
You can now judge whether a number fits, surface what it assumes and where its limits lie, and sidestep the correlation and trend traps. All of that careful appraisal is only worth something once it produces a clear statement: given everything you have found, what does the evidence honestly support? This chapter builds the bridge from a thorough appraisal to an honestly stated claim, carrying the right amount of confidence and no more.
By now you can take a number apart. But an appraisal that stays a private tangle of doubts helps no one decide anything. The final move is to gather what you found, the fit, the assumptions, the limits, the traps you checked, and say plainly what it all adds up to.
That statement has two halves that always travel together. The first is the claim itself: what, on balance, the evidence supports. The second is your confidence: how strongly it supports it. A claim offered without a confidence level invites the reader to supply their own, usually a more certain one than you intended. Stating both is what keeps an honest appraisal honest all the way to the finish.
The word that does the work in this chapter is confidence, used in a specific way.
Calibrated confidence is not a mood; it is a reading you can defend. Every claim you make should carry a level and a reason in the same breath: not just “I am fairly confident,” but “I am fairly confident, because three separate months point the same way and the main assumption holds.” The reason is what lets someone else check whether your confidence is earned.
Once you have appraised your evidence, place your confidence on a simple three-step ladder. Each rung has a signature you can recognize and a way to phrase it.
Confidence | When it fits | How to say it |
High | Several independent pieces point the same way, the assumptions hold, and the limits you found are minor | “The evidence strongly supports…” |
Medium | The evidence leans one way but rests on an assumption or two, or comes from a single decent source | “The evidence suggests…, though it depends on…” |
Low | The evidence is thin, mixed, or drawn from a weak or self-selected sample, and real limits remain | “There is an early sign that…, but it is far from settled.” |
The ladder is a guide, not a scale to calculate. What matters is that the words you choose, “strongly supports,” “suggests,” “there is a sign,” match the rung your evidence actually sits on, and that you can name why.
Two situations call for extra honesty. When the evidence is thin, say so and label the claim low confidence rather than inflating it to sound decisive. And when a piece of evidence points against your claim, name it out loud and let it do its work, lowering your confidence or, if it is strong enough, flipping the claim altogether. Hiding the inconvenient piece is the quiet form of overclaiming.
If my evidence turns out to be weak, should I just stay quiet about it?
Not at all. A weak finding, honestly labeled low confidence, is still useful, it tells whoever is deciding how much weight to put on it. What misleads people is not weak evidence; it is weak evidence dressed up as strong.
Picture someone who switched their weekly shop to a discount grocery store and now wants to claim, “switching stores is saving me money.” They have three pieces of evidence: last month’s total was lower than the month before; a friend swears discount stores are always cheaper; and they recall fewer painful moments at the till. How should they state what this honestly supports?
Start with the honest claim. The evidence leans toward a saving, but it is thin and partly shaky. One month against one month is a short time frame, the kind a single unusual week can swing. The friend’s opinion is not their own data. And a memory of fewer painful bills is exactly the sort of impression that forms whether or not it is true.
There is also a piece pointing the other way, once they look for it: that lower month was one where they skipped a big quarterly stock-up, so part of the drop is what they did not buy, not what they saved per item. Named honestly, that lowers the claim rather than sinking it.
So the calibrated statement is not “discount stores save me money,” stated flat as a fact. It is closer to: “My first month suggests I spent less, but this is low confidence, one month is too short to trust, and some of the drop came from buying less that month rather than paying lower prices. I will compare three similar months before treating the saving as real.” That sentence carries a claim, a confidence level, a reason, and the limit that caps it, which is everything an honest appraisal owes its reader.
That is the whole move. State what the evidence supports, attach the confidence it has genuinely earned, give the reason, and let the awkward pieces lower the claim instead of vanishing from it.
An appraisal is only finished when it becomes a clear claim paired with how confident you are in it.
Confidence is calibrated when it matches the strength of the evidence, and every level should come with a reason.
Overclaiming stretches a claim past its evidence, while refusing to commit at all wastes the appraisal just as surely.
Use a simple high, medium, low ladder and choose words that match the rung your evidence sits on.
When evidence is thin or points the other way, say so plainly and let it lower or flip the claim.
You can now appraise a piece of evidence and state honestly what it supports. The final chapter puts the whole method to work on one real-style recommendation from start to finish, and then hands the reins to you.