Costs
Part of How to read recommended books critically
A business book claim checklist: five kinds and twelve checks
A business book claim checklist of five claim kinds and twelve checks: sort the claim, ask which cases were looked at, and ask what would prove the author wrong.
Business writing makes claims, and the claims are not all the same kind. A definition, a mechanism, a correlation, a prescription and an anecdote each support a different amount of weight, and mistaking one for another is how a reasonable book produces an unreasonable decision.
This is a checklist for a claim rather than for a book. It works on a chapter, a talk, a post, or a paragraph someone quotes at you in a meeting.
What to take away
- Sort the claim before you evaluate it. Most disagreements about business advice are disagreements about what kind of claim was being made.
- An anecdote can show that something is possible. It cannot show that it is typical, and it certainly cannot show that it will work for you.
- The prescription is the part that costs you something, and it is usually the part with the least support behind it.
The five kinds of claim
| Kind | What it asserts | What it can support | What it cannot |
|---|---|---|---|
| Definition | This word means this | Clarity, shared terms | Any factual conclusion |
| Mechanism | This causes that, by this route | An explanation to test | That it happened here |
| Association | These tend to occur together | A place to look | A direction of cause |
| Anecdote | This happened once | Possibility | Frequency or typicality |
| Prescription | You should do this | Nothing on its own | Anything, without one of the above |
The bottom row is the finding. A prescription borrows its authority from
Five Kinds of Claim
Kind
- Definition
- word means this
- Mechanism
- causes that, by route
- Association
- occur together
- Anecdote
- happened once
- Prescription
- you should do this
Asserts
- Definition
- clarity, shared terms
- Mechanism
- explanation to test
- Association
- place to look
- Anecdote
- possibility
- Prescription
- nothing on its own
Can support
- Definition
- factual conclusion
- Mechanism
- that it happened here
- Association
- direction of cause
- Anecdote
- frequency or typicality
- Prescription
- anything, without above
Cannot
- Definition
- Mechanism
- Association
- Anecdote
- Prescription
The twelve checks
- Which kind is this?Sort it into the table above before anything else.
- Who is the claim about?A claim about large organizations is not a claim about a business with four people, whatever the wording. The
Twelve Checks for a Claim
- Which kind is this?
- Who is the claim about?
- What would opposite advice look like?
- Which cases were looked at?
- What happened to those that failed?
- Is the mechanism stated or asserted?
- Outcomes or opinions?
- What is the source of the story?
Where the genre reliably fails
Most business writing works backward from an outcome. Something succeeded, the
author looks for what it did differently, and the difference becomes the principle.
The move is not illegitimate; it is how hypotheses are generated. It becomes misleading when the hypothesis is presented as a finding, because the search only ever ran over successes. The recommended books guide covers the types of book and what each can honestly show.
A second failure is the shift from description to prescription between chapters. The description may be careful and the prescription arrives without new support, carried by the momentum of the story. Marking the sentence where that happens is a useful reading habit.
Applying the checklist without becoming useless
Twelve checks is too many for ordinary reading, and running all of them on
Three Checks for Ordinary Reading
What kind of claim is this?
sort it into the five kinds
read on, but mark it
everything will simply stop you reading.
The working version is three: what kind of claim is this, what cases were looked at, and what would make the author wrong. Those three catch most of it. The full list is for a claim you are about to act on, which is a much smaller set.
What to do with a claim that passes
Treat it as a hypothesis about your situation, not a conclusion. Turn it into a cheap, reversible test, and write down in advance what result would count as success. That turns reading into evidence, the only way a book becomes useful. The decision record fields hold the before and after.
Common questions
Does this mean business books are worthless?
No. They are good at vocabulary, at mechanisms worth testing, and at showing that something is possible. They are weak at telling you what is typical and weaker at telling you what will work for you.
How do I check a claim if the sources are not given?
You usually cannot, and that is the finding. A claim with no traceable source is an assertion, and treating it as one is not cynicism.
What about claims backed by research?
Better, and still worth asking who was studied and whether the result has held up. General discussion of why published results do not always replicate is available in the entry on the replication crisis.
Is there a shortcut?
Ask what the author would count as being wrong. It is one question, it takes seconds, and the answer sorts most of the genre. The business quotes guide applies the same principle to a much smaller unit of text, where the sourcing problem is at its worst.
