Turn a Broad Topic into a Single Scoped Question
In the last chapter you learned to separate the part of a concern that evidence can answer from the part that rests on values. That is real progress, but the evidence-answerable question you are left with is usually still too wide to research. “How much energy does streaming use?” sounds answerable, yet a hundred different answers would all be correct. This chapter gives you a repeatable way to shrink a broad topic down to one specific question you could actually go and settle.
Narrow a Broad Topic to a Single Question
A broad topic is a starting point, not a research question. Consider “streaming and the environment”, “screen time and sleep”, or “the real cost of eating well”. Each is a whole field, not something a single search could resolve. Before you can gather evidence, you have to choose which slice of the topic you actually mean. That slice is your scope: the specific part of a broad topic that your question covers, and, just as importantly, everything you are deliberately leaving out.
Narrowing is a small, repeatable process, and each step throws away part of the topic on purpose.
Start from the broad topic and say it in one plain sentence.
Pick a single angle, one relationship or one quantity you care about, and set the rest aside for now.
Rewrite that angle as one question, not two or three bundled together.
Check that a realistic search, count, or measurement could move you toward an answer.
Take the topic “streaming and the environment.” One plain angle is energy use. A first narrowing might be “How much electricity does streaming video use?” That already beats the topic, because it names one quantity (electricity) and one activity (streaming video). It is still too wide, but you now have something you can sharpen, which is exactly what the next section does.
Set the Boundaries: Who, What, Where, and When
A question gets its precision from its boundaries: the limits that fix who, what, where, and over what period the question applies. Two people can research the “same” question and reach different answers simply because they never agreed on these limits. Boundaries remove that ambiguity before it wastes your time.
Ask four plain questions of your draft.
Who or what is this about? A typical household, adults over 65, one product category.
What exactly counts? Streaming video specifically, or all internet use.
Where does it apply? On home Wi-Fi, on a mobile network, in one country.
Over what period? This year, the last five years, a single evening.
Add the answers to your draft and watch it sharpen. “How much electricity does streaming video use?” becomes “About how much electricity does one hour of streaming HD video use on a home Wi-Fi connection, for a typical household, in the last couple of years?” It is longer, but every extra word closes a door that would otherwise let in a misleading answer.
Spot and Remove Questions Evidence Cannot Answer
Even a scoped question can be quietly broken in ways that make evidence useless. Two faults are especially common, and both are easy to fix once you can name them.
The first is a smuggled-in answer: a question that hides its conclusion inside the wording, so the “research” can only ever agree with you. “Why is streaming obviously wrecking the planet?” has already decided the verdict, and no honest evidence-gathering could start from there. The second is boundless sprawl, a question so wide that no realistic search could settle it, like “Is technology bad for the environment?”
Here are the patterns to watch for, and the fix for each.
Loaded wording that assumes its own answer (“Why is X clearly the best?”). Strip the verdict and ask the neutral version (“How does X compare to Y on cost?”).
Two questions joined into one by “and” or a comma. Split them and research one at a time.
A vague yardstick, words like “good”, “bad”, or “better” with nothing to measure. Name the specific thing you would count or compare.
No boundaries fixing who, where, or when. Add the limits from the previous section.
Walk Through a Full Narrowing Example
Let’s take one topic all the way through, from a vague starting point to a single scoped question.
Start with something a friend says over dinner: “Streaming is terrible for the environment.” As it stands, that is not even a question, and “terrible” is a values judgment. So you first pull out the evidence-answerable angle you actually care about, the energy a streaming habit uses.
Now narrow it.
Plain topic: the energy used by streaming video.
Single angle: the electricity used while watching, not the energy to build data centers or devices.
One question, not several: how much electricity, per hour of viewing.
Then set the boundaries.
Who: a typical household viewer, not a data-center operator.
What: one hour of HD video, not 4K and not music.
Where: on a home Wi-Fi connection, not a mobile network.
When: the last couple of years, since figures shift as technology improves.
Put it together and you get one scoped, researchable question:
“About how much electricity does streaming one hour of HD video on home Wi-Fi use for a typical household, using figures from the last couple of years?”
Finally, run the fault check. Is there a smuggled-in answer? No, the question does not assume streaming is good or bad. Is it two questions in one? No. Does it have a measurable yardstick? Yes, electricity per hour. Does it have boundaries? Yes, on all four. Compare that with where you started, “streaming is terrible for the environment”, and you can see how much wandering you have just saved yourself. You now have a question you could hand to a stranger, and they would know exactly what evidence would answer it.
Let’s Recap!
A broad topic is a field, not a question; narrowing means choosing one angle and setting the rest aside on purpose.
Scope is the slice of a topic your question covers, and boundaries fix who, what, where, and when it applies.
Adding boundaries makes a question longer but far more precise, because each limit closes off a misleading answer.
A smuggled-in answer hides its conclusion in the wording, and a boundless question is too wide for any evidence to settle.
Name the fault, apply the fix, and a vague topic becomes a single question a stranger could answer.
Your question is now specific and bounded, but “answerable in principle” is not the same as “answerable with real evidence.” Next, you’ll pressure-test whether evidence could genuinely reach your question, and add the measurable element that lets numbers speak.
