POLS1009 Chap.3 Research Questions, Puzzles and Concepts
Research Questions, Puzzles and Concepts
Every paper in the course is steered by its research question, so this chapter covers how to find one, test it and build it out of usable concepts. It starts with rule zero: describe carefully before you explain, because a trend in raw counts can vanish once you divide by the number of units that could produce it.
A good question grows from well-informed surprise, and the chapter separates genuine puzzles from three tempting fakes. You will see why variation in an outcome, over time within one unit or across many units at one moment, is a dependable starting point, and where else questions come from: the world, the literature, and replication.
Topics are turned into questions with a boundary, a direction and a finish line, checked for scope against Baglione's six tests, and opened with a word that fixes the kind of paper you write. The chapter closes on concepts and the ladder of abstraction, where narrowing a definition sharpens it but leaves fewer cases to study, and on the step from concept to measured variable.
What this chapter covers
- 01
Rule zero and the denominator problem
- 02
Well-informed surprise and what makes a genuine puzzle
- 03
Three fake puzzles and the Week 6 review shorthand
- 04
Variation over time and across units as starting points
- 05
Sources of questions and what to do when yours is already studied
- 06
Topics versus questions, scope and Baglione's six checks
- 07
Starting words, keeping the answer out, bounding the scope
- 08
Concepts, the ladder of abstraction and the step to measurement
Worked example · free
Choosing a rung on a ladder of abstraction for participation
- 1Attributes. Each rung down adds one attribute: acting to influence government (1), doing so through elections (2), casting a national ballot (3).
- 1Coverage. Cases fall from 820 to 610 to 540, so the first step removes 210 residents and the second removes 70.
- 1Fit. The question is about election activity of any kind, so electoral participation, covering 610 residents, is the rung that matches it.
- 1Above and below. Political participation would add 210 residents whose activity lies outside elections; voting in national elections would drop 70 residents who campaign or vote in other elections, cutting away part of the phenomenon.
Key terms
- Rule zero
- A working rule that tells you to confirm, through careful description, that a pattern really occurred before you try to explain why it occurred.
- Fake puzzle
- An observation that looks surprising but does not defy a real expectation, such as ordinary variation, one deviant case or a misbehaving corner of a strong pattern.
- Concept
- A basic unit of reasoning that groups facts under a name, from which research questions, theories and hypotheses are all assembled.
- Explanatory question
- A research question, usually opening with why, how, under what conditions or to what extent, that asks for the causes of an outcome rather than a description of it.
- Denominator problem
- The error of reading a change in raw counts as a real trend without dividing by the number of units able to produce the event, such as people or countries.
Research Questions, Puzzles and Concepts FAQ
What makes a good research question in political science?
The course uses six checks: interesting to you, important to others, counterintuitive to informed readers, short, direct rather than multi-part, and doable with your time, data and skills. Scope matters too: a question about all governments is a career, while one about a single decision on a single day is too thin for a paper.
What is the difference between a research puzzle and a fake puzzle?
A genuine puzzle runs against established knowledge, accepted wisdom or a theory's prediction, so it surprises people who know the field. Fake puzzles only look odd: outcomes that differ from place to place, a lone exception to a probabilistic pattern, or a small sub-sample that behaves differently from the overall trend.
Is it bad if someone has already studied my research question?
Usually not. Prior work suggests the idea is worthwhile and gives you a literature to engage with. You need to improve on existing answers in at least one way, through a better theory, stronger evidence or a fairer test. A question with no prior work deserves more suspicion, since it may be trivial, obvious or unanswerable.
Can a research question be descriptive or normative?
Yes, but it produces a different kind of paper. Who, what, when, where and does questions lead to descriptive analysis, and should questions lead to normative debate. Both have value. A paper meant to explain an outcome needs a why, how, under what conditions or to what extent opener.
How do I narrow the scope of a research question?
Draw the boundary for a reason you can defend. A limit is justified when the phenomenon exists only in that place or period, or when reliable data cover only that span. Choosing cases because they fascinate you is risky, because what makes them fascinating may be linked to the variables you study.
Exam move
Practise on questions you invent. Write five research questions on any political topic, then put each through the same routine: apply rule zero by asking what the base would be if the pattern were expressed as a rate, decide whether the surprise would impress someone who knows the literature, check the starting word, and make sure no favourite answer is hiding inside the wording.
Score each against the six checks and rewrite the weakest. For concepts, take one key term from each question and draw a three-rung ladder, writing down the attribute you add at each step and naming a case that drops out, then decide which rung your question really needs.
Working through Research Questions, Puzzles and Concepts in POLS1009? Sia is AskSia’s AI Arts and Humanities tutor — ask any POLS1009 Research Questions, Puzzles and Concepts question and get a clear, step-by-step explanation grounded in how POLS1009 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.