- 3.0 Table of Contents
- 3.1 Learning Objectives
- 3.2 Summary
- 3.3 Revisiting Research Methods
- 3.4 Constructs and Clarity
- 3.5 Making a Contribution
- 3.6 The Independent Research Project
- 3.7 Insight into Project Methodologies
- MLO2 - Transform a research topic into a research design
- MLO2 - Consider the design implications of different modes of computer science research
- MLO4 - Critically reflect on time and resource limits in relation to research projects
- MLO4 - Develop a research plan for a researchable problem
A good research project will have interesting research questions or hypotheses, be well designed, communicate clear constructs to the reader, and make a contribution to a field of knowledge.
The Independent Research Project (IRP) requires consideration as to what is possible, practical, and required. Throughout the project, five elements need to be managed: time, scope, quality, resources, and cost.
Depending on the project, traditional, agile, or hybrid project methodologies may be employed.
- Required: Chapters 2, 3, Projects In Computing And Information Systems: A Student's Guide, Dawson
- Required: Chapters 1, 7, Research Design: Qualitative, Quantitative, and Mixed Methods Approaches, Creswell and Creswell
- Extension: none
The advantages and disadvantages of various research approaches were detailed in the Research Methods module:
Research methods can be classified according to their timeframe. Which to choose depends on the nature of the research and the intended outcome.
- Cross-sectional studies provide a snapshot of a situation or event as a particular point in time
- Longitudinal studies provide a long term picture, where data is gathered over a period of time
Research design, questions, and hypotheses are all crucial components of academic research process.
Simply, research design is a blueprint for research: a strategy that guides the researcher in planning and implementing a study in a way most likely to achieve the intended goal.
It outlines:
- What data to gather
- How to collect it
- How to analyse it
In mixed methods research, the design will ensure that both qualitative and quantitative aspects complement each other in understanding the investigated phenomenon.
A research question is a clearly defined question that the study seeks to answer. It could be:
- Open ended
- Exploratory
- Qualitative or quantitative in nature
However, hypotheses are specific predictions about the possible outcomes of the research.
- In quantitative studies, hypotheses are often statements about the expected relationships between variables
- Mixed methods research may have both quantitative and qualitative hypotheses
Research is iterative and rarely neat, however the process can broadly be represented as:
A research project should start with a clearly defined research questions or hypothesis.
- The question should be specific and narrow the focus of the study
- Generally, qualitative research questions start with words like how and what
- Mixed methods research questions should directly address the integration both elements
- Hypotheses should be testable predictions about the expected outcomes
- These are generally articulated as either alternative or null hypotheses
Both questions and hypotheses should be directly linked to the context of the study.
Research symmetry implies that irrespective of what the findings are, they should have a similar value. This means that every hypothesis should be framed in a way that it yields meaningful and valuable insight.
- Required: Toward Principles of Construct Clarity: Exploring the Usefulness of Facet Theory in Guiding Conceptualization, Zhang et al
- Extension: none
With units of analysis and research objects previously discussed, the construct of the research needs to be considered. This relates to:
- The key concepts and topics of the research
- These are not directly observable
- However, they can be inferred from observed and measured data
Construct clarity refers to the practice of carefully identifying strong categories that condense concepts and processes as distinctly as possible.
Clear constructs are important for other researchers: they enable the identification fo problems, potential links to background concepts, and concepts deserving further investigation.
Therefore, for clarity, the research domain and the techniques employed should be stated as early as possible in the proposal and paper. For example, "this research project aims at investigating possible solutions to extract usable work from a physical system, quantifying the performance studying variations of the entropy of the system" identifies the domain in a physics paper clearly.
To aid in achieving construct clarity, Facet Theory can be employed. This is a meta-theoretical framework that provides structure to complex, multidimensional concepts:
- A facet of habitat would result in the frog and duck being grouped together
- However, a zoological classification would result in the duck and hen being together
- Careful consideration of the topic and expected outcomes is therefore important
Facet Theory can be applied to research using the following phases:
- Identify the domain - determine the conceptual universe to be investigated, eg customer satisfaction
- Determine the facets - within this domain, exhaustively identify the facets that can be studied independently, eg product quality, customer service, price
- Identify the elements - determine the elements within each facet, eg product quality may be made up of durability, functionality, aesthetics
- Create a mapping sentence - a comprehensive statement that outlines the structure of the concept according to the facets, elements, and their interactions
- Generate a data collection instrument - create a data collection instrument, such as a survery or questionnaire, that captures data on all facets and elements
- Analyse data - use statistical methods to analyse the data, taking into accountt he structure provided by the facets and elements
- Verify model - verify the theory using empirical data, potentially through comparing the model's predictions with the observed data or coherence with theories
- Required: Chapter 2, Projects In Computing And Information Systems: A Student's Guide, Dawson
- Required: Chapters 3, 5, Writing For Computer Science, Zobel
- Extension: none
Good research involves original investigation, wider contribution, expanding knowledge, and gaining understanding.
Contribution in research is one of the main criteria for judging a paper as it reflects the key properties of originality and validity.
Part of the reason for undertaking a literature review is to:
- Understand the contribution made by the reviewed publications
- Help articulate the contribution the proposed project would make
There are four types of contributions:
- Confirming knowledge - corroborating or verifying existing knowledge, usually through replication with different samples or methods. The contribution is confirmation of the robustness of the existing knowledge
- Enchanging knowledge - building on existing knowledge, possibly by extending theories to new contexts or discovering nuances. The contribution lies in enchancing complexity and depth of understanding.
- Challenging knowledge - findings that contradict or challenge existing theories, models, or assumptions. The contribution is in prompting re-evaluation and innovation in thinking.
- Rejecting knowledge - disproving or rejecting established knowledge, possibly leading to discarding of old theories or models. The contribution is in the updating of understanding.
Each type of contribution is valuable and vital to progress of the field. Together, they represent a dynamic and iterative process of building knowledge.
The significance of a contribution could be measured by its impact or the amount of change that would follow from the paper being published.
- Required: Chapters 3-4, 7, 8, Projects In Computing And Information Systems: A Student's Guide, Dawson
- Chapters 1-6, Dissertations and Project Reports: A Step by Step Guide, Cottrell
- Extension: none
Undertaking a research project involves balancing competing requirements:
- What is possible - conceptually and theoretically possible
- What is practical - real world limitations on time and resources
- What is required - degree requirements, career development etc...
Time is an important resource and effective time management is important throughout long term research projects:
- Personal time - identify issues that may arise, such as a lack of routine or everyday commitments
- Allocating time - consider the broad project processes involved and use planning tools for scheduling
- Pacing work - working little and often, contingency planning, and realistic mini-goal setting can help
- Maintaining momentum - motivation can be maintained through strict routines and acknowledging achievements
Procrastination in writing up projects is common. Therefore, it is important to plan for writing early on. Good enough standards and section deadlines should be decided on.
It is important to understand the role of a supervisor in the project: they provide guidance, direction, and feedback, but not day to day supervision. Establishing good everyday habits is down to the researcher.
The Independent Research Project (IRP) is a module wherein the student will produce an individual piece of research under the guidance of a supervisor.
The Research Proposal (RP) module is 8 weeks and the Independent Resarch Project is 16: a total of 24 weeeks working on the project. This is divided into 5 phases:
- Phase 1 - definition (RP)
- Phase 2 - planning (RP)
- Phase 3 - initiation (IRP)
- Phase 4 - control (IRP)
- Phase 5 - closure (IRP)
Within each phase there may be multiple activities, such as research design, literature reviewing, data collection, writing drafts etc.
In Computer Science there may be an opportuntiy to build a prototype as part of the project. This would require project methodologies to be adopted alongside research methodologies.
The IRP requires the submission of a final research proposal. It is expected that this will be the proposal developed for the assessment in the preceding Research Proposal module, however this can be updated, adapated, or changed.
The timetable for submitting a proposal is constrained by the time it takes to get ethical approval, if needed.
Taking time to create diagramatic schedule can help visualise the intended research steps:
The control phase involves managing and controlling five project elements: time, scope, quality, resources, and cost.
- The IRP must be completed in 16 weeks, so time is fixed
- Similarly, cost is fixed due to being a student without research funding
- This leaves scope as the main adjustable element
Time, scope, and quality should be monitored closely to ensure tasks are completed according to the project plan and objectives.
Project management involves making tradeoffs between elements as the project progresses. These are summarised below, with the exception of cost which is not within the control of the project:
The goal is to ensure the project is completed on time, with the best possible effort.
The closure phase refers to the point where all research activities are completed and the project itself needs to be finalised.
Data collection and analysis needs to be checked to ensure that all data has been collected and analysed according to the research design.
Similarly, the conclusions based on the data analysis needs to be finalised. This may involve going through several drafts before considering this section completed.
The final report needs to be written, conforming to the relevant standards. This should include the research question or hypothesis, methodology, data analysis, results, conclusion, and recommendations.
Every project has risks. The research proposal will need a risk log included for the supervisor to review.
Risks are not stationary and so this log may need updating as the project progresses, as not all risks can be anticipated and some may escalate in importance.
The risk management process can be illustrated as four stages:
- Identify risks - identifying potential risks that could affect the project
- Assess impact of risks - assessing the impact of each risk and determining which are most critical
- Alleviate critical risks - creating strategies and contingency plans to deal with each critical risk
- Control risks - managing and monitoring risks, possibly checking for risk triggers or precursors
- Required: Chapters 6, Projects In Computing And Information Systems: A Student's Guide, Dawson
- Extension: none
Project methodologies are primarily covered during the IRP, but they can provide helpful insights as to how to plan a project.
Generally, projects run in a sequential order. However, both traditional and agile methodologies are suitable in Computer Science for projects that involve software development, possibly even a hybrid of the two.
Conventional waterfall type models follow a sequential, one directional flow of stages. These are:
- Less flexible
- Don't typically respond well to change
- Suitable for projects with well understood and unchanging requirements
Incremental development involves breaking projects into smaller pieces with smaller aims.
- This results in releasing an operational product each increment
- Additionally, feedback and can be incorporated into the process
Prototyping is an iterative refinement of a product, circling a final version.
- It is appropriate when there is uncertainty in the requirements
- Additionally, it can help avoid spaghetti code by ensuring logcial structures at the start
Agile methods emphasis quick, iterative development cycles.
- Values continuous delivery, flexibility, and regular communication
- This is ideal for projects with unclear or changing requirements
Ultimately, choice depends on the nature of the project, risks involved, time frame, and resources available. Generally:
- Waterfall approaches are suitable for projects with unchanging requirements
- Incremental approaches are good for projects with sub-deliverables
- Prototyping approaches are useful for proof of concept situations