Test Data & Plans

Testing "10" and "20" isn't enough. You must actively try to break your own algorithm by throwing Boundary and Invalid/Erroneous data at it.

The Three Data Types

Imagine an algorithm that asks for a student's test score, which must be between 1 and 100.

Normal

Standard, expected data.

50, 75, 80

Boundary

The absolute extreme valid edges.

1, 100

Invalid / Erroneous

Rule-breaking OR wrong type.

200, 0, "Ten"

Examiner's Eye - The Myth of Invalid vs Erroneous

Historically, "Erroneous" meant wrong data type and "Invalid" meant out of bounds. The official OCR guidance explicitly flags this as a widespread myth. Both terms are now completely synonymous. ANY test data that breaks the rules is considered both Invalid and Erroneous. (Also, you no longer need to know the phrase "invalid boundary"!).

Example Test Plan (Rule: Score 0-100)

Test Number Data Category Data Used Expected Outcome
1 Normal 50 Data accepted.
2 Boundary 0 and 100 0 accepted. 100 accepted.
3 Invalid/Erroneous 101 or "Hello" Type/Range error. Loop triggers.

The Validation Sandbox

Validation Rule: 1 <= Age <= 100.
Type absolutely anything into the box below. The engine will instantly classify your test data.

Awaiting Data...

Check Your Understanding

1. An algorithm asks for a User Rating between 1 and 5 (Integer). A tester inputs '99'. What category of test data is this?

2. The rule is "Password must be exactly 8 characters long". Which of the following is an example of Boundary Data?

Written Exam Scenario (AO2/AO3)

Stretch (Grade 9)

"A login programme requires a string username. The validation rule states the username must be between 3 and 10 characters long, inclusive. Draw a brief test plan table providing exactly one piece of Normal, Boundary, and Invalid test data that could be used." (3 marks)