ranking schemes, and every scheme reports the same 54% citywide share; the 16-tract scheme is just one more way of grouping the same votes.

Studied alongside

Other concepts from Thinking Geographically.