What Is a Statistical Question?
Consider two questions: "How tall am I?" and "How tall are the students in my class?" The first is answered by measuring one person once — the result is a single fixed value. The second requires measuring every student, and those measurements will differ. Heights of 152, 158, 163, 160, and 155 cm are not identical. That spread of values is what makes the second question statistical.
Variability is not a problem to avoid. It is the reason statistical analysis exists. When observations differ, you need tools — means, medians, distributions, proportions — to describe what you found. A question that anticipates this need is a statistical question.
A statistical question is a question that expects data to vary and requires those data to be collected and analyzed. For example, "How tall are the students in my class?" is statistical because students have different heights. "How tall am I?" is not statistical because it asks for one fixed measurement rather than a distribution of varying observations.
The concept of a statistical question is introduced in CCSS.Math.Content.6.SP.A.1. The standard asks students to recognize that a statistical question anticipates variability in the data and accounts for that variability in the answer — not simply that it has "many" possible answers.
What Makes a Question Statistical? Variability.
Variability means that the values or responses collected are not all identical. When you ask students how many books they read last month, you might get: 2, 5, 1, 3, 4, 0, 7, 2. Those eight numbers are not the same. That's variability.
Variability can appear across many different kinds of observations:
- People: Different students have different heights, sleep durations, and homework times.
- Objects: Different classroom doors might not be identical in width once you measure them carefully.
- Events: Different school days see different numbers of absent students.
- Repeated trials: A machine filling bottles won't produce identical fill volumes every single cycle.
- Time periods: One person's daily commute varies from Monday to Friday.
In every case, collecting those observations and seeing how they spread out is the core task of statistics. The statistical question is the one that starts that process.
Statistical vs. Non-Statistical Questions
A non-statistical question asks for a single fixed fact, one defined value, or one calculation. It does not require collecting and analyzing a set of varying observations. The table below compares the two:
| Feature | Statistical Question | Non-Statistical Question |
|---|---|---|
| Data structure | Requires a set of observations | Often answered by one lookup or measurement |
| Variability | Observations are expected to vary | One fixed value is the target |
| Data type | Numerical or categorical | Can involve numbers too (see note) |
| Analysis needed? | Yes — summarize or compare the variation | No distribution of varying data required |
| Example | "How old are the students in my school?" | "How old am I?" |
"How many days are in October?" contains a number (31) but it is not statistical — no varying dataset is needed to answer it. What determines statistical status is whether the answer requires collecting and analyzing observations that vary, not whether numbers appear in the question or answer.
How to Tell If a Question Is Statistical
Grammar and wording don't determine whether a question is statistical. The question "How many…?" can be either type depending on context. What matters is whether the answer needs varying observations. Run through these three checks:
Does it require a set of observations?
Not a single fact, formula, or one-time measurement. You need to collect data from multiple people, objects, events, or repeated trials.
Do you expect the observations to vary?
Different students will give different answers. Different days will show different values. Different trials will produce different results. If every observation is expected to be identical, it's less likely a statistical question.
Will the answer summarize or describe that variation?
Through counts, percentages, averages, distributions, comparisons, or relationships. If the answer involves describing a pattern across varying observations, it's statistical.
If you answer yes to all three, the question is statistical. This works better than asking only "Does the question have more than one answer?" — because a trivia question might have multiple guesses, yet the correct answer is still one fixed fact.
Examples of Statistical and Non-Statistical Questions
The best way to build intuition is through pairs. Each one below contrasts a statistical question with a closely related non-statistical one.
Age
Height
Sleep
Favorite color
Temperature
School absences
Can a Statistical Question Have One Final Answer?
Yes — and this trips up many students. The final result of a statistical analysis is often a single summary: an average, a percentage, a most common category. That does not make the question non-statistical.
"What is the average height of Grade 6 students at this school?" may produce one answer: 157.4 cm. The question is still statistical because arriving at that number required collecting many different heights and computing a summary. The underlying observations varied; the mean is just how we describe that variation in a single value.
The same applies to other summaries. "What proportion of students walk to school?" produces one percentage. "What is the most popular lunch choice?" produces one category. Both questions are statistical because they rest on a set of varying observations gathered from multiple people.
What makes the question statistical is not how many numbers appear in the final answer — it is whether the process of answering it required collecting and analyzing varying data.
Statistical Questions with Categorical Data
A common error — especially in introductory courses — is assuming that statistical questions must produce numerical answers. They don't. Categorical data (names, labels, choices) can vary just as meaningfully as numbers.
| Question type | Example | Data produced |
|---|---|---|
| Quantitative | How many minutes do students spend traveling to school? | Numbers: 5, 12, 18, 9, 25… |
| Categorical | Which mode of transportation do students use to get to school? | Categories: bus, walk, car, bicycle… |
| Categorical (yes/no) | What proportion of students have a pet? | Binary: yes, no, yes, yes, no… (summarized as a percentage) |
All three questions anticipate variability. All three require collecting data from multiple students. All three are statistical. The types of data page covers categorical and quantitative data in detail; here it is enough to know that both can be the basis of a statistical question.
Yes/no responses work the same way. Each individual student answers either yes or no — one fixed response for that person. But across 30 students, you get a distribution of yes and no answers that varies from student to student, and that distribution tells you the proportion who have a pet. That proportion is a statistical result, and the question is statistical.
One Person Can Produce Statistical Data
Statistical questions don't always involve many people. Repeated observations of a single subject across time or trials also produce varying data.
"How many steps does Maria take each day over 30 days?" is statistical. Maria is one person. But her daily step counts — 7,200; 9,400; 6,800; 11,200; 8,000 — vary from day to day. That variation is what makes the question statistical. A distribution forms. Summaries are possible. The question asks about a pattern across varying observations, not a single fixed value.
The same applies to machines, locations, and experiments. "How long does this machine take to fill a bottle across 100 production cycles?" asks about one machine. But the fill times vary: 2.3 s, 2.5 s, 2.4 s, 2.6 s. Studying that variation is a statistical task.
The key is varying observations — not the number of people, objects, or locations involved.
Types of Statistical Questions
Statistical questions can be organized into four categories based on what they aim to learn. These are not the only possible groupings, but they cover the main kinds you will encounter at the introductory level:
Descriptive
Describes the distribution of one variable in one group. No comparison made.
Comparative
Compares a variable across two or more groups. Do not infer causation unless the study design warrants it.
Association / Relationship
Explores whether two variables vary together. Uses "associated with" or "related to," not "causes."
Repeated Observation
Tracks how a variable changes over time or across repeated trials. One subject, many moments.
Each type connects to a different kind of analysis. Descriptive questions produce distributions and summary statistics. Comparative questions call for group comparisons. Association questions use correlation or regression. Repeated-observation questions use time-series or within-subject methods. The study design section explains how to match question types to appropriate data collection plans.
Statistical vs. Non-Statistical: 12 Side-by-Side Examples
| Statistical question ✓ | Non-statistical question ✗ | Why the difference? |
|---|---|---|
| How old are the students in my school? | How old am I? | Many varying ages vs. one person's fixed age. |
| How many siblings do Grade 6 students have? | What is 6 × 8? | Sibling counts vary; the multiplication has one fixed answer. |
| What colors do students choose for their backpacks? | What color is this pencil? | Many category responses vs. one fixed observation. |
| How many minutes do students spend on homework each night? | How many minutes are in one hour? | Homework time varies; 60 minutes is a fixed fact. |
| What percentage of students in this school bring lunch from home? | Does this specific student bring lunch? | A proportion across many students vs. one person's yes/no. |
| What are the shoe sizes of students on the basketball team? | What is Mia's shoe size? | Shoe sizes vary across a group vs. one person's size. |
| How do daily noon temperatures vary during October? | What was the temperature at noon today? | Thirty varying readings vs. one specified reading. |
| How many steps does Maria take each day over 30 days? | How many steps did Maria take yesterday? | Thirty varying daily counts vs. one day's fixed value. |
| What proportion of 100 coin flips land heads? | What is the theoretical probability of heads on one flip? | Observed varying outcomes vs. a fixed theoretical calculation. |
| How many books do Grade 6 students read each month? | How many planets are in the Solar System? | Reading counts vary by student; the planet count is a defined fact. |
| Which lunch option is most popular among students? | What is the name of the school cafeteria? | Many varying student preferences vs. one fixed name. |
| How does this machine's fill time vary across 100 cycles? | How long was this one bottle fill? | A distribution of 100 varying times vs. one recorded value. |
How to Write a Good Statistical Question
A good statistical question makes clear what is being observed, who the observations come from, and what you want to learn. A loose formula that works at the beginner level:
Example built from that formula:
"What is the median number of minutes Grade 6 students at Lincoln School spend reading at home on a typical weekday?"
Population: Grade 6 students at Lincoln School. Variable: minutes spent reading. Context: weekday, at home. Feature: median (typical value). Variability: different students spend different amounts of time — that's the whole point.
You don't need to name a specific statistic (mean, median) in the question itself. "How many hours do sixth graders sleep each night?" is perfectly valid. The analysis method is chosen after the data are collected.
Writing templates
How to Turn a Non-Statistical Question Into a Statistical One
Most non-statistical questions can be made statistical by changing the observational scope — expanding from one person or moment to a group or repeated observations.
In each case the transformation is the same: instead of fixing the observation to one person, moment, or instance, you open it up to a collection of observations that are expected to differ.
Statistical Questions for 6th Grade
Grade 6 students work with statistical questions on topics close to their daily experience. Here are examples using the most common classroom contexts. Each one anticipates variation.
The Statistical Investigation: Question Comes First
A statistical question is not just a curiosity. It determines what data you need to collect, from whom, and how you will analyze the results. Asking a vague question leads to vague data and inconclusive analysis. A clear statistical question anchors the entire investigation.
From Question to Answer
This connects to the broader study design process. Questions also determine whether the data will be categorical or quantitative, which guides visualization choices — bar charts for categories, histograms or dot plots for numerical values. See data visualization for how chart type follows from data type.
And the question determines whether the analysis will describe one group or infer something about a larger population from a sample. That distinction is covered in descriptive vs. inferential statistics. When the full population cannot be observed, a sample is drawn — but the question still frames what that sample is meant to tell us.
Common Mistakes About Statistical Questions
| Mistake | Why it's wrong | Correction |
|---|---|---|
| Statistical questions must have numerical answers | Categorical data — colors, names, choices — can also vary | Variability, not data type, is the defining feature |
| The final answer must include multiple values | A statistical analysis often ends with one summary (mean, percentage) | One summary value is fine; the underlying observations still varied |
| Statistical questions always involve many people | One person's commute over 30 days, or one machine over 100 cycles, can be statistical | Repeated observations of one subject also produce varying data |
| Yes/no questions cannot be statistical | Binary responses across a group vary and can be summarized as proportions | "What proportion of students have a pet?" is statistical |
| If it contains a number, it's statistical | "How many days are in October?" has a number (31) but needs no varying data | Numbers in the question or answer don't determine statistical status |
| The word "average" makes a question statistical | Only when the average is computed from varying observations | Ask: what observations produce the average? If they vary, it's statistical |
| Every survey item is a statistical question | A single survey item asked of one person produces one fixed response | The statistical question is the broader investigative question answered by many respondents' varying replies |
| Causal language is fine in any statistical question | "Does screen time cause lower grades?" implies causation that a survey cannot establish | Use "associated with" or "related to" unless the study design supports causal inference |
Practice: Which Questions Are Statistical?
For each question below, decide whether it is statistical or non-statistical, then reveal the answer and explanation. Answers include the reasoning — not just yes or no.
Fixed factual answer — Paris. No varying dataset is needed to answer it.
Travel times vary from student to student: 5, 12, 20, 8, 35 minutes. Collecting and summarizing those varying values is required.
Fixed fact: 60. No data collection or analysis of varying observations is involved.
Each student's response (yes/no) varies. The percentage is a summary of those varying binary responses across the whole school.
One person, one fixed value. No varying dataset needed.
Shoe sizes differ across team members. Collecting and describing that distribution is the task.
Specified location and time — one target reading. Once defined, the answer is one fixed historical value.
Thirty-one daily readings expected to vary. A distribution forms; center and spread can be described.
One hundred observed flips produce varying outcomes (H, T, H, H, T…). The proportion is a summary of those empirical results. This is different from asking for the theoretical probability, which is a fixed mathematical value.
Answered mathematically: 1/6. No empirical data collection is required. A related statistical question would be: "What proportion of 600 rolls of this die come up 6?" — that involves observed, varying data.
Reading counts differ from student to student: 0, 1, 2, 3, 5. Collecting and summarizing those varying counts is the statistical task.
Fixed arithmetic: 100. No data collection. The presence of numbers does not make a question statistical.
Although the final answer may be one category (pizza), it requires collecting preference data from many students whose answers vary. One summary answer does not make a question non-statistical.
This is a defined administrative count for a specific moment — one fixed value. Compare: "How many students are absent each day this month?" (counts vary daily) — that would be statistical.
One machine, but 100 varying fill times: 2.3 s, 2.5 s, 2.4 s, 2.6 s… Studying that variation is a statistical task. Many people are not required — varying observations are.
Frequently Asked Questions
It depends on who is being asked. Addressed to one specific person, it produces one categorical response — not statistical. Asked across an entire class ("What are the favorite colors of students in Grade 6?"), responses vary and can be counted and summarized. That version is statistical. Context matters.
For one named person, no — it produces one fixed value. "What are the ages of students in this school?" is statistical because ages vary across students. The scope of the question determines statistical status.
No. Statistical questions can be answered through surveys, physical measurements, experiments, repeated trials, administrative records, sensors, and many other methods. Defining a statistical question as a "survey question" is an oversimplification. The investigation method is chosen after the question is formed.
Variability means that the observations are not all identical. Heights: 152, 160, 165, 158 cm — not all the same. Favorite colors: blue, green, blue, red — not all the same. Both sets contain variability. Variability is not purely random noise; it can arise from genuine differences between people, objects, or conditions, from changing circumstances, or from measurement differences.
A research question may be broadly framed ("What factors affect student achievement?") and not yet ready for data collection. A statistical question is framed specifically enough that data variation can be collected and analyzed to answer it ("How does reported study time relate to test scores among Grade 6 students at this school?"). Statistical questions are a more precise, data-ready form of research questions.
Technically yes, if the average is of a fixed, non-varying set. But in practice, "What is the average…?" almost always involves computing a mean over observations that differ — which makes it statistical. The safer check: identify what observations produce the average. If they vary across people, objects, or trials, the question is statistical.
Grade 6 examples that work well: "How many hours do students in our class sleep on school nights?", "How do students in our grade travel to school?", "What proportion of students in our class own a pet?", "How many minutes do students spend on homework each night?", "Which lunch option do students prefer?", and "What is the distribution of shoe sizes in our class?" These use everyday topics with natural variation that students can observe directly.
Key Takeaways
- Variability is the defining feature. A statistical question anticipates that its observations will differ across people, objects, events, or trials — and requires that variation to be collected and analyzed.
- The final answer can be one number. A mean, percentage, or most common category is still a statistical result if it summarizes varying underlying observations.
- Categorical data counts. Questions producing colors, names, or yes/no responses can be fully statistical when responses vary across a group.
- Many people are not always required. Repeated observations of one person, machine, or location also produce varying data.
- Numbers don't determine statistical status. "How many planets are in the Solar System?" contains a number but is not statistical. "How many books do students read each month?" is.
- The question comes first. It determines what to measure, from whom, and how to analyze the results — the entire investigation flows from it.
- Association is not causation. A statistical question can ask about a relationship between two variables without implying that one causes the other.
For a broader introduction to the field, see what statistics is. For what happens after the question is formed — planning your data collection and study — visit the study design overview.