ATS2946 Cheatsheet · Critical thinking: How to analyse arguments and improve your reasoning skills

0 · Exam Mapcourse + brief

“This final assessment is open book for unit material. You may access course materials during the course of completing the final assessment.” (ATS2946 Final Assessment page, 2026)

Use this sheet as a desk reference. The final rewards visible reasoning: define the controlling concept, select the fact that activates it, show the inferential link, test a credible alternative, then qualify the conclusion.

Revision cycle that transfers

Allocate revision time to the reasoning operation the task asks for, not merely its topic. Before writing, identify the required product, scope and audience; after writing, change one assumption and decide whether the recommendation should survive.

Final-task priority: the Take Home Final Assessment is 30%. Use the assessment weights as a planning map, then follow the current task instructions for scope, format and submission. This page supplies retrieval and decision structure; it does not replace those instructions.

1 · Argument or Explanation?C1 boundary

2 · Reconstruct the ArgumentC1 · P1–P6

Argument = a conclusion together with one or more premises offered as reasons for accepting it. A premise is a statement presented as support; the conclusion is the claim the audience is asked to accept on that basis.

Extraction sequence

1. Write the main conclusion as one precise claim.
2. List each stated premise separately; strip away examples and qualifications only after preserving what they change.
3. Supply the unstated proposition needed to connect premise to conclusion.
4. Draw support or opposition; show whether reasons must combine.
5. Test a counter-case and name evidence that would justify revision.

Answer spineP1: [material fact] + P2: [stated/implicit link] → C: [qualified claim]

Representation discipline

An argument map is a visual representation of how premises support, oppose or combine in relation to conclusions. Its value is diagnostic: it exposes where evidence ends and inference begins. After drawing it, ask which single arrow is least secure, what counter-case attacks it, and what new evidence would restore it. Preserve the writer’s qualifications before compressing; otherwise the reconstructed conclusion may become stronger than the original claim.

3 · Support, Validity, AssumptionC2

Deductive support aims to make the conclusion unavoidable if the premises are true. Validity is a property of form: true premises cannot occur with a false conclusion. It does not, by itself, establish that the premises are acceptable.

Assumption = an unstated proposition needed for the stated reasons to support the conclusion as intended. Recover it by asking: what must also be true for this move to work? Then reverse or deny that proposition and trace whether the conclusion survives.

3b · Fallacy DiagnosisC2 boundary

A fallacy is a recurring pattern of defective reasoning whose weakness must be explained in context. Naming the fallacy is not the evaluation. Show the precise damage: which reason becomes unacceptable or irrelevant, which option was wrongly excluded, or why the stated support cannot carry the conclusion.

Required sentence“This move weakens [acceptability / relevance / support] because …; if [counter-case/evidence], the conclusion should be revised.”

Four-pass defect test

4 · Complex Support StructuresC3

Intermediate conclusion = a claim supported by earlier premises that also functions as a premise for a later conclusion. Mark it twice: first as an output, then as an input. If it fails, every later claim depending on it loses support.

Linked support requires premises to be combined before they provide the intended reason. Removing either linked premise defeats that reason. Keep independent reasons on separate lines so one may survive when another fails.

Build the complete map

Number every proposition before drawing arrows. Start at the final conclusion and work backwards: which claim immediately supports it, and what supports that claim? An intermediate conclusion should appear once as the endpoint of earlier reasons and again as the premise for the later step. If reasons are linked, mark the joint support and test the pair together; if they are independent, test each route separately.

5 · Normative InferenceC3 · P3–P5

A normative claim states what ought to be done, valued or permitted, not merely what is the case. A principle is the general evaluative standard connecting descriptive facts to that normative conclusion.

Normative bridgeFact: [harm/benefit is supported].
Principle: [general evaluative standard].
Therefore: [actor] ought to [action], unless [review condition].

Boundary: descriptive evidence cannot complete a normative inference until the relevant value or principle is stated and defended. Test at least two feasible options against the same criteria and stakeholder positions; otherwise the comparison is one-sided.

When it bites

Use this frame whenever evidence about harm becomes a policy recommendation. Mark each intermediate conclusion; expose the contested value principle; then change one condition and ask whether the recommendation should travel.

5b · Digital Evidence ProcessC5 process check

6 · Statistical GeneralisationC4

Sample = observed cases used to support a claim about a wider population. Representativeness asks whether relevant features are preserved. Sampling bias is systematic selection that makes observed cases differ from the population.

Size ≠ representativeness. A larger sample reduces random variation but cannot repair systematic selection bias. For a voluntary poll, ask separately about random variation and whether volunteers differ systematically from the target population.

7 · Causation + Digital EvidenceC5

A causal claim says changing one factor produces or helps produce change in another. Correlation is measured association and alone does not identify which factor produces change. A confounder is associated with both proposed cause and outcome and can mislead.

8 · Universal Answer Frameall chapters

1. Decision: answer the question before displaying theory. Name the claim, actor and scope.
2. Definition: give the exact controlling concept, not a family of labels.
3. Application: select the material fact and explain the mechanism connecting it to the defined element.
4. Comparison: test a credible alternative under the same criteria; state what it excludes or preserves.
5. Qualification: identify uncertainty, a review condition and the evidence that would require revision.

Paragraph frame“The better view is [decision]. [Concept] means [definition]. Here, [fact] activates [element] because [mechanism]. However, [alternative] changes [link] because [reason]. Therefore [qualified C], unless [review trigger].”

Select and connect facts

Transfer check: reverse the strongest fact. Trace the change through the defined element, mechanism and conclusion. Revise only if it changes evidence, causal link or boundary; otherwise explain why the conclusion holds.

9 · Drill: Scattered Public ClaimC1 worked · 4 marks

Prompt: reconstruct a public claim whose reasons are scattered across examples, qualifications and implied assumptions. The four-step allocation below is practice only, not an official University scheme.

Compact application

Classify: state which sentence is the conclusion and which scattered statements function as premises. Reconstruct: add the missing proposition that makes those reasons bear on that exact claim. Test: treat any example as support only if it changes acceptability, relevance or strength. Conclude: keep the result inside the boundary: reasons to believe a claim are not the same job as explaining why an event occurred.

End productP1 [stated reason] + P2 [implicit bridge] → C [precise public claim]; counter-case [changed fact] → revise [link/C].

Map audit

Argument map = a visual representation of how premises support, oppose or combine in relation to conclusions. Use it as a constraint: state the supported consequence, remaining uncertainty and revision condition.

10 · Drill: Workplace ProposalC2 worked · 4 marks

Prompt: evaluate a workplace proposal that relies on a false choice and an unstated prediction about staff behaviour. The allocation is practice only.

Compact application

The proposal presents two options as exhaustive. First identify the excluded alternative rather than merely writing “false choice”. Next state the unstated prediction about staff behaviour that must be true for the proposal to work. Suppose every stated premise is true: if the conclusion can still be false because another live option or behaviour remains, the deductive support fails. Explain the effect on relevance, acceptability or support.

Finish the diagnosis“The label matters because [excluded option / denied assumption] leaves open [counter-case], so the premises do not make C unavoidable. Evidence about [staff behaviour] would justify revision.”

Apply each test separately

Trap: naming a fallacy is not an evaluation. Identify the move, then explain how it weakens relevance, acceptability or support; state the evidence that would repair the defect.

11 · Drill: Policy RecommendationC3 worked · 4 marks

Prompt: map a policy argument moving from evidence about harm to a recommendation through a contested value principle. The allocation is practice only.

Compact application

Separate the claim that harm exists from the claim that a policy ought to follow. Mark the intermediate conclusion supported by the evidence. Then state the principle that converts that conclusion into a recommendation and defend it. Compare a credible rival option under the same evaluative standard; do not change criteria to favour one side. Finish with a conditional recommendation and the changed fact that would reopen it.

Chainevidence of harm → intermediate C
intermediate C + defended principle → normative C
rival principle / changed condition → retest action

Symmetrical comparison

Linked support means premises must be combined before they give the intended reason. Test each alone and together; if one is removed, state whether the intermediate conclusion and recommendation still follow.

12 · Drill: Voluntary Online PollC4 worked · 4 marks

Prompt: evaluate a public claim drawn from a small voluntary online poll. The allocation is practice only.

Compact application

The observed cases are voluntary respondents; the conclusion concerns a wider population. Identify the feature on which volunteers may differ and explain why it matters to the claim. Treat small size and voluntary selection as different defects: more cases can reduce random variation, but systematic selection can persist at any size. The responsible conclusion is limited to the observed cases unless evidence establishes relevant representativeness.

Sentence frame“Because entry depended on [selection process], the sample may differ from [population] on [relevant feature]. Increasing size alone would not remove that systematic difference.”

Generalisation audit

Representativeness = how far a sample preserves features relevant to the population inference. Name the feature, show how voluntary entry distorts it and limit the conclusion’s scope.

13 · Drill: Causal HeadlineC5 worked · 4 marks

Prompt: assess an article combining a platform trend, an algorithmic summary and a strong causal headline. The allocation is practice only.

Application: treat the trend as association unless changing the proposed cause is shown to produce the outcome. Test platform selection and factors associated with both. Audit the source and each transformation; then name the alternative cause and evidence that could distinguish it.

Qualified end“The evidence supports inquiry into X–Y, not the unqualified causal headline, because [selection/confounder/alternative] remains. Revise if [specified evidence] excludes it.”

14 · Fast Recall + Trap GridC1–C5