How to Think Critically and Make Well-Reasoned Decisions

A guide to Framing the Question and Reconstructing the Argument, Gathering and Evaluating Relevant Evidence, Testing the Reasoning and the Strongest Alternatives, Choosing an Action and Reviewing the Reasoning

Written by Prof. Julian Sorel, Ethics & Philosophy

Last updated: November 6, 2026 Sources listed

Download checklist (PDF)

Interactive Step Checklist0 of 20 steps completed (0%)

Critical thinking is not permanent skepticism or winning every debate; it is careful, goal-directed thinking about what to believe or do. This 20-step guide in our Philosophy and Religion collection turns that idea into a practical worksheet, drawing on the Stanford Encyclopedia of Philosophy's overview of critical thinking. Use it for everyday claims and decisions, while seeking qualified professional advice where specialist stakes require it.

How do you think Critically and Make Well-Reasoned Decisions?

Overview: How to Think Critically and Make Well-Reasoned Decisions

Howcompass | How to Think Critically and Make Well-Reasoned Decisions

Method1

Framing the Question and Reconstructing the Argument

Bearded man in a green sweater sits at a desk writing with a pen, chin on his hand, with icon cards, a blue lamp, mug and plants around him.

Howcompass

Turn a vague topic into one focused question

Write one belief or decision as a focused question. Replace a broad topic such as 'remote work' with a question that has a clear context, such as whether a particular team should adopt a specific hybrid schedule for a trial period. Keep the question neutral enough that more than one answer remains genuinely possible. The same framing habit anchors the other guides in our Philosophy and Religion collection.

3Howcompass

Define ambiguous terms and the discussion's scope

Define ambiguous terms and the scope of the discussion. Explain key words in observable or operational terms, identify who and what the question covers, and set a relevant time frame. If two people use the same word differently, record both definitions instead of allowing a verbal misunderstanding to masquerade as a factual disagreement. Precise shared definitions are also the opening move in our values-clarification exercise.

Woman in a green shirt and navy skirt stands placing sticky notes on three arched wall panels marked with an eye, a network and scales; notebook and lamp below.

Howcompass

Split observations from interpretations and values

Separate observations, interpretations, and value judgments. List what was directly observed or reliably reported in one column, the inferred explanation in another, and the goals or principles used to judge it in a third. Facts can constrain a decision without automatically settling which values should take priority. Our guide to keeping a reading journal uses a similarly divided page so description stays apart from reaction.

Seated woman in green holds a pen to her chin while studying a sheet of four picture cards with arrows pointing down to a summary card; books and mug nearby.

Howcompass

Reconstruct the argument as numbered premises

State the conclusion and the reasons offered for it. Write one sentence beginning 'Therefore…' followed by numbered premises beginning 'Because…', removing rhetoric and unrelated background so the reasoning becomes visible. Our guide to argument reconstruction provides a closer reading method for longer texts.

Woman in green and navy trousers kneels on grass holding a magnifying glass beneath a footbridge made of folded paper cards over a rocky gap, facing an open blue door.

Howcompass

Surface hidden assumptions and possible defeaters

Expose hidden assumptions and possible defeaters. Ask what must also be true for the premises to support the conclusion, then add any unstated assumption to your numbered list. Finish by writing one piece of evidence or one condition that would reasonably make you revise the conclusion; if nothing could, notice that the position may be closed to examination. A timed debate night with friends is a fast way to hear which of your assumptions others do not share.

Method2

Gathering and Evaluating Relevant Evidence

Stacked leather books, an open notebook with pen and a teacup sit on a stone sill by a sunset canal; a hand holding an open book enters at the right.

Howcompass

Read the original source in full context

Consult primary sources and read the surrounding context. Locate the original study, dataset, transcript, policy, or complete passage rather than relying on a screenshot or a secondhand summary. Check publication dates, corrections, methodology, and what the source actually measured before comparing it with the claim being made. Our guide to reading a research paper shows how to work through an original study without being misled.

Two bearded scholars in robes read a large open book at a table in a columned cloister while a standing robed man holds a scroll; a fountain glitters.

Howcompass

Judge each source's expertise, method, and incentives

Assess the source's expertise, method, and incentives. Ask whether the author has relevant subject knowledge, how the information was obtained, whether the reasoning can be inspected, and what financial, political, or social interests might shape the presentation. A conflict of interest is a reason for closer scrutiny, not automatic proof that a claim is false. The same source checks are set out step by step in our guide to vetting a claim online.

Woman with a dark bun in a green shirt lifts a sticky note toward arched panels showing an eye, a molecule and scales, notebook and lamp on the desk.

Howcompass

Weigh evidence by quality instead of volume

Grade evidence by relevance and reliability rather than volume. Decide whether each item directly bears on the exact claim, whether it was produced by a sound method, and whether independent sources converge without simply copying one another. The philosophical overview of evidence and justified belief is a useful reminder to consider your total evidence, not only the convenient subset. Our guide to scoring a questionnaire honestly models the same discipline of weighing each item rather than counting items.

10Howcompass

Scrutinize samples, baselines, and statistics

Inspect samples, comparisons, and alternative explanations. Check who or what was included, what baseline or comparison group was used, whether important cases are missing, and whether correlation is being treated as causation. When numbers appear, look for denominators, uncertainty, and absolute as well as relative differences before drawing a practical conclusion. The method checklist in our guide to reading a research paper covers samples, baselines, and misleading comparisons in more depth.

Red-haired woman in a tan sweater shows a card to three smiling friends at a table stacked with cards, an hourglass and cups, bookshelves and sunset windows behind.

Howcompass

Search for evidence against your first answer

Search deliberately for evidence that challenges your first answer. Use search terms that express the strongest reasonable objection, consult a credible source with another perspective, and record findings with the same care as confirming evidence. Change your confidence when the evidence warrants it; the aim is an accurate map, not loyalty to your initial guess. When the challenge comes from other people's convictions, our guide to conversations across different beliefs shows how to hear the objection without turning it into a fight.

Method3

Testing the Reasoning and the Strongest Alternatives

Kneeling woman in green uses a magnifying glass over a bridge of folded paper cards spanning a chasm, with an open door and potted plant on the far cliff.

Howcompass

Test whether the conclusion follows from the premises

Test whether the conclusion follows from the premises. For a deductive argument, ask whether accepting the premises would leave any possible room for the conclusion to be false; for an inductive argument, ask how strongly the evidence raises the conclusion's probability. Keep the separate question of whether the premises themselves are true or justified visible on the page. Arguing both sides at a debate night with friends is a practical stress test for whether a conclusion really follows.

Seated woman in green rests a pen near her face while looking at picture cards joined by arrows to a lower card, beside stacked books and a blue mug.

Howcompass

Explain the faulty step instead of labeling it

Check for fallacies without using labels as a shortcut. Look for false dilemmas, circular reasoning, cherry-picking, irrelevant attacks, and hasty generalizations, then explain exactly where the inference fails. Use our logical-fallacies guide and the Internet Encyclopedia of Philosophy's fallacy overview to inspect the reasoning rather than merely attaching an impressive-sounding name.

Woman in green kneels on grass at a cliff edge, magnifying glass in hand, examining a bridge of folded paper cards spanning a gap toward an open blue door.

Howcompass

Build counterexamples that target the real claim

Construct counterexamples and edge cases for broad claims. Try a case where the key definition becomes unclear, a situation where the proposed rule produces an unacceptable result, or a similar scenario where the stated principle would not apply. A counterexample should target the actual argument, not an exaggerated version that is easier to defeat. Explaining your reasoning to peers in a study group often surfaces the edge cases you would miss alone.

Robed men gather around a table in a stone arcade at sunset; one seated scholar holds a book open while another stands with a scroll, near a fountain.

Howcompass

Rewrite the opposing case at its strongest

Steelman the strongest reasonable opposing position. Rewrite the other side using its best evidence, clearest definitions, and most charitable interpretation until a thoughtful supporter could accept your summary. Then compare that version with your own position; if the disagreement remains, identify the precise premise, fact, or value that separates them. The same charitable exchange is the core skill in our guide to communicating better and resolving conflict.

Four laughing people sit around a wooden table with a card game, an hourglass, coffee cups and an open book, in front of tall windows showing a sunset town.

Howcompass

Probe remaining gaps with Socratic questions

Use Socratic questions to test what remains unclear. Ask 'What do we mean?', 'How do we know?', 'What follows?', 'What is the strongest objection?', and 'What would change this conclusion?' Our guide to practicing the Socratic method helps keep the exchange curious rather than hostile; suspend judgment when the available evidence is not yet sufficient.

Method4

Choosing an Action and Reviewing the Reasoning

Woman with a dark bun in a green shirt raises a sticky note toward arched panels marked eye, molecule and scales, notebook and lamp on the desk.

Howcompass

Map everyone affected by the decision

Map the people, duties, and consequences affected by the decision. Include people who are absent or less powerful, distinguish direct from indirect effects, and note rights, promises, fairness, and long-term consequences alongside immediate benefits. Our stakeholder-matrix guide provides a structured way to make these ethical considerations visible.

Three young people lean over an inlaid board game on a wooden table, one man placing a piece while a blonde woman holds cards, with wine, beer and coins nearby.

Howcompass

Generate genuinely different options before choosing

Generate several genuinely different options before choosing. Include a small reversible trial, a delayed decision with a deadline for new evidence, and a way to reduce the worst foreseeable harm where these are feasible. Avoid treating the first two ideas that came to mind as an exhaustive choice. Our guide to building an interest inventory demonstrates the same principle of widening the option set before narrowing it.

Woman in green sits with a pen at her chin before a sheet of four photo cards with arrows pointing to a chart card; books and mug nearby.

Howcompass

Compare options against explicit criteria

Compare the options with explicit criteria and uncertainty. Choose a short list of relevant criteria, explain their priority, and rate each option using the evidence you have while marking unknowns rather than inventing precise scores. Test how the ranking changes under a reasonable pessimistic and optimistic assumption before committing to an irreversible course. Our guide to scoring a simple questionnaire honestly offers a similar template for rating options without faking precision.

Woman with glasses and a green sweater writes with a pen at a desk lit by a brass lamp, globe and open books before gothic sunset windows.

Howcompass

Record the decision with reasons and a review date

Record the decision, supporting reasons, and a review date. Write what you chose, the strongest evidence for and against it, your level of confidence, and the observable signs that would trigger a change. A short entry in a morning and evening reflection journal can preserve your reasoning so later hindsight does not rewrite it.

Carved dark and light chess pieces stand arranged on a wooden board at a stone sill, a die beside it, sunset town and bridge beyond the arch.

Howcompass

Grade the process separately from the outcome

Review the quality of the process separately from the outcome. When results arrive, ask whether you used the information and methods available at the time, sought meaningful objections, and responded appropriately to uncertainty; good luck does not validate a careless decision, and bad luck does not automatically disprove a sound one. Update your belief and next action using the new evidence, then repeat the process at the point where circumstances changed. A weekly values review journal gives that process review a fixed rhythm and a written record.

Frequently Asked Questions

Question: Does critical thinking mean doubting everything?

No. It means calibrating confidence to the quality and completeness of the reasons and evidence. Some conclusions deserve strong confidence, while others warrant a tentative view, further inquiry, or no judgment yet.

Question: How is a fact different from a value judgment?

A factual claim describes what is or was the case and can be investigated with evidence. A value judgment concerns what is good, fair, important, or ought to be done; evidence can inform it, but the underlying principles also need explicit examination.

Question: What should I do when two credible sources disagree?

Compare their exact claims, evidence, methods, expertise, and relevant context rather than counting sources. Look for the reason behind the disagreement, seek additional independent evidence, and adjust your confidence or delay action when appropriate.

Question: Is critical thinking only for major life decisions?

No. The same short process of framing a question, checking the reasons, and looking for disconfirming evidence improves ordinary choices about what to believe online, what to buy, or how to plan a week. Match the depth of the analysis to the stakes and reversibility of the decision, and practice the habit on small choices so it is ready when larger ones arrive.

Tips

Use a one-page reasoning sheet with headings for question, premises, evidence, objections, decision, and review date so the process remains practical.
When a discussion becomes heated, summarize the other person's position and ask them to correct your summary before offering your rebuttal.
Match the depth of analysis to the stakes and reversibility of the decision; not every small choice needs a full evidence review.

Warnings

Do not use critical-thinking language to dismiss another person's identity, lived experience, or good-faith participation in a discussion.
A fallacy label does not establish that a conclusion is false; it shows a problem with a particular way of supporting that conclusion.
For medical, legal, safeguarding, financial, or other high-stakes specialist decisions, use this process to prepare questions and seek a qualified professional rather than substituting it for expert advice.