A critical thinking model is a framework that breaks good reasoning into parts or steps you can check one at a time. The best known include the RED model (recognise assumptions, evaluate arguments, draw conclusions), the Paul-Elder framework of elements and standards of thought, Bloom’s taxonomy, the Toulmin model of argument and Socratic questioning. Each gives you a structure to follow when an argument or decision is too complex to judge at a glance.
If you are learning about these models because you have a critical thinking test coming up, the RED model is the one to start with, because it underpins the Watson Glaser. You can try the format in a free critical thinking test written in its style.
The theory behind critical thinking
Most accounts of critical thinking share three ideas.
- It is a set of skills. These include interpreting information, analysing arguments, evaluating evidence, recognising assumptions and drawing inferences. Our guide to what critical thinking skills are covers each one.
- It depends on dispositions. Skills are only useful if you are willing to use them, including on your own views. Open-mindedness, fairness and a willingness to change your mind are as much a part of critical thinking as logic.
- It uses standards. Reasoning can be judged against standards such as clarity, accuracy, relevance and logical consistency. Critical thinking means applying those standards deliberately, not just reacting.
Behind these ideas sits a basic distinction between two kinds of reasoning:
- Deductive reasoning moves from premises to a conclusion that must be true if the premises are true. “All invoices over £5,000 need two signatures. This invoice is over £5,000. So it needs two signatures.”
- Inductive reasoning moves from evidence to a conclusion that is probable but not certain. “Every invoice I have checked this month was paid on time, so the next one probably will be.”
Much of critical thinking is knowing which kind of reasoning you are dealing with, and how much confidence it justifies.
The RED model
The RED model is used by TalentLens, the publisher of the Watson Glaser, to describe the skills that test measures.
- Recognise assumptions: notice what is being taken for granted, and separate fact from opinion.
- Evaluate arguments: judge whether arguments are relevant and well supported, without letting your own views get in the way.
- Draw conclusions: reach conclusions that follow from the evidence, and no further.
The model maps directly onto the test. Recognising assumptions has its own section, as does evaluating arguments. Drawing conclusions covers the inference, deduction and interpretation sections. Our guide to what the Watson Glaser test is explains each section with an example.
The Paul-Elder framework
This framework splits critical thinking into three parts.
- Elements of thought: the parts every piece of reasoning contains. These are its purpose, the question at issue, the information used, the interpretations and inferences drawn, the concepts relied on, the assumptions made, the implications and consequences, and the point of view taken.
- Intellectual standards: the tests you apply to each element, such as clarity, accuracy, precision, relevance, depth, breadth, logic, significance and fairness.
- Intellectual traits: the habits of mind that develop with practice, such as intellectual humility, integrity, perseverance and fair-mindedness.
In practice you pick an element and apply a standard to it. For example: “Is the information accurate?” or “Is the point of view fair to the people affected?” It is a thorough model, which makes it well suited to checking an important report or decision.
Bloom’s taxonomy
Bloom’s taxonomy is a model of thinking skills that is widely used in education. In its revised form it lists six levels, from simpler to more demanding: remember, understand, apply, analyse, evaluate and create. It is not a critical thinking model as such, but it is often used to describe one, because analysing and evaluating sit near the top. It is a reminder that you need to understand information before you can judge it. We explore that link further in analytical thinking vs critical thinking.
The Toulmin model of argument
The Toulmin model breaks an argument into six parts, which makes it easy to see where it is strong and where it is weak.
- Claim: the conclusion being argued for.
- Grounds: the evidence offered to support it.
- Warrant: the link that explains why the grounds support the claim. It is often unstated.
- Backing: further support for the warrant.
- Qualifier: how strongly the claim is made, such as “probably” or “in most cases”.
- Rebuttal: the circumstances in which the claim would not hold.
Worked example
A customer service manager argues: “We should keep the helpline open until 8pm. Of the 5,000 calls we missed last month, 1,400 came in between 6pm and 8pm.”
- Claim: the helpline should open until 8pm.
- Grounds: 1,400 of 5,000 missed calls came between 6pm and 8pm. That is 1,400 ÷ 5,000 = 0.28, or 28% of missed calls.
- Warrant (unstated): callers in that window need help at that time and would be served if the line were open.
- Backing: this would be stronger with evidence that those callers did not get through later, for example by checking how many called back the next day.
- Qualifier: a fair version of the claim is “extending the hours would probably reduce missed calls”.
- Rebuttal: the argument weakens if many of the 1,400 calls were the same people trying repeatedly, or if most of them got through the next morning.
Laying the argument out like this shows that the weak point is the warrant, not the figure. That tells the manager exactly what to check before deciding. Spotting an unstated warrant is the same skill as recognising assumptions in the RED model.
Socratic questioning
Socratic questioning is a method rather than a list of parts. You test a claim through a series of probing questions, such as:
- What exactly do you mean by that?
- What is the evidence for it?
- What are you assuming?
- What would someone who disagrees say?
- If this is true, what follows from it?
It works well in conversation and in teams, because it tests an idea without attacking the person who holds it. It is also a good way to question your own thinking.
Which model should you use?
Use the model that fits the job.
| Situation | Useful model |
|---|---|
| Preparing for a Watson Glaser or similar test | RED model |
| Checking an important report or decision thoroughly | Paul-Elder framework |
| Taking apart a single argument | Toulmin model |
| Testing an idea in a discussion | Socratic questioning |
| Planning learning or training | Bloom’s taxonomy |
Whichever you choose, the value comes from applying it regularly until the questions become habit. Our guide on how to improve critical thinking skills shows how to build that habit.
Put the models into practice
Models are easiest to learn when you apply them to real questions. A timed critical thinking practice test lets you use the RED model on test-style questions and check your reasoning against a worked explanation for each one.
