PSD model

Teppo Räisänen | University of Oulu
Practical Findings from Applying
the PSD Model for Evaluating
Software Design Specifications
Teppo Räisänen
Tuomas Lehto
Harri Oinas-Kukkonen
University of Oulu, Finland
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Outline





Background
Requirements engineering
Case
Results
ELM and PSD
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Outline





Background
Requirements engineering
Case
Results
ELM and PSD
THIS IS AN ACTUAL
PRESENTATION I GAVE IN
PERSUASIVE 2010 CONFERENCE
IN COPENHAGEN, DENMARK
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Background
 We have applied the PSD model in a similar
fashion for a half a dozen traditional persuasive
cases
 Alcohol, weight-loss, exercise behaviors
 Research collaboration with an Oulu-based
technology company who have been developing
a reference platform for mobile Internet device
 = a mobile device that can be used to design mobile
devices
 Cf. Google’s Android or Nokia’s Symbian
 In this paper
 Analysing the software requirements in the design
specification documents
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Background
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Background
 Any changes in
the SW are more
expensive
the later in the
process they are
introduced
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Background
In addition to this,
behavior and
attitude change should
be taken into account
early in the design
process
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Requirements engineering
 Early part of the software system development
process
 ”Branch of software engineering concerned with
the real-world goals for, functions of, and
constraints on software systems. It is also
concerned with the relationship of these factors
to precise specifications of software behavior,
and to their evolution over time and across
software families.” [1]
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Requirements engineering





Eliciting requirements
Modeling and analyzing requirements
Communicating requirements
Agreeing requirements, and
Evolving requirements
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Requirements engineering
 Communicating requirements
 Communication with stakeholders
 Agreeing requirements
 Establishing a consensus that the requirements and
models elicited provide an accurate account of what
the stakeholders want
 Evolving requirements
 Handling changes in the requirement documents
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Requirements engineering
 Eliciting requirements
 Identify all the stakeholders (e.g. customers,
developers, users)
 Identify the objectives and tasks of the system
 Prototyping, brainstorming, focus groups
 Modeling and analyzing requirements
 Produce abstract descriptions of the requirements so
that they would be amenable to interpretation.
 Data modeling, domain modeling, behavior modeling
 If we design persuasive systems domain modeling
and behavior modeling are crucial
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Requirements engineering
 Domain modeling helps understand the
persuasion context
 PSD: The intent, the event, and the strategy
 Behavior modeling helps identify opportunities
for different behavior change types
 Potentially even Kairos
 Microsuasion vs macrosuasion
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Case
 We utilized the PSD Model to evaluate
requirements specifications of a mobile Internet
device
 We received many requirements documents
related mainly to user interface and user
interaction issues
 They were full of live-sized pictures,
descriptions of the touchpad gestures, and user
interface elements
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Case
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Case
 Two researchers independently evaluated the
documents through the PSD framework
 A joint understanding was formed through faceto-face discussion
 All persuasive principles from the PSD Model
were considered and the SW was graded in
terms how well the requirements seemed to
support them
 The company was then provided with the results
and a list of suggestions on how to improve the
persuasiveness and usability of the system
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Results
 The case demonstrated that the PSD Model
works well for evaluating SW requirement
 It was somewhat difficult to identify the primary
task of the case system
 Mobile phone, Internet device, social networking
 It was little bit of a challenge to analyze primary task
issues
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Results
 The case also demonstrated how various PSD
principles work together
 Many times pairing PSD principles has good
synergy






Suggestion-reward
Suggestion-personalization
Self-monitoring-reminders
Similarity-liking
Competition-cooperation
Simulation-rehearsal
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Results
 Sometimes principles do not complement each
other
 Abundant use of reduction makes tunneling nearly
irrelevant
 Rather than trying to utilize all principles, a
coherent set of features should be applied
 Emphasizing synergy
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Results
 Another challenge was that some of the PSD
model’s social support principles are highly
interlinked with each other
 Social learning vs social comparison
 Some aspects are difficult to evaluate at
specification level
 Credibility support
 No explicit evaluation heuristics have been
defined yet
 Without heuristics the evaluation is more subjective
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Theories and PSD
 PSD Model is a meta-level model
 Practitioners can utilize the model and gain benefits
from it
 Researchers still need a theoretical framework to gain
full benefits of the model
 Some suitable theories




Elaboration Likelihood Model (ELM)
Theory of Reasoned Action (TRA)
Theory of Planned Behavior (TPB)
Social Cognitive Theory (SCT)
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Elaboration Likelihood Model (ELM)
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Elaboration Likelihood Model (ELM)
Look at the yes arrows in the model above
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Elaboration Likelihood Model (ELM)
PSD can support most of them
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
ELM and PSD
Dialogue support
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
ELM and PSD
Dialogue support
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Social
support
Teppo Räisänen | University of Oulu
ELM and PSD
Primary
task
support
Dialogue support
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Social
support
Teppo Räisänen | University of Oulu
ELM and PSD
Primary
task
support
Credibility
support
Dialogue support
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Social
support
Teppo Räisänen | University of Oulu
Conclusion
 PSD Model can be used to evaluation purposes
 It could benefit from specific heuristics
 Persuasive principles used in design should be
selected so that they have increased synergy
when utilized together
 By using the PSD Model together with suitable
behavior and attitude change theories we can
achieve better results
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS
Teppo Räisänen | University of Oulu
Questions?
[email protected]
THANK YOU!
OULU ADVANCED RESEARCH ON SOFTWARE AND INFORMATION SYSTEMS