Separate a constraint from a preference
An unaffordable required payment may be a constraint. A preferred class format may be a preference. Giving both a numerical weight can make a serious constraint disappear inside a favorable total.
Imagine a fictional program that scores highly on research fit and location but requires attendance that conflicts with an unavoidable care responsibility. A high aggregate score does not make the timetable workable.
A smaller matrix can be more honest
| Criterion | Offer A | Offer B |
|---|---|---|
| Required schedule | Confirmed | Partly unknown |
| Training relevant to goal | Strong in one area | Broader, less depth |
| Confirmed support | Written amount and duration | Renewal conditions unclear |
| Unresolved question | Supervisor availability | Placement timing |
This example uses descriptions rather than invented precision. Numerical scores can still be useful if the person understands what they mean and tests how changing the weights affects the ordering.
Unknown is not a zero
A missing answer about funding is uncertainty, not evidence that funding is absent. It is also not a reason to assume the best. Marking the unknown identifies a question worth resolving before commitment.
University requirement and program FAQ pages provide facts to populate a comparison. They do not supply the personal weights. Cornell's degree requirements and Brown's master's FAQ illustrate the kinds of official information involved, rather than a ranking of those institutions.
Try changing one assumption
If a small change in a weight reverses the result, the matrix is revealing a close choice. If an unconfirmed assumption drives the result, obtaining that information may be more useful than refining the scoring system.
A good matrix makes the reasons for a choice easier to explain. It does not make a difficult decision morally or mathematically compulsory.
