Research Methods for Computer

Science
One of the primary goals of acad
emic training is to learn how to le
arn, i.e., to learn how to continuo
usly absorb new knowledge.
 The process of exploring the unk
nown, studying and learning new
things, building new knowledge a
bout things that no one has unde
rstood before - that is what we th
ink of as performing research.
The activity of a diligent and systema
tic inquiry or investigation in an area,
with the objective of discovering or r
evising facts, theories, applications e
tc. The goal is to discover and disse
minate new knowledge
 Research Process
 Realize there is a problem
 Look to see if anyone has solved it
 Develop a plan to solve it
 Solve it
 Evaluate your solution
 Disseminate your solution
 CS Research
 Inherently multidisciplinary (except for CS Ed)
 Other fields are not multidisciplinary
 Planned
 Specific Purpose
 Clear start and finish dates
 Finite resources (time, people, and money)
 Continuing Educational Goals
 Refine your critical thinking skills
 Refine your ability to work independentl
y or on a team
 Increase your understanding of how to
use research tools
 Refine your oral and written presentatio
n skills
 Continuing Research Goals
 Deepen understanding of a topic
 Contribute to the community’s underst
anding of a topic
 Key Players
 The researcher
 The supervisor
 The evaluator
 Assessment
 Relevance
 Originality
 Significance
 Not plagiarized
 Clearly presented
 Consistent presentation
 Apparent understanding of topic
 Ability to handle references and citations
 Ability to meet deadlines
 Ability to adhere to formal requirements
 Research Institute
 Establish new knowledge and make it a
vailable to the public
 Not driven by profit
 Researchers free to choose research top
ic
 Commercial Organization
 Research related to business goals
 Goal is to generate profit for the business
 Researchers must choose topics that will deve
lop new product for the organization
 Frequently the details are not disseminated to
the public
Regardless of the type of rese
arch, it is critical that it be reli
able
 Research Methods
 Procedures for collecting data, formulati
ng a hypothesis, testing a hypothesis, in
terpreting results, and drawing conclusi
ons
 Quantitative Research
 Develop models, theories, and hypotheses de
scribing a phenomenon
 Quantitative methods involve measuring som
ething that will help develop the model/theor
y/hypothesis
 Large samples needed
 Experiments are repeatable
 Qualitative Research
 Discover underlying meanings and patterns of
relationships without using or developing mat
hematical models.
 Qualitative methods involve fieldwork which c
ould include interviews or personal observatio
ns
 Usually small sample sizes
 Experiments not necessarily repeatable
Qualitative research is frequen
tly a precursor to quantitative
research.
 The Research Team
 Researcher
 Supervisor(s)
 Evaluator(s)
 The Researcher
 Discuss with your supervisor what supervi
sion and work schedule is good for you a
nd them.
 Discuss the topic and timetable with the s
upervisor
 Stick with the schedule and don’t disappe
ar from the supervisor’s radar
 Keep systematic records of work
 Submit written material with enough time
for the supervisor to read it.
 The Researcher
 Discuss the final submission details with t
he supervisor
 Don’t ignore criticisms or guidance from t
he supervisor
 Make sure you don’t do anything illegal
 Remember you are the driver
 Let the supervisor know of any problems
 Do your best.
 The Supervisor

 Know the rules and standards of the orga

nization regarding research
 Make sure the supervisees know the rules
and standards
 Discuss dates and work schedules
 Give needed guidance
 Continuously update skill set
 Schedule regular meetings
The Quality Evaluator Examine
r
 Evaluates the completed project.
 Evaluates based on the contribution, complexi
ty of the problem, the usefulness of the soluti
on, and the presentation style.
 Very objective … well, it’s supposed to be :-)
The Quality Assuror Examiner
 Reviews and gives feedback at various c
heckpoints in the research.
 Can eliminate big problems from occurri
ng at the end of the project
 Can be too subjective, but not nearly as
subjective as the supervisor.
It is usually easier to be either
a Quality Assuror or a Quality
Evaluator, but not both. So ch
oose one of each if you can.
If you get the chance to be an
evaluator it can improve your
own research
Typical Evaluation Criteria

 Does the title correspond well to the cont

ents?
 Does the abstract give a complete and co
rrect picture of the contents?
 Can the abstract be understood without r
eading the paper?
 Does the introduction explain clearly what
the investigated problem is?
 Does the introduction give a complete an
d brief overview of the whole project?
Evaluation Criteria Continued

 Does the author explain all terminology in

a clear and precise way?
 Are terms used in a consistent way?
 Does the author make use of relevant sci
entific conventions in the choice of termin
ology and definitions?
Evaluation Criteria Continued

 Is the description of each method sufficie

ntly clear and detailed to allow replication?
 Are there any technical flaws where a met
hod is applied in an incorrect way?
Evaluation Criteria Continued

 Are the presented results relevant?

 Is the analysis done correctly?
 Are the analysis techniques relevant?
 Are tables and figures used in a correct and relev
ant way?
 Are all results presented objectively?
 Does the text match the tables/figures?