NCEA 91908: analyse an area of computer science (Level 3 Digital Technologies)

What NCEA 91908 assesses

NZQA's standard (version 1) says you analyse an area of computer science by explaining:

• the key aspects of the area
• the relevant algorithms or other mechanisms behind it
• how the area is used, implemented or occurs, giving examples
• key problems or issues related to the area and how they have been, or may be, addressed

Achieved is that explanation. Merit analyses in depth: a detailed explanation of how the technical capabilities and limitations of the area relate to humans, giving examples, and comparing and contrasting different perspectives. Excellence analyses critically by drawing insightful conclusions about the area.

The areas in 2026 and 2027

You answer on one area, chosen from the list in NZQA's specification for that year. The list changes:

For complexity and tractability, NZQA lists time complexity and Big O notation, best, worst and average case, the classes P, NP and NP-complete, approximation algorithms and heuristics, and problems such as the travelling salesperson and knapsack. For formal languages it lists regular expressions, finite-state automata and context-free grammars.

The NCEA 91908 exam in 2026 and 2027

NZQA's 2026 timetable puts Level 3 Digital Technologies on Tuesday 10 November, in the afternoon session. The 2027 timetable is not out yet: NZQA says exams run from 9 November to 3 December 2027 and the subject timetable is due in December 2026. The exam is online, so you need a laptop or desktop computer. NZQA's 2025 report says it had two parts: short answers on key mechanisms, algorithms, applications and issues, then deeper analysis and evaluation including limitations, societal impacts and human and computer interaction. Aim for 1,500 words at most in total, and be ready to apply your area to unfamiliar contexts and to overlaps with areas such as artificial intelligence.

How NCEA 91908 is marked

NZQA's schedule scores your whole response from N1 to E8. N1 and N2 are relevant comments that do not show understanding of the whole response, A3 and A4 explain the key aspects, algorithms or mechanisms and how the area is used, M5 and M6 add the detailed link to humans and the comparison of perspectives, and E7 and E8 draw insightful conclusions.

NCEA 91908 cut scores

These are NZQA's cut scores for the total out of 8, taken from each year's assessment schedule. They have not changed from 2021 to 2025. The 2026 cut scores will appear in NZQA's 2026 schedule once the exam has been marked.

Worked examples at each grade

These examples use complexity and tractability. Searching a sorted list of 1 000 000 numbers shows the idea:

def binary_search(items, target):
    low, high = 0, len(items) - 1
    while low <= high:
        mid = (low + high) // 2
        if items[mid] == target:
            return mid
        if items[mid] < target:
            low = mid + 1
        else:
            high = mid - 1
    return -1

Achieved: explain the key aspects

Linear search checks each item in turn, so in the worst case it makes 1 000 000 comparisons: its time complexity is O(n). Binary search halves the range after each comparison on a sorted list, so the worst case is about log₂(1 000 000) ≈ 20 comparisons: O(log n). Both give the right answer, and an everyday example is looking up a name in a sorted contacts list.

Merit: link capability and limitation to people

Binary search needs sorted data, so it only pays off if the list is already sorted or is searched many times, since sorting costs O(n log n). A phone app searching 1 000 000 contacts feels instant with binary search, but a user with a few hundred contacts will not notice the difference. Compare perspectives: a developer values the speed, while a small business may prefer the simpler code.

Excellence: an insightful conclusion

Brute force for the travelling salesperson problem checks every route: (n − 1)! ÷ 2 distinct routes, which is 181 440 for 10 cities and about 6.1 × 10¹⁶ for 20. The growth is O(n!), so the problem is intractable for large inputs, and no known algorithm solves it in polynomial time. A nearest-neighbour heuristic runs in O(n²) and gives a good but not necessarily optimal route. The insight is that the practical question is not "which algorithm is correct" but "what is a good enough answer in the time and energy we have", which is why delivery companies use heuristics. Notice how complexity decides what is feasible, not just what is fast.

Try counting halvings in our Search Algorithms course, then press Mark to check your answer in the course:

Exam answers

The answers and a worked mark scheme come with today's paper once it is marked. We do not publish official or past paper answers. Every question on the daily paper is our own.

Use the daily paper as a practice exam

Uni-Sine publishes a free daily paper on the NCEA Level 3 Digital Technologies track. It is a short timed paper of 8 questions worth 25 to 35 marks, not a full NCEA exam, and it is not written by NZQA. Its topics include searching, sorting and complexity, regular expressions, finite state machines and networking, which line up with the complexity, formal languages and protocols areas. It asks short questions, while 91908 asks for 800 to 1,500 words, so also practise writing a one-page analysis of your chosen area. See the NCEA Level 3 Digital Technologies guide for the whole subject.

Common mistakes in NCEA 91908

NZQA's 2025 report says candidates who were Not Achieved commonly:

• Gave vague, general answers without showing computer science understanding.
• Missed the mechanism. They failed to identify or explain the required algorithms.
• Wrote at length instead of with accuracy. The report says the assessment rewarded accuracy, specificity and clarity rather than volume.
• Stopped at description. Excellence needed technical reasoning, calculations or comparisons, and synthesis beyond isolated explanations.

Read next

New Zealand · NCEA Level 3NCEA Level 3 Digital Technologies: standards and examNCEA Level 3 Digital Technologies explained: the computer science and reflective analysis externals, the 2026 and 2027 topics, exam dates and examples.