A No-Nonsense Look at Introduction to Statistical Investigations
I ran into this curriculum about six years ago when a colleague recommended it as a replacement for the traditional textbook route. The official title is Introduction Statistical Investigations Nathan Tintle, and it's published through Macmillan Learning. It targets AP Statistics and introductory college-level stats courses, built around an inquiry-based model rather than the standard lecture-and-problem-solve structure most of us grew up with. At its core, the book is a collection of structured activities where students encounter data first, wrestle with questions about it, and gradually build the statistical vocabulary and methods around their discoveries. The theory comes after the intuition. That flips the typical textbook approach on its head. Chapters progress through core AP Statistics topics: study design, distributions, probability, sampling, inference, regression, and chi-square tests. Each chapter contains multiple activities, each one designed to be worked through in class with some combination of group discussion, technology, and guided questions. The accompanying technology piece is called statKey. It's a free web-based tool that handles simulations, randomization tests, confidence intervals, and probability distributions without requiring Minitab or TI-84 specific button sequences. That was one of the selling points for my department. We spent less time teaching calculator navigation and more time talking about what the numbers actually meant.
How It Works in Practice
Each activity follows a rough pattern: there's a scenario or dataset, a set of investigative questions that start simple and get more technical, space for students to record findings, and synthesis questions at the end that connect the activity back to formal statistical concepts. The teacher's role shifts from delivering content to circulating, probing student reasoning, and filling in gaps during the wrap-up discussion. The activities are intentionally incomplete without student engagement. You cannot just assign the readings and expect comprehension. Students need to do the simulations, make the plots, and argue through the conclusions. I've seen departments try to use it as a self-study resource and fail because the design assumes classroom interaction. It's not a supplement you can quietly layer onto an existing lecture course. It replaces the lecture entirely.
A Specific Problem I Ran Into
During my second semester using the curriculum, I hit a snag with the randomization test activities in the inference chapters. Students were supposed to use statKey to build null distributions, but a significant number of them were treating the simulation output as a final answer without connecting it to the formal p-value definition. They'd run 1000 reps, get a distribution, and then just report the proportion without understanding that the randomization distribution was an approximation of what would happen under the null hypothesis. Some were also misreading the app's interface and accidentally computing one-sided p-values when the question asked for two-sided. My workaround was to add a mandatory written step before students submitted any simulation result: they had to write one sentence describing what the null hypothesis was in plain language and one sentence describing exactly what the simulated statistic represented. It took five extra minutes per activity but eliminated roughly 80 percent of the misinterpretation errors I was seeing. The curriculum itself doesn't build in that requirement, so I had to create it myself.
Get the Full Details
Common Pitfalls
The pacing assumption is aggressive. The book is written assuming roughly one activity per class period over a full semester. If you're trying to cover the same material in fewer periods or alongside a traditional homework load, you will fall behind. I've watched instructors skip activities and try to reconstruct the inquiry portion with worksheets, and the result is worse than teaching the content directly. The inquiry is the mechanism, not optional decoration. Calculator dependency drops. Because statKey handles so much computation, students who rely on this curriculum may struggle on the AP exam where they need to use a TI-84 or similar device. I made sure to schedule periodic calculator-practice sessions outside the main activities. Without that, students can pass the course using statKey but still score poorly on the free-response section because they can't manually compute a t-test or construct a confidence interval under exam conditions. Some activities are unevenly developed. The early chapters on data exploration and study design are strong and well-paced. The later chapters on chi-square and complex inference procedures feel more rushed. The regression chapter in particular has thinner activity support than the inference chapters, which surprised me given how heavily the AP exam weights regression analysis.
What Beginners Miss
Most people coming into this curriculum expect it to be easier than a traditional textbook because the language is simpler and the problems are more concrete. That's true on the surface, but the depth requirement is actually higher in some ways. Students need to articulate reasoning in written form throughout, not just produce correct numerical answers. The written communication demand catches students off guard. They can compute a p-value but then can't explain in two sentences what it means in context. Another thing that catches people: the curriculum assumes statistical literacy, not just procedural fluency. When a student gets a question about sampling bias, the book expects them to discuss how the data collection method could distort results, not just identify the bias type from a list. That's a harder skill to develop and harder to assess quickly.
Where It Falls Short
The curriculum is not designed for self-directed learners or for students who need remedial math support. The pace assumes a baseline comfort with algebra and a willingness to engage with ambiguity. Students who struggle with open-ended problem solving tend to disengage. I've had to supplement with additional direct instruction for those students, which somewhat defeats the purpose of the inquiry model but is necessary in mixed-ability classrooms. There is also limited material on machine learning, predictive modeling, and modern data science applications. If your course aims to bridge into those areas, you will need external resources. The curriculum is firmly rooted in classical statistical inference and will not prepare students for anything beyond that scope. The instructor materials are adequate but not exceptional. The solutions are correct, but the pedagogical notes that explain why each activity matters and what misconceptions to watch for are thinner than I'd like. I found myself writing my own implementation notes over successive semesters rather than relying on the published guidance.
Where to Access It
The main text and statKey platform are available through Macmillan Learning's website. statKey itself is free and does not require a textbook purchase. The full curriculum with instructor materials requires a license through the publisher or your institution. You can find the landing page at macmillanlearning.com, and statKey is accessible directly at openstat.org/statkey. There is no legitimate third-party download of the full curriculum, and anything offering that is likely pirated or outdated. If you're considering this for a course, the realistic time investment for a first-time instructor is about 15 to 20 hours during the first semester to familiarize yourself with the activities, set up statKey in your classroom environment, and build the supplementary calculator practice sessions I mentioned. After that, maintenance drops to maybe 2 to 3 hours per week. The curriculum is not a set-it-and-forget-it product, but it stabilizes after the initial ramp-up. The students who thrive with it are the ones who enjoy discussing data and don't mind being wrong in the process. The ones who prefer clear rules and right-or-wrong answers tend to resist it. That's not a flaw in the curriculum, it's just a mismatch in fit.
