FRQ 6 – Investigative Task

Overview

The AP Stats Investigative Task is Question 6, the final free-response question on the AP Statistics exam. You get 25 minutes for it, it's scored on a 0-4 scale, and it counts for 25% of your free-response score (the FRQ section is worth 50% of the exam, so the Investigative Task is roughly 12.5% of your total grade). It's the single most heavily weighted question on the test.

What makes Question 6 different from FRQs 1-5: it drops you into an unfamiliar context. You'll see a setup you probably haven't practiced, like a new kind of graph, a strange probability game, or a non-standard way of analyzing data. The task isn't to recall a memorized procedure. It's to take the statistics you already know and apply it to something new.

That's the whole point. The Investigative Task rewards clear statistical reasoning over guessing the "right" canned method. There's often more than one valid approach, so showing how you think matters more than landing on a hidden answer key.

This page is a deep dive on Question 6. For the other five free-response questions, see the FRQs 1-5 guide, and for the full exam layout start at the AP Statistics Exam page.

How the AP Stats Investigative Task Is Scored

Question 6 is scored holistically on a 0-4 scale, the same scale used for every AP Stats free-response question. Readers judge your whole response together rather than awarding a point for each sub-part. Each part of your answer gets rated as Essentially correct (E), Partially correct (P), or Incorrect (I), and those component ratings combine into your final 0-4 score.

Two things follow from holistic scoring. First, attempting every part beats perfecting one part and leaving others blank, because readers reward statistical thinking across the whole response. Second, communication is part of the grade. If a reader can't follow your logic, you don't get credit for it, even if your math is right.

Heads up: starting with the May 2027 exam, AP Statistics changes a lot. The free-response section will have 4 questions instead of 6, the exam goes fully digital in Bluebook, and several topics are cut. None of that applies to the 2025-26 exam, so everything on this page reflects the current test.

How to Approach the Investigative Task, Step by Step

Use your 25 minutes deliberately. This question gives you more time per question than Part A on purpose, because you have to understand a new situation before you can solve it.

Minutes 0-5: Read deeply and plan

Read the entire question twice. The first pass gives you the big picture; the second catches the details that matter later. Phrases like "selected without replacement" or "measured to the nearest gram" are usually doing real work.

As you read, answer three questions:

• What core concept is hiding here? Even a weird context connects to something familiar: variability, distribution, association, inference, or probability. Naming the core unlocks what you already know.
• What information am I given? List the facts, data, and conditions. Small details often resurface in later parts.
• What am I actually asked to do? Break each part into specific tasks. You might need to interpret an unusual graph, propose a method, or evaluate someone else's reasoning. Knowing exactly what's wanted prevents wasted effort.

Don't lock into a strategy yet. Just jot which concepts seem relevant.

Minutes 5-20: Work through the parts

Start with the part that feels most accessible. Early wins build momentum for the harder pieces. Investigative Task parts don't always build linearly, so part (b) sometimes reshapes how you see part (a). Stay flexible.

For each part, reread the prompt, sketch an approach, execute it with clear work shown, then check that your answer actually addresses what was asked. If you stall, switch angles. Can you work backwards from what you need? Solve a simpler version first? Connect it to a standard scenario you know cold?

Three moves rescue most stuck moments:

Build a bridge to a familiar concept. Suppose data is given as deviations from the average instead of raw values. The presentation is new, but you're still working with center and spread; the deviations are just already centered at zero. Map the strange surface onto a structure you know.

Try a tiny example. If a part defines a new measure of spread that compares every value to every other value, don't reason about it abstractly. Use {1, 3, 7} and compute the pairwise differences: |1-3| = 2, |1-7| = 6, |3-7| = 4. Now you can see what the measure captures and how it relates to range or standard deviation.

Organize probability concretely. For an unfamiliar game or random process, list the sample space, draw a tree diagram or table, and lean on the multiplication and addition rules as building blocks.

Minutes 20-25: Review and tighten communication

With answers in place, make your reasoning easy to follow. State any assumptions. If you picked one approach over another, say why in a sentence. Define any notation you invented, because a reader can't credit "X" if they don't know what X means.

Then sanity-check your numbers. Is a probability between 0 and 1? Is a mean inside the range of the data? These quick checks catch arithmetic slips. If a fresh idea hits you, add it; even "Another valid approach would be..." signals broader thinking.

Patterns That Show Up on Question 6

Every Investigative Task is unique, but a few flavors recur. Recognizing them helps you start faster.

Novel graphical displays. The prompt shows data in an unfamiliar visual and asks you to interpret it. Start by reading the axes and scale, then figure out what a single point or element represents. Build from concrete examples up to the general pattern, and ask what this would look like in a standard display.

Modified inference procedures. A familiar inference setup with a twist, like testing whether an 8-sided die is fair, or comparing groups whose data collection doesn't quite fit standard assumptions. The logic of inference holds even when details change: you still have hypotheses, a test statistic measuring deviation from the null, and a p-value quantifying evidence. Adapt the framework, don't abandon it.

Probability with a twist. A new game, a modified random process, or a layered conditional probability. Track what you know carefully, list the sample space even if it's big, and check your answer with a complementary approach when you can.

Unusual relationships. A relationship between variables that doesn't fit standard correlation or regression. Try transforming variables to linearize, consider categorical summaries if a continuous approach fails, and use graphs to see the pattern before you compute.

Worked Reasoning Example

Here's how the "bridge to a familiar concept" move looks in practice on an example similar to what you might see. (This is illustrative, not an actual exam item.)

Say a question introduces a "fairness index" for a new carnival game and asks whether the game is fair to players. The phrase "fairness index" sounds foreign, but strip away the wrapper and it's asking about expected value and variability, concepts you know well.

A clean response would:

1. Define fair in statistical terms: a player's expected net gain is 0.
2. Build the probability distribution of the player's outcome from the game's rules.
3. Calculate the expected value. If $E(X) < 0$, the game favors the house.
4. Connect back: "Because the expected payoff is negative, a player loses money on average, so the game is not fair."

Notice the response names the concept, shows the calculation, and explains the conclusion in plain words. That combination of correct work plus clear communication is exactly what earns a 4.

Common Mistakes

• Overthinking the novel context. Students fixate on every quirk of the scenario and miss the simple statistics underneath. Fix: strip away the unfamiliar vocabulary and ask what fundamental question is really being asked.
• Abandoning statistical principles. Faced with something new, some students stop reasoning and just guess. Fix: do the opposite. Fall back on fundamentals. What would a graph show? What summary statistic helps? How can you quantify uncertainty?
• Weak communication. Logic that feels obvious to you can look like a leap to a reader, and holistic scoring won't credit reasoning it can't follow. Fix: explain connections, define your notation, and show intermediate steps.
• Getting stuck on one approach. If your first method isn't working after 5-7 minutes, you're burning time. Fix: switch tools. There's usually more than one valid path, so flexibility beats stubbornness.
• Leaving parts blank. Because scoring is holistic, a partial attempt on every part outscores one perfect part and three empty ones. Fix: write something reasoned for each part, even if you're unsure.
• Skipping the sanity check. Probabilities above 1 or a mean outside the data range slip through when you don't review. Fix: spend the last minutes checking that answers make sense in context.

Practice and Next Steps

You can't rehearse the exact Investigative Task you'll get, but you can build the flexible thinking it tests. Take procedures you know and adapt them: if you can test whether a coin is fair, work out how you'd test a 12-sided die; if you can compare two means, figure out how you'd compare two medians. For each standard method, ask yourself why it works and which assumptions matter, because understanding the "why" is what lets you adapt under pressure.

Then practice under real conditions. Run timed FRQ practice with instant scoring and pull more prompts from the FRQ question bank and past exam questions to get used to unfamiliar setups. Take a full-length practice exam so 25 minutes on Question 6 feels normal, and use the AP score calculator to see how the Investigative Task fits into your overall score. To shore up the fundamentals behind it all, review the key terms glossary and the rest of the AP Statistics Exam prep guides.