Directly targets **DECEMBER 2024 EXAM QUESTION FOUR [20 MARKS]** and **QUESTION FIVE [20 MARKS]**. Master the distinction between research methodologies and software engineering frameworks.
Part 1: Research Methodology vs. Software Development Methodology
M3-Q01[DEC 2024 EXAM Q4] [DEFEND & COMPARE]
Use a suitable example to distinguish between a Research Methodology and a Software Development Methodology. [20 Marks]
NUST Model Answer [20 Marks Structure]:
1. Definitions & Primary Purpose (6 Marks):
Research Methodology: The overarching strategy and logical sequence governing data collection, empirical testing, hypothesis evaluation, and validation of new knowledge (e.g., Applied Quantitative Research or Experimental Design).
Software Development Methodology: The engineering framework governing how software code, artifacts, modules, and systems are specified, built, tested, and maintained (e.g., Iterative Prototyping, Agile / Scrum, or Waterfall).
2. Direct Comparison Matrix (8 Marks):
Core Objective: Research Methodology answers a scientific question or validates a hypothesis; Software Dev Methodology delivers a functional software product.
Evaluation Metric: Research Methodology measures empirical validity, statistical significance, precision/recall, or latency reduction; Software Dev Methodology measures sprint velocity, code coverage, defect count, and feature completion.
Sequence & Control: Research Methodology isolates independent and dependent variables under controlled conditions; Software Dev Methodology manages backlog items, sprints, and code refactoring.
3. Illustrative CS Example (Biometric Exam Authentication System) (6 Marks):
Research Methodology Chosen: Applied Experimental Research. Why: To empirically measure whether dual-factor NFC + fingerprint authentication reduces impersonation fraud rates while maintaining throughput under 3.5 seconds per student.
Software Development Methodology Chosen: Iterative Prototyping. Why: To build the physical hardware driver integration in small iterative feedback cycles (NFC tap module → Fingerprint SDK integration → SQLite local sync).
⚠️ Examiner Warning: Stating "Agile is my research methodology" is an immediate failing error. Agile is a software development methodology, NOT a research method.
Part 2: Mixed Research Methods Defense (Dec 2024 Q5)
M3-Q02[DEC 2024 EXAM Q5] [DEFEND & COMPARE]
"The mixed research methods have gained more popularity than both the qualitative research methods and the quantitative research methods because of their structure." DEFEND. [20 Marks]
NUST Model Answer [20 Marks Defense Structure]:
1. Thesis & Core Definition (4 Marks):
Mixed methods research explicitly integrates both quantitative (numerical/statistical) and qualitative (contextual/textual) data collection and analysis within a single study. It provides a more comprehensive understanding of complex computer science problems than either approach alone.
2. Architectural Superiority Over Monomethods (8 Marks):
Overcoming Quantitative Monomethod Limitations: Quantitative benchmarks (e.g., measuring app latency spikes) reveal what happened, but fail to explain why users abandon the system or how they experience latency psychologically.
Overcoming Qualitative Monomethod Limitations: Qualitative interviews provide rich context regarding user frustration, but lack generalizability and statistical rigor to prove systemic performance trends across thousands of users.
Triangulation & Convergence: Mixed methods enable researchers to cross-verify findings from multiple sources, enhancing internal and external validity.
3. Defense via the Three Structural Mixed Method Designs (8 Marks):
Explanatory Sequential (Quantitative → Qualitative): Collects numerical benchmark data first (e.g., telemetry logs showing 28% drop-off on biometric entry), followed by qualitative interviews to explain why the biometric scanner failed for specific skin tones.
Exploratory Sequential (Qualitative → Quantitative): Conducts interviews with system invigilators first to discover unseen exam security vulnerabilities, then designs a quantitative survey across 1,000 students to measure how widespread those vulnerabilities are.
Convergent Parallel (Quantitative + Qualitative Simultaneously): Concurrently collects sensor throughput metrics (latency in ms) and user anxiety ratings during live exams, merging both datasets during interpretation to evaluate holistic system performance.