Skip to content

AP Human Geography · technique

How to attack stimulus-based multiple choice

On this page · 5 sections

Section I is 60 multiple-choice questions in 60 minutes, worth 50% of your composite score. Every question has five answer choices, A through E. There is no penalty for a wrong answer, so nothing should be left blank.

Roughly 30 to 40 percent of those questions hang off a stimulus — call it 18 to 24 of the 60. College Board draws them from maps, tables, charts, graphs, images, infographics, and landscapes. A useful way to sort them as you study: quantitative sources, such as a choropleth or flow map, a data table, a population pyramid, a line graph, or a cartogram, and qualitative ones, such as a satellite image, a landscape photograph, an infographic, a model diagram, or a short quoted passage. That split is a study frame, not a published proportion. Stimuli often arrive in sets — one figure, two or three questions. Plenty are standalone: one figure, one question, one point. When you do get a set, you pay the reading cost once and collect two or three points from it.

Scoring is flat. Reading a number off an axis earns the same single raw point as predicting a government's policy response. So a stimulus question is not harder by definition; it is a question where part of the information has been handed to you instead of memorized. The usual failure is not ignorance. It is answering from memory when the data was sitting in front of you, or importing outside knowledge into a stem that told you to stay inside the figure.

The steps

  1. 1

    Read the labels before you read the data

    Spend ten seconds on the frame: title, axis labels, units, legend intervals, year, source, and the geographic unit being mapped (country, state, metro area, census tract).

    Units decide answers. A map of population growth in absolute numbers and a map of natural increase in percent put different countries on top. India adds more people per year than any other country. Niger has one of the world's highest total fertility rates and one of the highest rates of natural increase, so it sits near the top of the percent map. On a choropleth, confirm whether darker means more or less; the legend sometimes runs the other way. On a cartogram, area encodes the variable, not land, so a physically small country can dominate the image. On a log axis, equal spacing means equal multiplication, not equal addition.

    Read two things, not one. The geographic unit tells you the scale of analysis, because scale of analysis is the level at which the data are grouped: country = national, state or province = subnational, county or census tract = local. The extent of the figure tells you only how much of the world is being displayed. A world choropleth shaded by country is national-scale data shown at a global extent, so its scale of analysis is national, not global. The CED tests scale of analysis by name, so read both and keep them apart.

    A pattern can reverse when you switch scales. The United States has a low national population density and extreme local density in Manhattan. A country can post rising national GDP per capita while its interior provinces fall behind its coastal ones. So read the stem for the scale it demands. A choice that is true of the world can be false of the one region the question asked about. Read the frame first and you answer from the right variable, at the right scale.

  2. 2

    Say the pattern in one sentence before you look at the choices

    Force a prediction: what varies, across what space, in which direction. Say it out loud when you practice at home, and say it silently as one complete sentence in the proctored room.

    "Total fertility is highest across Sub-Saharan Africa and lowest in Europe and East Asia." "Land value drops steeply within two miles of the CBD, then flattens." "Migration arrows converge on three destination states." A prediction turns the answer choices into one match and a set of mismatches. Skip it, and you read a screen of confident statements written in course vocabulary and pick whichever sounds most like your teacher. That is exactly what distractors are engineered to sound like. The prediction also catches your own misreading: if you cannot state the pattern in one sentence, you have not actually read the figure yet, and no amount of staring at the choices will fix that.

  3. 3

    Classify the verb: read it, name it, or explain it

    Every stimulus question is one of three tasks, and they have different rules.

    Read it — "According to the table, which country has the highest percentage of its population over age 65?" The answer lives entirely inside the figure. This is the only task you can finish from the figure alone. You still need to know what the label means, but you need nothing from outside about causes, and importing it is a trap here.

    Name it — "The pattern shown best illustrates which of the following?" You attach a course term to the pattern: hierarchical diffusion, primate city, distance decay, Stage 3 of the demographic transition. Here the outside knowledge you supply is definitional. The figure never supplies the vocabulary or a model's defining criteria; you bring those.

    Explain or predict it — "Which best explains the pattern shown?" or "...most likely consequence of the trend shown?" The figure sets the situation; the causal content comes from the course.

    So the split is not knowledge versus no knowledge. Name-it and explain-it both require course content, in different kinds: name-it needs definitions and model criteria, explain-it needs causes and consequences. Mis-sorting these is the most common way a well-prepared student loses stimulus points. Underline the verb in the stem before you touch the choices.

  4. 4

    Put the figure on the model shelf

    Most APHG stimuli are a real-world instance of one of the course's named models. Naming the model tells you which vocabulary should appear in the right answer.

    Unit 1. On any map, name the projection and the map type first. Mercator is conformal: it keeps compass direction and local shape. But it inflates area toward the poles, so Greenland looks the size of Africa. Gall-Peters keeps area honest and stretches shape instead. Robinson is a compromise: it distorts area, shape, and distance a little rather than keeping any one of them exact. Goode's homolosine is not a compromise. It is an interrupted equal-area projection, spliced from the sinusoidal near the equator and the Mollweide toward the poles. The oceans are cut open so the landmasses keep true area and still look like themselves, which is why atlases use it for world distribution maps. Then name the map type: a choropleth shades a rate by unit, an isoline map joins points of equal value, a graduated symbol grows with the value, and a cartogram sizes area by the value. A dot map places one dot for each quantity the legend names, such as 1 dot = 10,000 people. Read that dot value before you compare densities. Regions sort three ways: formal shares a measurable trait, functional has a node and a zone of influence, perceptual or vernacular lives in people's heads.

    Unit 2. Population pyramid to demographic transition stage, dependency ratio, and any narrow cohort left by a war or a birth-control policy. Migration flow map to Ravenstein, push-pull factors, and the gravity model's size-over-distance logic. A line graph of food supply against population to Malthus versus Boserup. Malthus has food rising slower than people; Boserup has that same population pressure pushing farmers to invent better methods.

    Unit 3. Language-family or religion map to hearth plus diffusion type: relocation diffusion travels with the migrants themselves, while expansion diffusion spreads outward from the hearth and splits into contagious, hierarchical, and stimulus. Landscape photograph of the built environment to folk versus popular culture, placelessness, ethnic enclave, and sequent occupance.

    Unit 4. Boundary map to antecedent, subsequent, superimposed, or relict, plus the defined-delimited-demarcated sequence. State-shape map to compact, elongated, fragmented, perforated, or prorupted. Electoral-district map to gerrymandering and redistricting. Membership map of an organization like the EU or ASEAN to supranationalism versus devolution and centripetal versus centrifugal forces.

    Unit 5. Concentric rings of land use around a market to von Thünen, where transport cost and perishability set the ring order. Aerial or satellite photo of farmland to survey system: long lots are narrow strips reaching back from a river or road, metes and bounds follows natural landmarks into irregular shapes, and township and range stamps a grid of squares across the land.

    Unit 6. City-size distribution to rank-size versus primate. A honeycomb of hexagons to Christaller's central place theory, where threshold is the minimum number of customers a service needs to survive and range is the farthest people will travel to reach it. Urban land-use map to Burgess rings, Hoyt sectors, multiple nuclei, the galactic edge-city pattern, the Griffin-Ford Latin American spine, McGee's Southeast Asian model built around an old colonial port zone instead of a single CBD, or the Sub-Saharan African model with its three centers: colonial, traditional, and market. A graph of land value falling with distance from the CBD is bid-rent. The users who earn the most per acre outbid everyone else for the center, so land value drops steeply outward and land use changes with it. That same auction logic orders von Thünen's farm rings.

    Unit 7. Weighted locational triangle to Weber's least-cost model, bulk-reducing versus bulk-gaining industries. Development indicators to Rostow's stages, Wallerstein's core-semiperiphery-periphery, HDI, and the Gender Inequality Index.

    Name the model in five seconds, then check the choices for its terms used correctly.

  5. 5

    Eliminate against the stimulus first, the vocabulary second

    Run two passes over the answer choices. First pass: does this choice contradict what the figure actually shows? Cross it out even when it is a true statement about the world. On a stem containing "according to," "shown," or "in the map," correctness means supported by this figure, and the wrong choices are often accurate geography that this particular figure does not demonstrate.

    Second pass on the survivors: is the technical term used correctly? Distractors are built by taking the right answer and swapping one word for its near neighbor — contagious for hierarchical, density for distribution, formal for functional region, devolution for supranationalism, site for situation. APHG choices frequently attach a reasoning clause after a comma, and the planted error is often in that clause, not in the term at the front. Read the whole choice to its last word.

  6. 6

    Run the clock: one read per figure, two passes, nothing left unanswered

    You average one minute per question. Sets are where you bank time. A three-question set gets a 3:00 budget. It costs about 25 seconds of stimulus reading plus roughly 35 seconds per question, so 25 + 105 = 2:10. That banks about 50 seconds. Spend the surplus on isolated hard items later.

    A standalone stimulus question gets 1:00 and no discount, because the reading cost is not spread across anything. Same 25 seconds of reading plus 35 seconds of answering, and you break even. So cap the read at 25 seconds whether the figure carries one question or three. Do not study a map you are only being asked one thing about.

    Set a hard 90-second ceiling on any single question. At 90 seconds, pick the best surviving choice, flag it, and move. On pass two you will often catch the units error in five seconds, because you are no longer locked into the reading that trapped you.

    You take this exam in Bluebook. The flag control is labeled "Mark for Review" — use it on anything you skip, and use the answer eliminator to cross off dead choices. Before time expires, open the review page. It lists every question in the section, and it shows which ones are still blank and which ones you flagged. Use it to confirm nothing is unanswered and to jump back to flagged items. Answer-sheet misalignment is not a risk on a digital exam, so the only self-inflicted zero left is a blank. A blank scores zero every time. A guess costs you nothing. A blind guess among five choices lands 20% of the time, and knocking out two first raises that to 33%.

Worked through

**Example 1 — a data table**

Stimulus: a table titled "Selected Demographic Indicators, 2023," four unidentified countries. The table is illustrative. It was written for this walkthrough, and the shape reflects the real demographic transition, but the numbers are invented for instruction and are not official estimates. Read them; do not memorize them.

| Country | TFR | % under 15 | % over 65 | Natural increase | Life expectancy | |---|---|---|---|---|---| | W | 5.1 | 43 | 3 | +2.7% | 61 | | X | 2.4 | 29 | 6 | +1.4% | 71 | | Y | 1.7 | 18 | 18 | +0.1% | 81 | | Z | 1.2 | 13 | 28 | −0.4% | 84 |

Labels first. The units are mixed on purpose: TFR is children per woman, two columns are shares of total population, and natural increase is an annual rate of change that excludes migration. One year, so this is a snapshot, not a trend.

Pattern in one sentence: moving from W to Z, fertility falls, the population ages, natural increase turns negative, and life expectancy climbs. That is the demographic transition laid out across four rows.

Question 1. "Which country's data are most consistent with Stage 2 of the demographic transition model?" The choices each pair a country with a reason, and W appears twice. One W choice says birth and death rates are both low. The other says birth rates stay high while death rates fall.

Verb check: name it. You supply the model's criteria; the table supplies the numbers. Stage 2 is the rapid-growth stage — death rates have fallen while birth rates stay high, so natural increase spikes and the base of the pyramid is very wide. W has 43% under 15, the highest TFR, the highest growth rate, and the lowest life expectancy in the table. Answer: the W choice that cites high birth rates alongside falling death rates. The other W choice names the right country but describes Stage 4, so the reason clause sinks it. Notice what you did not need. You never identified the country.

Question 2. "Country Z's government is most likely to respond to the pattern shown by adopting which of the following?" Among the choices on offer: pronatalist incentives paired with expanded labor immigration; antinatalist campaigns promoting smaller family size; large-scale primary school construction; restrictions on rural-to-urban migration; export-processing zones to absorb surplus rural labor.

Verb check: predict it. Causal course content is now required. Z's births no longer replace its deaths — natural increase is −0.4% — so absent immigration its population will shrink. Be precise about what that column does and does not say. Natural increase excludes migration, so whether Z's total population is actually falling depends on net migration, which this table does not report. What the table does establish is natural decrease plus 28% over 65 against a working-age share of 59% — an old-age dependency ratio near 47 per 100. A TFR of 1.2 will push that ratio higher as today's small cohorts reach working age.

Governments in that position pay people to have children and open channels for foreign workers: Japan's childcare subsidies, Hungary's income-tax exemption for mothers of four, Germany's recruitment of skilled migrants. Answer: pronatalist incentives paired with labor immigration. The immigration half is exactly the flow the natural-increase column leaves out.

The traps are worth naming. School construction is a genuine policy response — for W, whose 43% under 15 is a classroom-shortage problem. Right fact, wrong row. The antinatalist choice inverts the sign: Z's TFR of 1.2 is already far below the 2.1 replacement level, so suppressing births deepens the problem. Migration restrictions belong to a different topic and have no support in the table. Export-processing zones are a real development strategy, but they solve a labor surplus. Z's problem is the opposite: too few working-age people, not too many. One stimulus, one reading, two points, about two minutes.

**Example 2 — a choropleth map**

Stimulus: a world choropleth titled "Percent of Population Living in Urban Areas, 2020," shaded by country in five classes, darkest = highest percentage.

Frame first. The variable is a percentage of each country's own population, the unit of analysis is the country, so the scale of analysis is national, displayed across a global extent. Darker means more urban here — check that, because legends sometimes run the other way.

Pattern in one sentence: South America, North America, Western Europe, and the Gulf states shade dark, while most of Sub-Saharan Africa and South Asia shade light.

Question 1. "According to the map, which of the following regions contains the largest number of countries in the lowest urbanization class?"

Verb check: read it. Stay inside the figure. Sub-Saharan Africa and South Asia carry the lightest shading, and Sub-Saharan Africa holds far more countries. Answer: Sub-Saharan Africa. Nothing you know about any individual country is needed.

Question 2. "Which best explains why India appears in a light class on this map even though India contains one of the world's largest urban populations?"

Verb check: explain it. The map reports a rate, not a count. India's urban share in 2020 was about 35%, just over a third, but 35% of roughly 1.4 billion people is still nearly 500 million city dwellers. A country can be lightly shaded on a percent-urban choropleth and still hold more urban residents than every dark country in Western Europe combined. Answer: the map normalizes by each country's own population, so it measures urban share, not urban totals.

Name the trap: choropleths show rates, cartograms show magnitudes. A cartogram of total urban population would swell India and China and shrink Uruguay, and it would look nothing like this map of the same subject.

Common mistakes

  • Overriding the figure with memory. Stimuli use unlabeled countries, a specific year, and one narrow indicator. If the table reports 2023 total fertility and you answer from a crude birth rate you memorized last month, you will pick a true sentence about the wrong measure.

  • Choosing a true statement the stimulus does not support. When the stem says "according to," "shown," or "in the graph," correctness means visible in this figure. Distractors are often accurate geography; only one choice is accurate and demonstrated here.

  • Misreading the unit or the scale. Rate versus absolute count, per 1,000 versus percent, GDP versus GDP per capita at PPP, cartogram area versus land area, and log versus linear axes each change which region appears to lead. Check the legend and axis label before you rank anything.

  • Answering at the wrong scale of analysis. A stem that specifies a region, a country, or a city is asking whether the choice is true at that scale. Global patterns reverse locally, and a choice that is true worldwide can be false inside the one place the question named.

  • Carrying question one's answer into question two of the same set. Sets deliberately shift the task: read the figure, then name the concept, then predict a consequence. The stimulus stays constant; the verb does not.

  • Swapping near-neighbor terms. Contagious versus hierarchical diffusion, density versus distribution, formal versus functional versus vernacular region, site versus situation, von Thünen's agricultural rent versus Weber's industrial location. Distractors are manufactured from exactly these pairs, so verify the term rather than trusting that it sounds right.

  • Sinking four minutes into one dense map or cartogram. Every question is worth one point. Four minutes spent on a hard figure costs you the three fast vocabulary questions at the end of the section that you would have gotten right.

Checklist

  • Can you state, within ten seconds of seeing a figure, what it measures, in what units, for what year, and at what scale of analysis?
  • Can you describe the pattern in one sentence before you read a single answer choice?
  • Can you tell whether a stem wants you to read the figure, name the concept, or explain the cause — and say which one of the three can be answered from the figure alone?
  • Can you name the model behind a population pyramid, a ring diagram of farmland, a city-size graph, a religion map, and a flow map of migrants without pausing?
  • Can you look at a boundary and call it antecedent, subsequent, superimposed, or relict — and look at a state's outline and call it compact, elongated, fragmented, perforated, or prorupted?
  • Can you eliminate a choice that is true geography but unsupported by the stimulus, and explain in one sentence why it fails?
  • Can you distinguish contagious from hierarchical diffusion, density from distribution, and site from situation using one concrete example each?
  • Can you finish a three-question stimulus set in about 2:10 while reading the stimulus only once?
  • Can you point to a timed practice section where you enforced the 90-second ceiling, used Mark for Review, and finished with the review page showing nothing unanswered?

More AP Human Geography technique