Problem Solving & Quality · Observe

Fact–Opinion Sort

Tag every statement on the board as measured fact, observed fact, assumption or opinion — then go and check the assumptions that matter.

Time30 min
FormatTeam
StageObserve

Fact–Opinion Sort: what it is and why it works

The Fact–Opinion Sort classifies every statement made about a problem into one of four categories: measured fact (from a record or instrument), observed fact (seen by a named person), assumption (plausible but unchecked) or opinion (a judgment or preference). Each fact is given its source and date. The team then circles the assumptions that would change the analysis if they turned out to be false, and assigns each one a quick check. A note moves into a fact column only when evidence comes back.

Problem-solving meetings mix data, memory and conviction on the same board, and after a few sessions the most often repeated hypothesis starts to be treated as settled. The sort makes that drift visible and cheap to correct. It does not argue with anyone; it only asks for a source. It is quicker than a full fishbone and works well as a checkpoint before root-cause analysis, right after a gemba walk has produced raw observations, or in the middle of an investigation that has stalled. The assumptions it flags often become candidates for on/off cause verification, and its verified facts feed the Ishikawa diagram, where each cause is marked verified, refuted or still open.

What you need

All statements, theories and data points collected so far
Sticky notes in at least two colors, and a board with four columns
Access to records and instruments to trace sources
The people who made the statements, or their notes

What you get

A board sorted into measured facts, observed facts, assumptions and opinions
A source and date for each fact
A short list of critical assumptions
A check action with an owner for each critical assumption
A visual reading of how solid the analysis is

When to use it

When the loudest hypothesis has quietly become the official truth.

How to do it, step by step

Write every statement made about the problem on its own sticky note.
Sort the notes into four columns: measured fact, observed fact, assumption, opinion.
For each fact, add the source — record, instrument, witness — and its date.
Circle the assumptions that, if false, would change the analysis.
Assign a quick check to each circled assumption and move the note only when evidence comes back.

Worked example: High-pH excursions at an industrial wastewater plant

Illustrative scenario — figures are realistic but not from a real company.

The wastewater treatment plant at a chemicals site has exceeded its pH 9.0 discharge limit four times in two months. After three meetings, the team's official explanation is that 'the acid dosing valve is sticking'.

The facilitator wrote each of the 22 statements from the meeting notes on its own note, including 'the valve is sticking', 'it only happens at night' and 'Unit B dumps cleaning solution without warning'.
Sorting gave six measured facts (effluent pH trend, acid dosing totals, four exceedance timestamps), five observed facts, eight assumptions and three opinions. 'The valve is sticking' turned out to be an assumption: nobody had tested it.
Sources were added to each fact: the pH trend from the historian, exceedance times from the permit log, the night-only observation from shift logs.
Two assumptions were circled as critical: that the valve sticks, and that Unit B discharges cleaning solution at night. A valve stroke test and a review of Unit B's batch records were assigned.
The valve stroked normally under test. Unit B's records showed caustic clean-in-place discharges within an hour before three of the four exceedances.

Result. The analysis moved from the valve to the scheduling of Unit B's caustic cleaning discharges. Routing them to an equalization tank and adding a call-ahead rule ended the exceedances over the following quarter. The team kept the board's color mix as a review habit: facts outnumbered assumptions only after the checks came back.

Open it in the FieldKit Problem Solving app → Get the Problem Solving template pack — join the waitlist

Common pitfalls and how to avoid them

Treating one person's observation as a measured fact.Keep the categories separate and record who observed what, and when; observations can be wrong without anyone being dishonest.
Arguing too long about categories.When in doubt, classify the note as an assumption and check it.
Trying to check every assumption.Circle only those that would change the conclusion if false, and check those first.
Discarding opinions.Opinions often carry experience; rewrite the useful ones as testable assumptions instead of ignoring them.

Frequently asked questions

What is the difference between a fact and an assumption in problem solving?

A fact can be traced to a source: a measurement, a record or a named observer at a known time. An assumption is a statement believed to be true but not yet checked against such a source. Assumptions are not wrong by definition, only unverified. The aim is to check the ones that matter before the analysis relies on them.

Why separate measured facts from observed facts?

Because their reliability differs. A measured fact comes from an instrument or a record with known accuracy. An observed fact depends on a person's perception and memory, which can be affected by timing, attention and expectations. Both are valuable, but when they conflict, the team needs to know which is which and check both.

How does a fact–opinion sort help avoid confirmation bias?

It makes every statement show its source. Once the favorite hypothesis is labeled as an assumption on the board, everyone can see it has not been tested, and a check can be assigned. The sort also puts the facts that contradict the hypothesis in plain view, which is exactly what confirmation bias tends to push aside.

Origin

SiliciumHex original, in the spirit of TQM “management by facts”.

Related methods

More in “Observe”

Go where the problem happens and collect facts, counts and timelines — not opinions.