Contributed on July 8, 2013

5.1K

1.9K

COGNISE@CAiSE2013

Description

Slides of my presentation at COGNISE workshop at CAiSE conference, 17 June 2013, Valencia, Spain

Transcript

Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 201318 June, 2013FACULTY OF ECONOMICS AND BUSINESS ADMINISTRATIONCognitive Aspects ofStructured Process ModelingJan Claes, Frederik Gailly and Geert Poels Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 20132/14OutlineProcess models, process model qualityProcess of Process Modeling, data collectionVisualization with PPMChartsObservations, Structured Process ModelingTheoretical explanation, hypothesesMethodology Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 20133/14Business Process Modeling (BPM)Business process model Graphical, abstract representation of process Important tool for analysis and improvementBusiness process model in BPMN notation Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 20134/14Process model qualityQuality dimension PurposePRIMARYRESEARCHFOCUS Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 20135/14Process of Process Modeling (PPM)Properties oftextual descriptionProperties ofmodelerProperties ofmodeling processProperties ofresulting modelPRIMARY RESEARCH FOCUSProperties ofreal processProperties ofobservation processProperties of softwareand modeling languageProperties ofmodel readerProperties ofreading process Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 20136/14Process of Process Modeling (PPM)Observational modeling sessions People construct models Every action on modeling canvas is logged Modeler information is collected through survey Different datasets• 120 students in Eindhoven 2010• 14 experts in Berlin 2010• 14 experts in Eindhoven 2011• 118 students in Eindhoven 2012 Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 20137/14Process of Process Modeling (PPM)Properties of the modeling processActivity Timestamp AttributesCreate start event 10:00 Id = 1; x = 10; y = 10Create activity 10:04 Id = 2; x = 40; y = 10; name = “Receive order”Create edge 10:05 Id = 3; from = 1, to = 2Move activity 10:07 Id = 2; x = 15; y = 10Create gateway 10:08 Id = 4; x = 65; y = 10; type = “XOR”Create edge 10:09 Id = 5; from = 2, to = 4Create activity 10:24 Id = 6; x = 80; y = 0; name = “Reject order”Create activity 10:25 Id = 7; x = 80; y = 20; name = “Prepare order”Create gateway 10:27 Id = 8; x = 105; y = 10; type = “XOR” Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 20138/14 CREATE_ACTIVITY CREATE_START_EVENT CREATE_END_EVENT CREATE_AND CREATE_XOR CREATE_EDGE MOVE_ACTIVITY MOVE_START_EVENT MOVE_END_EVENT MOVE_AND MOVE_XOR DELETE_ACTIVITYDELETE_START_EVENT DELETE-END_EVENT DELETE_AND DELETE_XOR DELETE_EDGE NAME_ACTIVITY RENAME_ACTIVITY NAME_EDGE RENAME_EDGEProcess of Process Modeling (PPM)Visualization in PPMCharttimemodelelements Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 20139/14ObservationsChunked modelingStructured process modelingModeling pausesDelayed startPRIMARY RESEARCH FOCUS Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 201310/14Structured Process Modeling (SPM)Structured process modeling Creating blocks ‘as a whole’ (before moving on tothe creation of the rest of the model)ACTACTX XACTProcess model block Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 201311/14Theoretical explanationCognitive aspects Cognitive Load Theory (CLT)limited capacity of working memory Cognitive Fit Theory (CFT)effect increase if task representation/modeler fits Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 201312/14HypothesesH1: Structured process modeling relates toprocess model quality improvementH2: The quality improvement will be higher fornovices than for experts(expertise in the case domain, in modeling, and/or in themodeling language)H3: The quality improvement will be higher ifthe task representation fits with the technique(i.e. a “breadth-first” ordering of the task description) Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 201313/14MethodologyControlled experiment Pre-test Treatment Experiment Compare correctnessand understandabilityCognitive Fit Theory Repeat with experts 3 different taskdescription ordersPre-test: case 1ST1 & QT1TreatmentExperiment: case 2 Experiment: case 2SC1 & QC1ST2 & QT2 SC2 & QC2Fake treatmentTreatment group Control group(QT2 – QT1) > (QC2 – QC1)ST2 > ST1 & QT2 > QT1SC2 ≈ SC1 & QC2 ≈ QC1Pre-test: case 1Tool tutorial Tool tutorial Ghent university, Economics and Business Administrationjan.claes@ugent.be – www.janclaes.infoCOGNISE@CAiSE 201314/14Contact informationJan Claesjan.claes@ugent.behttp://www.janclaes.infoTwitter: @janclaesbelgiumThanks for your attention!Do you know a better cognitive explanation forthe observed effect of structured processmodeling on process model quality?