Research Methods In Computer Science


Research classification Research methodsResearch Methods in Computer ScienceLecture 6: Research methodsUllrich HustadtDepartment of Computer ScienceUniversity of LiverpoolUllrich HustadtResearch Methods in Computer Science102 / 117

Research classification Research methodsPreviously . . .13Scientific methodElements14Intellectual discoveryDeductionAbductionInductionProcess model15Problem solvingUllrich HustadtResearch Methods in Computer Science103 / 117

Research classification Research methodsTopics16Classifying research17Research methodsOverviewExperimentsQuestionnairesUllrich HustadtResearch Methods in Computer Science104 / 117

Research classification Research methodsClassifying research (1)Research can be classified from three different perspectives:1FieldPosition of the research within a hierarchy of topicsExample:Artificial Intelligence Automated Reasoning First-Order Reasoning Decidability2ApproachResearch methods that are employed as part of the research processExamples:Case study, Experiment, Survey, Proof3NaturePure theoretical developmentReview of pure theory and evaluation of its applicabilityApplied researchUllrich HustadtResearch Methods in Computer Science105 / 117

Research classification Research methodsClassifying research (2)Pure theory:Developing theories and working on their consequences, with regard toexperimentation or applicationDescriptive studies:Reviewing and evaluating existing theories, including describing thestate of the art, comparing predictions with experimental dataExploratory studies:Investigating an ‘entirely’ new area of research, exploring a situation ora problemSee tory studies:Explaining or clarifying some phenomena or identifying the relationshipbetween thingsUllrich HustadtResearch Methods in Computer Science106 / 117

Research classification Research methodsClassifying research (2)Causal studies:Assessing the causal relationship between thingsNormative studies:Producing a theory of design (or of other development) likerecommendations, rules, standards, algorithms, advices or other tools forimproving the object of studyProblem-solving studies:Resolving a problem with a novel solution and/or improving somethingin one way or anotherDevelopment and Application studies:Developing or constructing something novelUllrich HustadtResearch Methods in Computer Science107 / 117

Research classification Research methodsOverview Experiments QuestionnairesQuantitative and qualitative research methodsQuantitative research methodsMethods associated with measurements (on numeric scales)Stemming from natural sciencesUsed to test hypotheses or create a set of observations for inductivereasoningAccuracy and repeatability of vital importanceQualitative research methodsMethods involving case studies and surveysStemming from social sciencesConcerned with increasing understanding of an are, rather than anexplanationRepeatability usually a problemUllrich HustadtResearch Methods in Computer Science108 / 117

Research classification Research methodsOverview Experiments QuestionnairesResearch methods (1)Action research:Pursues action (or change) and understanding at the same timeContinuously alternates between action and critical reflection, while refiningmethods, data and interpretation in the light of the understanding developedin the earlier cyclesExample: Reflective teachingCase study:In-depth exploration of a single situationUsually generates a large amount of (subjective) dataShould not merely report the data obtained or behaviour observed butattempt to generalise from the specific details of the situation observedExample: Case study of open source software developmentUllrich HustadtResearch Methods in Computer Science109 / 117

Research classification Research methodsOverview Experiments QuestionnairesResearch methods (2)Survey:Usually undertaken using questionnaires or interviewsQuestionnaire and interview design important!(See Dawson 2005 for details)Determination of sample size and sample elements important!(See specialist literature for details)Example: Survey on the popularity or use of programming languagesExperiment:Investigation of causal relationships using test controlled by the researcherUsually performed in development, evaluation and problem solving projectsExample: Evaluation of processor performanceUllrich HustadtResearch Methods in Computer Science110 / 117

Research classification Research methodsOverview Experiments QuestionnairesKey elements of an experimentA precise hypothesis that the experiment will confirm or refuteA completely specified experimental system, which will be modified insome systematic way to elicit the effects predicted by the hypothesisQuantitative measurement of the results of modifying the experimentalsystemUse of controls to ensure that the experiment really tests the hypothesisAnalysis of the measured data to determine whether they are consistentwith the hypothesisReport of procedures and results so that others can replicate theexperimentUllrich HustadtResearch Methods in Computer Science111 / 117

Research classification Research methodsOverview Experiments QuestionnairesKey issues for questionnairesConsider the following questionsWhat are the key issues for conducting a survey by questionnaire?Regarding the questionnaire itself, what types of questions do you knowand what is each of them used for?(7 minutes group discussion)Ullrich HustadtResearch Methods in Computer Science112 / 117

Research classification Research methodsOverview Experiments QuestionnairesKey issues for questionnairesDetermining the target audienceDetermining the most appropriate mediumAchieving an acceptable response rateEnsuring anonymity if necessaryObtaining additional information about the respondentsQuestionnaire designLayout and size (not too long, uncluttered)Question types(1) Quantity or information(5) RankingHow many hours . . .Rank in order of importance(2) Classification(6) Complex grid or tableGenderMultiple classifications(3) List or multiple choice(7) Open-endedHow do you keep informed?What do you think about . . .(4) ScaleHow easy is . . .Ullrich HustadtResearch Methods in Computer Science113 / 117

Research Methods in Computer Science Lecture 6: Research methods Ullrich Hustadt Department of Computer Science University of Liverpool Ullrich Hustadt Research Methods in Computer Science 102 /