Personal before business in requirements prior-IT-ization Johan F. Hoorn Vrije Universiteit Computer Science Information Management and Software Engineering
Contents Status Problem Analysis Model Predictions Method Results Conclusions Discussion Johan F. Hoorn, 2005 M M I
Status Sept. 1, 2001 – Aug. 31, 2005 Supervisors: Gerrit van der Veer Hans van Vliet Four international publications Industries: Human-Machine Interaction Johan F. Hoorn, 2005
Problem Requirements change Business model 1 Traditional office Mainframe with thin clients Business model 2 Flexible workplace Laptops with bluetooth A change request during development is extremely costly and frustrating Johan F. Hoorn, 2005
Nobody knows how changes in requirements priorities can be predicted
Analysis Where do change requests come from? Business model 1 Business model 2 Change in business sub goals - Main goals:Profit - Sub goals:Cost-effectiveness, efficiency How come business goals change? Change in personal sub goals (strategic management) - Main goals:Earn my living - Sub goals:Fire employees (not me), improve IT to guarantee same output Johan F. Hoorn, 2005
Model Change of Stakeholder Requirements (CoStaR) (Hoorn & Van der Veer, 2003a; 2003b) One of the hypotheses: Johan F. Hoorn, 2005 Goals RelevanceRequirements Stakeholder evaluation: Is a system feature important to my goals? Is a system feature trivial to my goals? (after Frijda, 1986)
Predictions If business or personal goals change, requirements prioritization – as an expression of relevance assessment – will change accordingly BM1: Traditional officeBM2: Flexible workplace 1 Mainframe1 Laptops 2 Thin clients 2 Bluetooth 3 Laptops3 Mainframe 4 Bluetooth4 Thin clients Priority change
Method (1) Requirements rank-ordering test System: Blackboard (BM1) vs. Didactor (BM2) Internet survey (the e-learning hoax) Stakeholders: Science students (N= 154) Four conditions of goal change: 1 From business egotistic to business altruistic (n= 36) 2 From business altruistic to business egotistic (n= 39) 3 From personal egotistic to personal altruistic (n= 43) 4 From personal altruistic to personal egotistic (n= 36) Johan F. Hoorn, 2005
Condition 1: From business egotistic… Students put rank numbers 1 up to 16 Different randomization between and within students
…to business altruistic For business altruistic to egotistic, the Motivation order was reversed Students, again, put rank numbers 1 up to 16
Uit eerder onderzoek onder studenten van de Vrije Universiteit is gebleken dat zij zelf meer willen kunnen profiteren van de kenniseconomie. Daartoe, vinden zij, moeten studenten hooggekwalificeerd de markt op kunnen. Om straks een goed betaalde baan te krijgen moeten overheidssubsidies ge- investeerd worden in hoogstaande technologie en leermiddelen die de individuele student ondersteunen. Uit de alternatieven kozen de studenten de nieuwe digitale leeromgeving The Didactor® als de meest geschikte kandidaat. Likewise for personal egotistic…
Uit eerder onderzoek onder studenten van de Vrije Universiteit is ook gebleken dat zij zich maatschappelijk verantwoordelijk voelen voor de ontwikkeling van de kenniseconomie. De studenten stellen dat de investeringen in hooggekwalificeerde afgestudeerden de maatschappij ook weer ten goede moet komen. Daartoe zijn hoogstaande technologie en leermiddelen noodzakelijk waarmee studenten elkaar kunnen ondersteunen. Uit de alternatieven kozen de studenten de nieuwe digitale leeromgeving The Didactor® als de meest geschikte kandidaat. …to personal altruistic and v.v.
Method (2) Calculating priority change Spearman’s rho ( S ) is a rank order correlation coefficient that analyzes whether a bivariate set of paired rankings correlates by rank sum S was calculated for each student in a condition S (altru to ego) S (ego to altru) (Business)(Personal) S (ego to altru) Johan F. Hoorn, ΣD 2 S = 1 – N(N 2 – 1)
Method (3) Four measures of priority change S 1 over data of those who filled in both lists (N= 103) S 2 over data of the 10 features that best contributed to S = -1 (N= 92) S 3 feature to feature rank-order total-scores,* using data of all those who filled in the first list (N= 154) S 4 feature to feature rank-order total-scores* over data of the 10 features that best contributed to S = -1 (N= 154) *(see paper or last slides) Rho moves between 1 and -1. The closer S approaches -1, the higher the disagreement between the two sets of ranked features (= priority change) Johan F. Hoorn, 2005
Original hypothesis: Johan F. Hoorn, 2005 Goals RelevanceRequirements Only S 3 (feature to feature rank-order total-scores),* using data of all students who filled in the first list (N= 154), rendered significant results *(see paper or last slides) Results (1)
Main effect (Business vs. Personal): F (1,146) = 4.09, p<.05, η p 2 =.03 The only significant difference Mean S 3 =.60 (priority change is low) Mean S 3 =.48 (priority change is high) Johan F. Hoorn, 2005 Results (2)
RE should be oriented to personal goals Only changes in personal goals had an impact on changes in requirements prioritization This effect occurred irrespective of the type of goal change (from egotistic to altruistic or v.v.) Business model change had less impact on changes in requirements prioritization Johan F. Hoorn, 2005 Conclusions
Effects were not too strong (η p 2 =.03). Replication in a business case is urgent Johan F. Hoorn, 2005 Discussion
Appendix (1) Calculating S 3 (feature to feature rank-order total-scores) Only the data of the first requirements list were used (N= 154) For each feature, the sum of rank-order scores was computed across all students in a condition (e.g., Business egotistic (Be) or Personal altruistic (Pa)) On the basis of the rank-order total score per feature (which were between 91 and 576), the 16 features were then rank-ordered from the lowest to the highest rank-order total score Johan F. Hoorn, 2005
Subsequently, the actual rank-order total score of a feature was replaced by the rank order number of their relative position in this general priority list. The feature with the lowest rank-order total score (= 91) received a 1 and the feature with the highest rank-order total score (= 576) received a 16 The feature to feature rank-order total-scores were established by calculating, for each student in a condition, s between Be (as based on the raw data) and the revised Ba (as based on the rank-order total scores) Ba (as based on the raw data) and the revised Be (as based on the rank-order total scores) Pe (as based on the raw data) and the revised Pa (as based on the rank-order total scores) Pa (as based on the raw data) and the revised Pe (as based on the rank-order total scores) Johan F. Hoorn, 2005 Appendix (2)