Unit 4 FRQ Practice — Concept Application
Master the AP Gov Concept Application FRQ using a polling scenario: methodology differences, margin of error, and tracking polls, scored on the 3-point rubric.
What you'll do in this lesson
A voice-first session with the Crimsora tutor on Unit 4 FRQ Practice — Concept Application, then targeted practice and FRQs — with the tutor adapting to where you get stuck.
What this lesson covers
The Concept Application FRQ is one of four free-response questions on the AP US Government exam, and it is often the most learnable because it follows a predictable pattern: read a short non-partisan scenario, then answer three tasks that each ask you to apply a course concept to that scenario. In this lesson you will attack a full Concept Application prompt about a Senate candidate staring at two contradictory polls released the same week. You will practice describing a methodological reason polls diverge, explaining how margin of error limits what a live-caller poll reveals, and explaining how tracking polls guide campaign spending. By the end you will know exactly how the 3-point rubric awards each point and how to write responses that graders can score quickly.
How the Concept Application FRQ Is Built
Every Concept Application question opens with a stimulus — a paragraph or short passage describing a real-world or hypothetical political situation. The prompt then poses three tasks labeled A, B, and C. Each task is worth exactly 1 point, for a total of 3 points, and the points are independent: missing part A does not prevent you from earning B and C.
The verbs in each task matter. A task that says describe wants a clear statement of a concept and what it looks like in the scenario. A task that says explain wants you to show a cause-and-effect relationship — how or why one thing produces another. Graders read for a specific concept plus a link to the stimulus. General knowledge with no tie to the candidate, the polls, or the campaign will not score.
The biggest scoring danger is writing about the topic in general terms instead of applying it to the exact facts you were given. Name the poll, the candidate, or the resource decision in every part.
The verbs in each task matter. A task that says describe wants a clear statement of a concept and what it looks like in the scenario. A task that says explain wants you to show a cause-and-effect relationship — how or why one thing produces another. Graders read for a specific concept plus a link to the stimulus. General knowledge with no tie to the candidate, the polls, or the campaign will not score.
| Task verb | What it demands | Common miss |
|---|---|---|
| Describe | Define a concept and identify it in the scenario | Defining but never connecting to the scenario |
| Explain | Give reasoning that links cause to effect | Restating the prompt without the mechanism |
Part A: A Methodological Reason Polls Differ
Two credible polls can report different results in the same week without either being fraudulent. Part A asks you to describe one methodological reason. Strong methodological factors include differences in the sampling frame (registered voters versus likely voters), sample size, the mode of contact (live caller, automated, online panel), question wording or order, weighting of demographic groups, and the timing of the field period.
To score, pick one factor and connect it to the scenario. For example: one poll may survey registered voters while the other screens for likely voters; because likely-voter models estimate who will actually turn out, the two surveys are measuring slightly different populations and can produce different candidate support numbers.
A common misconception is that any difference must mean one poll is biased or wrong. On the exam, avoid partisan or accusatory language — the scenario is deliberately non-partisan, and so should your answer be. Also avoid vague statements like "the polls asked different people." That is true but too thin; name the methodological mechanism (sampling frame, mode, weighting) and explain briefly how it changes the result. One well-developed sentence with a named factor plus a scenario link earns the point.
To score, pick one factor and connect it to the scenario. For example: one poll may survey registered voters while the other screens for likely voters; because likely-voter models estimate who will actually turn out, the two surveys are measuring slightly different populations and can produce different candidate support numbers.
A common misconception is that any difference must mean one poll is biased or wrong. On the exam, avoid partisan or accusatory language — the scenario is deliberately non-partisan, and so should your answer be. Also avoid vague statements like "the polls asked different people." That is true but too thin; name the methodological mechanism (sampling frame, mode, weighting) and explain briefly how it changes the result. One well-developed sentence with a named factor plus a scenario link earns the point.
Part B: Margin of Error and the Live-Caller Poll
Margin of error (MOE) expresses the range of sampling uncertainty around a poll's reported percentage, usually at a 95 percent confidence level. If a candidate polls at 48 percent with a margin of error of plus or minus 3 points, the true value likely lies between 45 and 51 percent.
Part B asks you to explain how margin of error affects what the live-caller poll can tell the campaign. The key insight graders want: when two candidates' support figures overlap within the margin of error, the poll cannot reliably say who is ahead — the race is a statistical tie. So if the live-caller poll shows the candidate at 47 percent and the opponent at 45 percent with an MOE of plus or minus 3, that 2-point lead is inside the error band and is not conclusive.
To explain rather than merely describe, walk the cause and effect: because the sample is only an estimate of the whole electorate, the MOE defines a confidence interval; because the candidates' intervals overlap, the campaign cannot treat the lead as real. A frequent error is treating MOE as a flaw or a measure of dishonesty. It is a normal feature of sampling. Tie the reasoning explicitly to the campaign's decision-making to lock in the point.
Part B asks you to explain how margin of error affects what the live-caller poll can tell the campaign. The key insight graders want: when two candidates' support figures overlap within the margin of error, the poll cannot reliably say who is ahead — the race is a statistical tie. So if the live-caller poll shows the candidate at 47 percent and the opponent at 45 percent with an MOE of plus or minus 3, that 2-point lead is inside the error band and is not conclusive.
To explain rather than merely describe, walk the cause and effect: because the sample is only an estimate of the whole electorate, the MOE defines a confidence interval; because the candidates' intervals overlap, the campaign cannot treat the lead as real. A frequent error is treating MOE as a flaw or a measure of dishonesty. It is a normal feature of sampling. Tie the reasoning explicitly to the campaign's decision-making to lock in the point.
Part C: Using Tracking Polls to Allocate Resources
A tracking poll surveys opinion repeatedly over time, often using a rolling sample, so the campaign can watch trends rather than a single snapshot. Part C asks you to explain how the campaign could use tracking polls to decide resource allocation — where to spend money, staff, ads, and the candidate's time.
The scoring logic requires a link between the data and an action. For example: if tracking polls show the candidate's support sliding in a particular region or among a particular demographic over several weeks, the campaign can shift advertising dollars and field organizers into that area to stop the slide. Conversely, if support is stable and strong somewhere, the campaign can pull resources from that safe area and redeploy them to competitive ground.
To earn the point, do not just define tracking polls — explain the mechanism connecting the trend data to a spending choice. Weak answers say "they can use polls to help the campaign," which restates the task. Strong answers name a specific resource (television ads, get-out-the-vote canvassing, candidate appearances) and describe how a movement in the tracking numbers triggers moving that resource. Because this is an explain task, the cause-and-effect chain from data to decision is what graders check.
The scoring logic requires a link between the data and an action. For example: if tracking polls show the candidate's support sliding in a particular region or among a particular demographic over several weeks, the campaign can shift advertising dollars and field organizers into that area to stop the slide. Conversely, if support is stable and strong somewhere, the campaign can pull resources from that safe area and redeploy them to competitive ground.
To earn the point, do not just define tracking polls — explain the mechanism connecting the trend data to a spending choice. Weak answers say "they can use polls to help the campaign," which restates the task. Strong answers name a specific resource (television ads, get-out-the-vote canvassing, candidate appearances) and describe how a movement in the tracking numbers triggers moving that resource. Because this is an explain task, the cause-and-effect chain from data to decision is what graders check.
Timing, Structure, and Rubric Strategy
You will have limited time across all four FRQs, so budget roughly 15 to 20 minutes for a Concept Application. Answer in order and label each part A, B, and C so graders can find your points instantly.
Do not write a formal essay with an introduction and conclusion — the Concept Application rewards direct, applied answers, not eloquence. Repetition of the stimulus wording without your own reasoning wastes time. The single most reliable way to gain points is to make the scenario connection explicit in every sentence that carries an idea. If you are unsure whether you have applied the concept, ask yourself whether your sentence would still make sense if the scenario were about something else; if it would, you have not applied it yet.
| Step | Action |
|---|---|
| Read | Underline the concept demanded and the scenario facts |
| Plan | Match one concept to each task |
| Write | One tight paragraph per part, name the concept, link to scenario |
| Check | Confirm each part references the candidate, poll, or campaign |
Key terms
- Concept Application FRQ.
- An AP Gov free-response question worth 3 points that presents a scenario and asks you to apply course concepts to it in three tasks.
- Sampling frame.
- The population a poll actually draws from, such as registered voters versus likely voters; differences here can cause polls to diverge.
- Margin of error (MOE).
- The range of uncertainty around a poll's reported percentage due to sampling, typically stated at a 95 percent confidence level.
- Confidence interval.
- The band around a poll result within which the true population value is likely to fall, defined by the margin of error.
- Live-caller poll.
- A survey conducted by human interviewers calling respondents, one of several polling modes that can affect results.
- Tracking poll.
- A poll that repeatedly measures opinion over time to reveal trends rather than a one-time snapshot.
- Statistical tie.
- A situation where candidates' support figures overlap within the margin of error, so no lead can be called reliable.
- Likely-voter model.
- A method of screening survey respondents to estimate who will actually vote, which can shift reported candidate support.
Worked example
In the final month of a competitive Senate race, a non-partisan candidate's campaign receives two polls released the same week. A live-caller poll of likely voters shows the candidate at 47 percent to the opponent's 45 percent, with a margin of error of plus or minus 3 points. An online-panel poll of registered voters shows the opponent ahead by 4 points. The campaign also has weekly tracking-poll data by region. Part A: Describe one methodological reason the two polls could show different results. Part B: Explain how the margin of error affects what the live-caller poll can tell the campaign. Part C: Explain how the campaign could use tracking polls to decide resource allocation.
Start with Part A. Identify a named methodological difference in the stimulus. The polls differ in both mode (live caller versus online panel) and sampling frame (likely voters versus registered voters). Choose one and link it: "One poll screens for likely voters while the other surveys all registered voters. Because likely-voter models estimate who will actually turn out and registered-voter samples include people who may not vote, the two polls measure different populations, which can produce different support numbers." That names the mechanism and ties it to the scenario, earning point A.
For Part B, use the numbers. The candidate leads 47 to 45, a 2-point margin, but the MOE is plus or minus 3. "Because the margin of error creates a confidence interval around each figure, the candidate's 47 percent could range from 44 to 50 and the opponent's 45 percent from 42 to 48. Since these ranges overlap, the 2-point lead falls within the margin of error, so the live-caller poll cannot tell the campaign it is actually ahead — the race is a statistical tie." This explains cause and effect and applies the numbers, earning point B.
For Part C, connect trend data to a spending decision. "If the weekly tracking polls show support declining in one region over several weeks, the campaign can move advertising money and field staff into that region to reverse the trend, while pulling resources from regions where its lead is stable. The trend data lets the campaign target spending where it will change the outcome." Naming a resource and the data-to-action link earns point C, for a full 3 out of 3.
For Part B, use the numbers. The candidate leads 47 to 45, a 2-point margin, but the MOE is plus or minus 3. "Because the margin of error creates a confidence interval around each figure, the candidate's 47 percent could range from 44 to 50 and the opponent's 45 percent from 42 to 48. Since these ranges overlap, the 2-point lead falls within the margin of error, so the live-caller poll cannot tell the campaign it is actually ahead — the race is a statistical tie." This explains cause and effect and applies the numbers, earning point B.
For Part C, connect trend data to a spending decision. "If the weekly tracking polls show support declining in one region over several weeks, the campaign can move advertising money and field staff into that region to reverse the trend, while pulling resources from regions where its lead is stable. The trend data lets the campaign target spending where it will change the outcome." Naming a resource and the data-to-action link earns point C, for a full 3 out of 3.
Practice questions
A live-caller poll reports Candidate X at 49 percent and Candidate Y at 46 percent with a margin of error of plus or minus 4 points. Which statement best explains what the campaign can conclude?
- Candidate X is definitively winning because 49 is greater than 46
- The race is effectively a statistical tie because the candidates' support ranges overlap within the margin of error
- The poll is biased because the numbers do not add to 100 percent
- The margin of error proves the poll used an unreliable sampling method
Answer: The race is effectively a statistical tie because the candidates' support ranges overlap within the margin of error
With a 3-point gap and a plus or minus 4-point margin of error, Candidate X's range (45 to 53) overlaps Candidate Y's range (42 to 50). Overlapping confidence intervals mean the poll cannot reliably identify a leader, so it is a statistical tie. The margin of error is a normal feature of sampling, not evidence of bias or bad methodology.
In a Concept Application response, explain how a campaign could use tracking polls to decide where to allocate advertising resources during the final weeks of a race.
Answer: A campaign can use tracking polls to detect trends in candidate support over time and shift advertising dollars toward areas or groups where support is declining, while reducing spending where its lead is stable.
A full-credit answer must connect the data to an action. Tracking polls survey opinion repeatedly, revealing movement rather than a single snapshot. If the data show support slipping in a region, the campaign concentrates ad spending there to reverse the decline; if a region is safe, it redeploys resources to competitive areas. The scoring key is the cause-and-effect link between the trend and the specific resource decision, not merely defining what a tracking poll is.
Which of the following is the strongest methodological reason two polls conducted the same week might report different results?
- One pollster personally supports a different candidate
- One poll surveyed likely voters while the other surveyed all registered voters
- One poll was released in the morning and the other in the evening
- One poll had a catchier name than the other
Answer: One poll surveyed likely voters while the other surveyed all registered voters
Differences in the sampling frame are a genuine methodological cause of divergence: likely-voter and registered-voter samples measure different populations and can yield different results. Personal preference is a partisan accusation, not a methodological factor, and release time or poll names have no bearing on measured opinion. The exam rewards naming an actual methodological mechanism.
FAQ
- How many points is the Concept Application FRQ worth and how are they divided?
- It is worth 3 points total, one point for each of the three tasks (A, B, and C). The points are scored independently, so you can miss one part and still earn the others. Label your answers by part so graders can find each point.
- What is the difference between describe and explain on this FRQ?
- Describe asks you to state a concept and identify it in the scenario. Explain asks you to show a cause-and-effect relationship — how or why something happens. Explain tasks require an extra step of reasoning, so add a because or so-that clause linking your concept to the scenario.
- Do I need to write an introduction and conclusion for the Concept Application question?
- No. It is not an essay. Write direct, applied responses labeled A, B, and C. An introduction and conclusion waste time you need for the other FRQs and earn no points. Focus every sentence on applying a concept to the given scenario.
- Why do two legitimate polls sometimes disagree, and is that a good FRQ answer?
- Legitimate polls disagree because of methodological differences like sampling frame, mode of contact, weighting, question wording, and field timing. Naming one specific factor and connecting it to the scenario is exactly what Part A rewards. Avoid claiming a poll is dishonest or partisan, since the scenario is non-partisan.
Learn this with a teacher, not a page
The Crimsora tutor teaches Unit 4 FRQ Practice — Concept Application live — explaining on a whiteboard, asking you questions, and adapting to where you get stuck.