Master Systematic Literature Reviews (SLR), PRISMA framework diagrams, IMRaD limitations, and Harvard citation rules (<5 years recency constraint).
Part 1: Systematic Literature Review & PRISMA Framework
M4-Q01[DRAFT & OUTLINE]
Explain the FOUR key stages of the PRISMA Framework (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) used in CS research. Draw a textual flow diagram of a PRISMA review starting from 120 database records down to 15 included studies.
NUST Model Answer [8 Marks]:
The 4 PRISMA Stages:
1. Identification: Searching digital databases (IEEE Xplore, ACM Digital Library, Springer, Google Scholar) using predefined search queries and boolean operators. (n = 120 initial records).
2. Screening: Removing duplicate records and screening titles/abstracts against inclusion/exclusion criteria. (n = 85 records screened after deduplication).
3. Eligibility: Assessing full-text papers for detailed quality and methodological relevance. (n = 40 full-text articles assessed).
4. Included: Final set of high-quality studies selected for synthesis in the literature review chapter. (n = 15 studies included).
What is the IMRaD Format (Introduction, Methods, Results, and Discussion) and what are its 3 primary limitations when applied to Computer Science research reports? [6 Marks]
NUST Model Answer [6 Marks]:
IMRaD Definition: A standard structure for empirical research papers comprising Introduction, Methods, Results, and Discussion.
3 Limitations in Computer Science Research:
Rigidity for System Artifacts & Engineering Architecture: IMRaD leaves no dedicated space for System Analysis & Design (Context Diagrams, ERD, Use Cases, System Architecture) which are essential intermediate artifacts in CS software/hardware research.
Inability to Accommodate Iterative Development: CS software methodologies (e.g. Iterative Prototyping, Agile) involve cycles of design-build-test, which clash with IMRaD's linear assumption that Methods precedes Results strictly once.
Obscures Algorithmic Pseudo-code & Mathematics: IMRaD forces complex algorithm derivations into a generic "Methods" section, making mathematical model formulation difficult to present cleanly.
M4-Q03[CRITIQUE & AUDIT]
HARVARD CITATION AUDIT: Review the following 3 reference entries from a student's proposal draft. Identify all citation errors based on NUST guidelines (< 5 years recency, Harvard format).
1. Smith, J. (2014) Network Security Essentials. New York: Wiley. 2. https://github.com/nfc-auth/project - an awesome repo for biometrics. 3. Moyo, T. and Ncube, S., 2024. Academic impersonation in higher education. African Journal of Higher Education, 8(3), pp.88-102.
NUST Model Answer [6 Marks]:
Audit & Corrections:
Entry 1: REJECTED FOR OUTDATED RECENCY. Published in 2014 (> 10 years old). NUST rules explicitly require references published within the last 5 years (2021–2026 for current work).
Entry 2: INVALID CITATION FORMAT. Bare URLs without author, year, title, or access date are illegal in Harvard style. Correct format: NFC Auth Team (2024). Biometric NFC Driver Repository. Available at: <URL> [Accessed 15 July 2026].
Entry 3: VALID HARVARD CITATION. Complies with < 5 years recency (2024), includes Authors (Moyo & Ncube), Year, Article Title, Journal Name (italicized), Volume(Issue), and Page numbers.