After years of helping introductory statistics students, I realized that “I don’t know” can mean very different things—and much of good teaching is figuring out what is actually preventing the student from answering.
Statistics was never supposed to become my thing.
I have a degree in mathematics, but mathematics was primarily a tool I used for the subjects I actually loved: physics and chemistry. When I started tutoring, those were the subjects I expected to teach, along with calculus. I had taken exactly half a credit of introductory statistics myself, and I didn’t particularly enjoy the subject matter.
What I did enjoy was teaching, and it just turned out that I was very good at explaining statistics.
Students started coming to me for help. Then their friends started coming. Before I really understood what had happened, I was TAing six statistics labs a week, tutoring a dozen students, and had somehow become “The Statistics Girl.”
The irony was that many of my students didn’t particularly want to be there either.
When a student complained, “Statistics is stupid,” I would tell them, “I agree. So let’s make sure you’re not sitting here with me again next year.”
I would remind them, “You have a choice. This could all be over for you in a couple of months. You only see me once a week. I TA six of these labs a week, and next semester it starts all over again. Escape while you can.”
“I Can’t Do This. I’m Bad at Math.”
One explanation came up constantly:
Student: “I can’t do this. I’m bad at math.”
Me: “Well, you’re in luck then, because this is most definitely not a math course.”
Of course, statistics has mathematical foundations, and statistics at higher levels can involve substantial mathematics. But that wasn’t the course sitting in front of these students.
Introductory statistics used simple mathematics—arithmetic, percentages, square roots, and other things they had learned long before they arrived in my lab. For the most part, we weren’t teaching them new mathematics. We were teaching them statistical concepts, terminology, reasoning, and interpretation.
Occasionally a student disagreed with my assessment. Since I usually had my own textbooks with me, I was not always successful at resisting the temptation to show them some tensor notation for comparison.
They generally became much more appreciative of introductory statistics after that.
The Math Box
Eventually, I started noticing something strange.
Students who believed they were bad at math—or were simply afraid of it—could create a psychological barrier that blocked surprisingly ordinary reasoning. Put a simple question inside what I came to think of as the “Math Box,” and suddenly it became a mysterious unknown, even though the same student could answer exactly the same question instantly outside that box.
A common example went something like this:
Me: “What does ‘at least 20’ mean?”
Student: “I don’t know.”
Me: “Yes, you do.”
Student: “I don’t know!”
Me: “Okay. If I told you I would pay you at least $20 to do a job for me, you complete the job, and when you return I hand you $19—is that okay?”
Student: “NO! You said you would pay me at least $20!”
Me: “Oh. So you do know what ‘at least 20’ means.”
And then came the realization.
They had known the answer all along.
That became one of my most common teaching methods. With simple questions like this, I tried not to give students the answer. Instead, I changed the context until they could see that the knowledge was already there.
Statistical terminology created the same problem.
Student: “I don’t know what an explanatory variable is.”
Me: “Do you know what a variable is?”
Student: “Yes.”
Me: “Do you know what the word explain means?”
Student: “Obviously.”
Me: “Do you know that explanatory is the adjective form of explain?”
Student: “Yes.”
Me: “So you know what explanatory means, and you know what variable means…but you don’t know what explanatory variable means?”
There it was again.
One of the hardest—and most important—parts of my job was convincing students that statistics wasn’t written in some secret code designed to trick them.
It was using English.
Sometimes They Didn’t See the Answer Because They Didn’t Look
Over time, however, I realized that “I don’t know” didn’t always mean the same thing.
Sometimes there was no Math Box to dismantle. Sometimes the answer was sitting directly in front of them.
Student: “I don’t know what significance level to use.”
Me: “Seriously?”
Student, slightly offended: “I can’t find it.”
Me: “Please read this sentence.”
Student: “At the 5% significance level…”
Pause.
Student, now slightly embarrassed: “I didn’t read that part.”
Me: “Reading the whole question can be surprisingly helpful.”
That distinction became important to me:
Some students couldn’t see the answer because they hadn’t looked. Others had looked, but the answer was sitting behind a blind spot.
For one, I needed to make them look.
For the other, I needed to help them see.
Bring a Highlighter
One of my most common recommendations was surprisingly low-tech:
Bring a highlighter to your test.
It came directly from the way I worked through questions myself. As I read a problem with a student, I would stop and identify each piece of information as it appeared.
So I taught them to do the same thing.
As you read the question, highlight the important information. Don’t wait until you’re halfway through the problem and suddenly realize you can’t remember whether the sample size was 50 or 500. Mark it the first time you see it.
This isn’t because students can’t read. Tests are stressful, statistics questions can contain a lot of information, and once you’re concentrating on calculations and terminology, it’s remarkably easy to forget something you understood perfectly well thirty seconds earlier.
The highlighter gives you a visual map back to the information you need.
Think of it as insurance for those moments when test anxiety temporarily convinces you that you have forgotten how to read.
Use the Words From the Question
Even when students read the entire question, they had a tendency to discard details they didn’t realize were important.
For example:
Question: What proportion of introductory statistics students at Saint Mary’s University use public transit daily?
Student’s answer: “15% of students use the bus.”
Me: “So 15% of students at all universities?”
Student: “No, just Saint Mary’s students.”
Me: “Oh, 15% of all Saint Mary’s students?”
Student: “Well, no. Just the introductory statistics students.”
Me: “So 15% of introductory statistics students at Saint Mary’s University have used the bus once?”
Student: “No, daily.”
Me: “So 15% of introductory statistics students at Saint Mary’s University use public transit daily. But according to your answer, it’s 15% of all students who have ever used the bus at least once.”
Student: “I thought it was close enough.”
It wasn’t.
In ordinary conversation, perhaps it would have been. But statistical conclusions depend on exactly who we are talking about, exactly what was measured, and sometimes exactly when or how often it was measured.
I eventually started telling students not to reinvent the wording of the question when writing their conclusions. In many cases, the question had already given them almost everything they needed.
If the question asked for “the proportion of introductory statistics students at Saint Mary’s University who use public transit daily,” then those words belonged in the answer. Replacing them with “students who take the bus” might sound close enough in everyday conversation, but it changes both the population and the variable being described.
My rule became simple:
When the question gives you precise language, use it.
Don’t make a statistics question harder by trying to put everything into your own words.
I also gave my students one final test before submitting an answer:
“Ask yourself whether there is any question I could ask that would make your answer incomplete.”
They knew exactly what I meant.
Which students? Which university? What behaviour? How often?
If I could ask a question that exposed missing information, the answer wasn’t finished yet.
When Understanding Wasn’t the Goal
Eventually, I realized that with some students I was simply banging my head against a wall.
I could explain the same concept six different ways, draw pictures, give examples, strip it down to plain English, and try again—and we still weren’t getting anywhere.
And sometimes there was another problem: they didn’t particularly care whether they understood it.
Some students were taking introductory statistics for one reason: their program required it. They weren’t interested in statistics, they weren’t planning careers involving statistics, and no inspirational speech from me was going to make them care.
And, truthfully, I understood.
I hadn’t planned to become The Statistics Girl either.
So eventually I stopped treating their lack of enthusiasm as another problem I needed to solve. They didn’t need to love statistics. They needed to pass statistics and move on to whatever they had actually come to university to study.
That’s when I started developing fill-in-the-blank templates.
Statistics is wonderfully repetitive. Many interpretations and conclusions follow the same basic structure; only the context and values change. So instead of asking a struggling student to construct an entire answer from scratch, I gave them the structure and taught them how to identify which pieces of information belonged in each blank.
It worked.
Some students eventually understood the reasoning behind the template. Some didn’t.
But both groups could use it.
And for a student taking introductory statistics because it was required for a completely unrelated degree, sometimes that was enough.
And in the end, every student I personally tutored made it through introductory statistics.
They didn’t all get there the same way. Some needed an explanation that finally made something click. Some needed help seeing what they already knew. Some needed to slow down and actually read the question. Some needed structure, repetition and templates. And some just needed someone to help them survive a course they had absolutely no interest in taking.
But they all made it through.
From Tutoring to StatLab
I don’t tutor nearly as much as I once did, but the methods I developed while working with those students never disappeared. They eventually became the foundation for the resources I now create through StatLab.
I also know that not every student has access to a tutor—or can afford to meet with one every week. So I’ve tried to make the same kinds of explanations and tools available at different levels of support.
My YouTube channel provides free explanations and worked examples. My downloadable study guides are built around the same templates I developed while tutoring, with individual guides priced under $20 and semester bundles costing roughly what a single tutoring session might cost.
For students who need more support, I’ve also created a companion course designed to be used alongside an introductory statistics course, with my explanations, step-by-step examples, templates, practice questions, and complete solutions.
The format has changed, but the goal hasn’t: give students practical help with the parts of introductory statistics that actually cause them trouble.
What “I Don’t Know” Really Means
Over the years, I realized that much of my job wasn’t putting new information into students’ heads. It was figuring out what “I don’t know” actually meant—and then finding a way forward.
Sometimes the knowledge was already there. Sometimes it wasn’t. Sometimes understanding came eventually, and sometimes a reliable structure was enough.
My students didn’t all need the same thing, and they didn’t all leave loving statistics.
I never needed them to.
I just wanted to make sure they weren’t sitting there with me again next year.