Selecting and Using Test Data

This lesson covers the selection and use of test data as required by OCR J277 Section 2.4. To test a program effectively, you need to choose appropriate test data that covers all possible scenarios. OCR requires you to understand three types of test data: normal, boundary, and erroneous/invalid.

---

Why Test Data Matters

Using the right test data is critical because:

• Testing with only "normal" data may miss bugs that occur at boundaries or with invalid input.
• Real users will inevitably enter unexpected data — your tests should anticipate this.
• A comprehensive set of test data gives confidence that the program handles all scenarios.

OCR Exam Tip: Test data questions are extremely common on Paper 2 and are typically worth 3–6 marks. You must know the three types of test data and be able to select appropriate examples for any given scenario.

---

The Three Types of Test Data

1. Normal Data

Normal data (also called valid data or typical data) is data that the program should be able to handle — data that falls well within the acceptable range and is in the correct format.

Normal data tests that the program works correctly under typical conditions.

Example: For a program that accepts exam marks from 0 to 100:

• Normal data: 45, 72, 88

---

2. Boundary Data (Extreme Data)

Boundary data tests the values at the edges of the acceptable range — the minimum and maximum values that should be accepted, and values just outside the range that should be rejected.

Boundary testing is crucial because many bugs occur at the boundaries of ranges (off-by-one errors).

Example: For marks from 0 to 100:

Value	Type	Expected Result
-1	Just below lower boundary	Rejected
0	Lower boundary (valid)	Accepted
1	Just above lower boundary	Accepted
99	Just below upper boundary	Accepted
100	Upper boundary (valid)	Accepted
101	Just above upper boundary	Rejected

OCR Exam Tip: When asked for boundary test data, always give values ON the boundary AND values just outside the boundary. For a range of 0–100, give: -1 (rejected), 0 (accepted), 100 (accepted), 101 (rejected). This tests that the boundaries are implemented correctly.

---

3. Erroneous/Invalid Data

Erroneous data (also called invalid data) is data that should be rejected by the program. It is data that is the wrong type, format, or is otherwise unacceptable.

Erroneous data tests that the program's validation works correctly and handles bad input without crashing.

Example: For a program that accepts exam marks (integer, 0–100):

• "abc" — wrong data type (string instead of number)
• "" — empty input (presence check failure)
• -50 — out of range
• 3.5 — wrong data type (real instead of integer)
• "fifty" — wrong data type

---

Summary Table for a Mark Field (0–100, Integer)