Separate Facts, Assumptions, Estimates, and Predictions

Show what you know, what you are trying out, and what you expect to happen.

“Attendance will double next month.”

That sentence sounds like a result. It is actually a forecast. Behind it may be a past attendance count, a larger invitation list, and a guess about how many people will come.

If an assistant blends those ingredients into one confident paragraph, readers may mistake a planning assumption for an established fact.

You can prevent that by labeling the parts. Clear uncertainty does not weaken a useful analysis. It shows the reader where the answer comes from and what could change it.

Four labels, one event

Imagine a fictional learning club whose last event had 60 recorded attendees. The organizer plans to invite those people again and send invitations to 100 new people.

Type	Meaning	Example
Fact	A statement supported by the relevant evidence	The supplied register records 60 attendees at the previous event
Assumption	A value or condition adopted for the analysis	Use a 70 percent return rate for a planning scenario
Estimate	An approximate quantity calculated or judged from available information	Under those assumptions, projected attendance is 62
Prediction	A statement about a future outcome	The next event is expected to attract about 62 people

The same number can appear in an estimate and a prediction, but the claims differ. A scenario calculation says what follows from chosen inputs. A prediction says something about what will actually happen.

Facts also have scope. “The register records 60” is a fact about a record. If the register missed people or counted duplicates, it may not equal actual attendance. Evidence quality still matters.

Write the relationship before filling in numbers

For the fictional club, use:

Expected attendees = previous-attendee pool × assumed return rate + new invitees × assumed response rate.

The central planning assumptions are a 70 percent return rate and a 20 percent response rate among the 100 new invitees.

The arithmetic is 60 × 0.70 + 100 × 0.20 = 42 + 20 = 62.

That is nowhere near a demonstrated doubling to 120. It is a conditional result based on assumptions that still need evidence.

Do not ask an assistant to fill in “reasonable” rates without saying how they will be used. If no basis is available, call them illustrative planning values. A plausible number is not a measured probability.

Expose assumptions that are hiding in the story

The formula assumes that the two invitation groups do not overlap. If 20 of the “new” invitees attended before, adding the groups without correction could double-count them.

It also assumes the organizer can accommodate the projected attendance and that the invitation process reaches the intended people. A list of 100 addresses is not proof that 100 people read a message.

Ask:

List the assumptions required for this calculation. Identify which are supported, which are planning choices, and which could materially change the result.

This is especially useful when a polished narrative skips directly from “we will send more invitations” to “we will have more attendees.”

Use scenarios honestly

Try three sets of assumptions:

Scenario	Return rate from 60	Response rate from 100 new invitees	Calculated attendance
Lower planning case	50%	10%	40
Central planning case	70%	20%	62
Higher planning case	90%	30%	84

The range from 40 to 84 is a scenario range. It comes from values chosen for exploration. It does not mean attendance has a 95 percent chance of falling inside it, and it does not rule out values outside it.

A statistical confidence interval is constructed using a specified method, data, and assumptions about sampling or a model. In its standard frequentist interpretation, the confidence level describes the long-run coverage of the method, not a probability invented for this particular forecast. NIST: What are confidence intervals?.

For this exercise, no statistical interval has been calculated. Calling the range “95 percent confidence” would add authority the analysis has not earned.

Find the assumption that matters most

Change one value while holding the others fixed. With a 70 percent return rate, raising the new-invitee response rate from 20 to 25 percent adds five attendees: 100 × 0.05 = 5.

Raising the return rate from 70 to 75 percent adds three: 60 × 0.05 = 3.

This is sensitivity analysis in plain language. It shows how an input affects the result. It does not establish which input is more uncertain or which outreach action would change it.

If materials cost $3 per attendee plus $100 in fixed expenses, the three attendance scenarios imply total costs of $220, $286, and $352. Those are conditional budgets, not vendor quotes or actual spending.

Now the organizer can see why uncertainty matters: it affects preparation and money, not just wording.

Match the sentence to the evidence

Replace:

Attendance will double, and the event will cost exactly $286.

With:

Using illustrative return and response rates, the central planning case produces 62 attendees and $286 in modeled costs. The three scenarios examined produce 40–84 attendees. Actual attendance and expenses remain uncertain; these are planning calculations rather than measured forecasts.

The revised version still gives a usable answer. It simply distinguishes its basis from its outcome.

Avoid vague hedges that conceal the useful information. “It might possibly be somewhat different” tells the reader less than “A five-percentage-point change in the new-invitee response rate changes the estimate by five attendees.”

Keep an uncertainty register

A short table can identify what to investigate next:

Unknown	Why it matters	Possible way to reduce uncertainty
Overlap between invitation groups	Could double-count potential attendees	Check the lists without exposing private details unnecessarily
Return and response rates	Drive the attendance estimate	Review comparable past events and current responses
Actual material price	Changes the budget	Obtain a current price for the required quantity
Capacity	Limits what can be accommodated	Confirm the usable venue limit

Do not collect information merely because it is available. Focus on uncertainties that could change the decision.

Students: distinguish the observation from the explanation

In a science report, “the plant grew four centimeters” is an observation if measured and recorded appropriately. “More light caused the growth” is a causal interpretation requiring a suitable comparison and method.

In a history essay, a documented date, an author's interpretation, and your own inference play different roles. Label them through clear wording and citations.

Ask an assistant to classify statements in your draft, then check its classifications. This is a way to examine your reasoning, not outsource the responsibility for it.

Your exercise

Classify these ten statements: (1) the register lists 60 attendees; (2) use a 70 percent return rate; (3) use a 20 percent response rate; (4) the two groups do not overlap for this model; (5) 42 previous attendees are projected to return; (6) 20 new attendees are projected; (7) the scenario total is 62; (8) 62 people will attend; (9) materials are assumed to cost $3 each; (10) the modeled central cost is $286.

The intended classifications are: fact about the record; assumption; assumption; assumption; conditional estimate; conditional estimate; calculated scenario result; prediction; assumption; conditional cost estimate. Explain any distinction you make between an exact calculation and an uncertain real-world estimate.

Completion check: Your rewritten forecast exposes its assumptions, uses appropriate units, and contains no invented probability or unsupported precision.