Subjects · Economics, Business & Finance
Economics, Business & Finance: the methods that matter most here
Know which methods suit a subject where a model's most important feature is what it's holding equal, and where "contested" and "settled" sound identical coming out of the same fluent mouth.
What you'll be able to do: Know which methods suit a subject where a model's most important feature is what it's holding equal, and where "contested" and "settled" sound identical coming out of the same fluent mouth.
What makes economics hard, specifically
A model is not the world, and its power depends on what it drops. Supply and demand is a real, useful model, and it is deliberately silent on expectations, market power, information and time. Students learn the model and forget it's a simplification with a scope; then they apply it outside that scope and get a confidently wrong answer.
Assumptions are stated in two words and unstated in twenty. "Ceteris paribus" (other things equal) is the whole discipline compressed into a phrase students learn to say without ever listing what "other things" means for the case in front of them.
A graph is an argument, not a picture. The slope of a demand curve is a claim about causality; the choice of axis, scale and starting point on a chart is a rhetorical decision. Reading a graph correctly and reading it critically are the same skill, and most students are taught only the first half.
Identification is the whole game and it's invisible in intro courses. Correlation is cheap. What distinguishes your causal story from the obvious alternative (reverse causality, a confound, a coincidence of timing) is usually the actual content of a good answer, and intro material mostly skips straight past it to the conclusion.
Positive and normative claims get silently merged. "This policy will reduce employment" is testable. "This policy is a bad idea" smuggles in a judgement about whose loss counts against whose gain. Students answer normative questions with positive analysis and don't notice the seam.
Accounting is a language, not arithmetic. The accounting equation, the T-account, the journal entry are a grammar for describing what happened to a business. Students who treat it as sums get every number right and can't say what a transaction means.
"Who bears the cost" is usually the whole question. Who writes the cheque and who actually pays are routinely different people, and most of tax, trade and pricing analysis is working out the gap between the two.
The ten
| # | Method | Why it earns its place here |
|---|---|---|
| 1 | C7 Ceteris paribus / assumption surfacing | The two words that carry the discipline; the marks and the real errors both live here |
| 2 | I1 Multiple representations | Equation, graph, table and words are one model, not four skills |
| 3 | M10 Chart deception analysis | A graph in this subject is an argument, and it's the cheapest argument to rig |
| 4 | M4 Plot my own data | Turn your own numbers into the model you're taught, and see it not quite fit |
| 5 | C5 Steelmanning | The subject's most interesting questions are the ones smart, informed people still disagree about |
| 6 | K5 Bias probing | Framing changes the answer here more than in most subjects, and you should know by how much |
| 7 | L11 Field disagreement mapping | The core hazard of the subject in one method: find out, don't guess, whether specialists agree |
| 8 | D2 Multi-actor perspective shift | Who bears the cost is usually the whole question, and it has a different answer for each actor |
| 9 | D13 Cause-consequence mapping | Identification: naming and ruling out the obvious alternative explanation |
| 10 | H6 Grading simulation | Positive/normative and identification are exactly where essay marks are lost |
The methods that work badly here
Glossary or flashcard memorisation of terms.
"Elasticity", "opportunity cost", "comparative advantage" all look like vocabulary, so students drill them like vocabulary. The failure is specific: you can produce a word-perfect definition of comparative advantage and still use it to mean "being better at things": the single most common conceptual error in the subject, and a flashcard cannot catch it, because the flashcard is exactly what taught you the word without the relation the word sits inside.
A term in economics means nothing outside its model. "Elasticity" is not a fact to recall; it's a ratio computed from a curve, and it changes value along a perfectly straight demand curve even though the slope doesn't, which is unlearnable from a definition and obvious from the graph. Learn the model; the vocabulary comes for free.
Asking the AI a policy or empirical question and accepting the answer.
This is the default way most people use a chat tool, and it is the specific failure this subject's primary article opens with. A fluent, complete, well-organised answer sounds identical whether the underlying question is settled or one specialists are still fighting about, and the default conversational mode gives you no signal to tell them apart. Every method below exists partly to put a check in front of that default.
The composed workflow
- Before accepting any conclusion: ask whether this is contested or settled among specialists, and how you'd check that independently (L11).
- State the model in at least two representations together, and its ceteris paribus conditions (C7 + I1).
- The moment a "should" appears, separate it from the positive claim next to it (feeds H6).
- Ask who bears the effect specifically, not what happens on net (D2).
- If it's contested, get the strongest version of the side you don't hold (C5), and check the answer doesn't just track the framing you used to ask it (K5).
- Before trusting a chart or a number, check what the axis, scale and date range are doing to the argument (M10), and where you can, replace it with your own data (M4).
That covers modelling, representation, disagreement, incidence and identification, which is most of what "thinking like an economist" actually means, and none of it is a memorised definition.
The ten articles in this subject
01 Ceteris paribus · 02 Multiple representations · 03 Steelmanning contested policy · 04 Bias probing and framing effects · 05 Chart deception · 06 Plot my own data · 07 Field disagreement mapping · 08 Who bears the cost · 09 Identification · 10 Positive versus normative