SESAMME: An iPad application for participatory systems
modelling
R. Richards *, C. Smith *, N.A. Setianto °°
* School of Agriculture and Food Sciences, The University of Queensland
> Center for Coastal and Marine Resources Studies, Bogor Agricultural University,
Indonesia
©The University of Jenderal Soedirman, Indonesia
Abstract:
The CCRES project is focused on enhancing livelihoods and food security, improving
community health and wellbeing, and sustaining coastal ecosystems in two case study
areas (El Nido, Philippines and Selayar, Indonesia). A key activity is to understand
how communities currently interact with coastal ecosystems and how these
interactions, along with external factors, have led to current socio-ecological
problems. This work is assisted by an iPad application (SESAMME — Socio-
Ecological Systems App for Mental Model Elicitation). SESAMME provides a
mechanism for eliciting the mental models of local stakeholders framed around
specific problems at specific geographical locations. It uses a ‘drag and drop’ icon
approach that provides a graphical and participatory means to mapping out
socio-ecological systems. It is used in conjunction with a methodological script
that guides participants through this process in a consistent and repeatable
manner. Overall, we found the use of an interactive app to elicit mental models to
be effective and engaging. This was based on comments received from the CMTs
who facilitated the FGDs and the comments from many of the FGD participants.
SESAMME has been developed for iPad and is planned to be available via the
Apple App Store after further refinement and testing has taken place.
Keywords: SESAMME; iPad app; rich pictures; community based system dynamics;
socio-ecological systems; mental models; focus group discussions
INTRODUCTION
Socio-ecological problems relate to the complex webs of relations between dynamic
ecosystems and the many scales of social networks (e.g. fishers, households, regional
institutions etc) that interact with them (Murray et al., 2009). They (socio-ecological
problems) are often framed as ‘wicked’ problems (Davies et al., 2015), or as the
stakes are raised, as ‘super wicked problems’ (Lazarus, 2009), characterised by
multiple spatial (local, regional, national and/or international) and temporal (e.g. days,
weeks, months, decades) scales, multiple dimensions (social, ecological, economics,
political etc), complex interactions and feedback pathways and conflicting
stakeholder interests. As such, there is an increasing awareness that these problems
need to be contextualised from a ‘systems’ perspective that acknowledges the
relationships between system structure and behaviour (Lazarus, 2009). This
characterisation of socio-ecological problems therefore lends itself towards using a
systems thinking and modelling approach (Sterman, 2000; Lazarus, 2009), which has
underlying tenets of holism and model structure (including feedbacks and delays).
There is also an increasing awareness that, in conjunction with using a systems
approach, active engagement with key stakeholders is needed when developing
models for managing socio-ecological problems (Davies et al., 2015). Such
‘participatory modelling’ is known to increase the sense of ownership, place and trust
in model outputs (McAllister et al., 2006) and help gain common understanding of a
problem (Senge and Sterman, 1992). Participatory systems approaches are commonly
applied in urban and regional planning and natural resource management with
established techniques to capture information from stakeholder groups (Mendoza and
Prabhu, 2006). This includes community-based system dynamics (CBSD) that draws
upon, and engages with, the knowledge domain of a community (Nadkarni and
Shenoy, 2004; Hovmand, 2014). Engaging with the community helps the community
itself make decisions based on their assumptions and inferences (i.e. mental models)
about how a system operates (Hovmand, 2014), building social capital in the process.
The application of participatory systems modelling is accelerating, however, the tools
available to help practitioners implementing the participatory systems modelling
process are limited. This is especially so for mental model elicitation where there
appears to have been little development of tools beyond the use of white boards,
butchers paper, marker pens and post-it notes (e.g. Richards et al., 2013). This is not
to denigrate the use of these tools, which are readily available and do not suffer the
‘slings and arrows of outrageous fortunes’ associated with more sophisticated
technology. However, the rapid emergence and near ubiquitous (as evidenced by
social network software such as Facebook) use of computer tablets and smart phones
(collectively termed here as ‘smart technology’) coupled with their increasing
functionality provides an opportunity to harness this technology for participatory
modelling.
In particular, there is capacity for developing and using mobile applications (apps) as
tools for improving the mental modelling process. The processing power and
capabilities of smart technology is increasing nonlinearly in a manner that parallels
computer development to the point where they themselves resemble powerful
computers. Furthermore, the availability of application programming interfaces
(APIs) means that the functionality of these apps can be expanded to include features
such as mapping (Apple, Google maps) and cloud database storage (e.g. Parse) that
cannot be provided by the traditional ‘pen and paper’ approaches. They (apps) can
also help with the “strive for visual simplicity” (Andersen and Richardson, 1997) by
providing active filtering of mental model components through hiding and un-hiding
selected elements. In the work presented here, this is exemplified by the ability (of
our developed app) to selectively show / hide parts of a system map, allowing this
visual simplicity to be achieved quickly and cleanly during a workshop.
In this paper, we present an iPad application (SESAMME — Socio-Ecological
Systems App for Mental Model Elicitation) that has been developed and used within
the Capturing Coral Reef & Related Ecosystem Services project. The purpose of
SESAMME is for eliciting rich picture mental models themed around specific socio-
ecological problems from community members during focus group discussions
(FGDs). The developed rich pictures are a precursor step in developing causal loop
diagrams (CLDs) and ultimately stock and flow models to be used for decision
support.
This paper is divided into the following sections:
- Brief description of the CCRES project
- Core modelling teams used to run the FGDs
- The FGD script used to guide the use of SESAMME
- Detailed description of SESAMME
- Summary of outputs derived from using SESAMME
- Conclusion
CCRES PROJECT
The CCRES project (www.ccres.net) is funded by the Global Environment Facility
(GEF), the World Bank and The University of Queensland (UQ), Australia. CCRES
is primarily focused on enhancing livelihoods and food security, improving
community health and wellbeing, and sustaining coastal ecosystems in two case study
areas (El Nido, Philippines and Selayar, Indonesia). One key activity of CCRES is the
Systems Analysis activity, which aims to understand how communities currently use
and interact with coastal ecosystems (i.e. coral reefs, mangrove forests, seagrass beds)
and how these interactions, along with external factors, have led to current problems
such as fish catch decline, mangrove loss, water pollution and food insecurity.
CORE MODELLING TEAMS
Central to the Systems Analysis activity of CCRES are the active engagement with in-
country partners, formation of core modelling teams (CMTs) and specification of
socio-ecological problems to be investigated by the CMTs.
In-country partners were engaged for both case study areas (El Nido, Philippines and
Selayar, Indonesia) to help coordinate, design and implement the FGD conducting
with community participants (Table 1). The partners were drawn from two
universities (one in the Philippines, one in Indonesia), local authorities and a non-
governmental organisation. Their involvement in the Systems Analysis activity was
critical because of their knowledge of the local issues (including socio-ecological
problems), values and the language of the local communities.
The in-country partner organisations formed CMTs, used to engage with the
communities and lead the design and implementation of the FGDs. Hovmand (2014)
provides an overview of the range of roles that can be incorporated within a CMT.
For the CCRES project, four key roles were identified for each CMT:
- Facilitator — leads the discussion during the FGD proper
- iPad operator — controls the SESAMME app during the FGD
- Recorder — takes notes during the FGD and manages the notes post FGD
- Runner ~ assists the other roles in running the workshop
Training workshops covering systems theory (systems thinking and system dynamics)
and the FGD methodology (assignment of roles within the CMTs and role-specific
training where required e.g. operating SESAMME; development of an accompanying
script for the FGDs) were conducted before the FGDs. These parallel the
requirements detailed in Hovmand (2014) and include:
(1) Raising / normalising the CBSD skills across the CMTs (systems modelling
capacity building);
(2) Identifying the roles and requirements for each CMT member before, during and
after FGDs;
(3) Developing a common language to be used by the CMT during the FGDs;
(4) Developing workplans (process maps) detailing the timing and location of FGDs;
(5) Developing the script that methodically details the FGD process; and
(6) FGD training and rehearsal.
The socio-ecological problems (Table 1) investigated in the CCRES project were
selected after a series of site visits to Selayar and El Nido, consultations with the in-
country partners and review of the results of prior community surveys.
Table 1. In-country partners used to form CMTs and the socio-ecological
problem assigned to each. Note there are two CMTs formed at Palawan State
University (PSUa and PSUb).
Socio-Ecological Problem In-country Partner Country
Food insecurity El Nido Foundation (ENF) Philippines
Mangrove loss Palawan Council for Sustainable Philippines
Development (PCSD)
Water quality Palawan State University (PSUa) Philippines
Fish catch decline Palawan State University (PSUb) Philippines
Coral reef fisheries decline Bogor Agricultural University Indonesia
(IPB)
SCRIPT
The FGDs are run according to a script, which is a logical and repeatable process that
allows the results of multiple FGDs to be compared. Such scripts provide a scientific
basis for group model building (Andersen and Richardson 1997, Hovmand, 2014).
Table 2 provides a summary of the main steps in the script used to elicit socio-
ecological systems information from FGD participants.
Table 2. Main steps in the script used to elicit socio-ecological systems
information from FGD particip
Step Description
Activity Activities related to the problem are identified e.g. fishing,
identification mangrove cutting (for mangrove loss)
Resource Resources affected by the Activities are identified e.g. reef fish,
identification mangroves, coral reefs. The current condition (state) of each
resource is also identified.
Trends in | Past, Expected Future and Desired Future trends are assigned to
Activities and | each Activity and Resource.
Resources
Pressures Pressures that have influenced the past trends in Resources and
Activities are identified (e.g. population growth, commodity prices,
climate change).
Trends in | Past, Expected Future and Desired Future trends are assigned to
Pressures each Pressure.
Interactions Direct interactions between Resources, Activities and Pressures are
identified. These are elicited as interaction sets (i.e. Resource —
Resource; Activity — Resource; Pressure — Activity; Pressure —
Resource, Activity — Activity, Pressure — Pressure interactions)
Decisions Decisions that stakeholders could make to address problematic
trends in Resources, Activities or Pressures are identified.
THE SESAMME APP
SESAMME is an app developed for the CCRES project to aid the elicitation of local
stakeholder mental models framed around specific socio-ecological problems at
specific geographical locations.
SESAMME uses a ‘drag and drop’ icon approach (see Figure 3 for example icons
used in SESAMME) coupled with the Apple maps application programming interface
(API) (comparable to Google Maps). The rationale for using an icon-driven approach
(as opposed to a text-driven one) is that it does not make assumptions about the level
of literacy of the participants.
It is important to note that whilst SESAMME was developed as an interactive tool for
creating rich picture representations of mental models for a given socio-ecological
problem, operation of the app within the Systems Analysis activity is restricted to a
trained operator (the iPad Operator in the CMT). This means that only the iPad
Operator needs to know how to operate SESAMME. To make the app accessible to
the FGD participants, the iPad that SESAMME is being run from is connected to a
projector and the iPad screen area projected onto a viewing area where it is visible to
all. The workshop participants provide the input regarding the rich picture map
composition (e.g. which icons to use, where to place them on the map, how they are
interconnected etc) via facilitated discussion.
App architecture
The base architecture of the SESAMME app is shown in Figure 1. SESAMME opens
with a splash screen whilst the app is launching. The app then opens the title screen,
which contains base information (app title, author names, development version). The
app is launched by touching the screen, taking the user to the log-in screen. Here, the
user must enter a user-name and can elect to use the app in ‘local’ (draws upon data
stored on the iPad) or ‘global’ (draws upon data stored on the ‘cloud’) mode. After
successful log-in, the FGD details screen is launched. This provides an opportunity
for the user to add details about the CMT (i.e. enter the names of the Facilitator, iPad
Operator, Recorder and Runner), the workshop (date, location, stakeholder group) and
up to 20 FGD participants. From here, the main interactive screen (Mapping area in
Figure 1) is opened.
Splash screen
Log-in
GD details Icon type select
(Activities Resources, Pressures, Decisions)
File manager
(Load, Save, Cloud access)
Side menu
Diagnostics
Settings
Figure 1. The architecture of SESAMME showing the main screens and the
pathways between them
Mapping Area
The mapping area provides the main focus screen for rich picture development during
the FGDs. As outlined earlier, the rich pictures are constructed according to a script
(see Table 2). As the rich pictures are constructed, icons are used to represent
Activities, Resources, Pressures and Decision and these linking these together with
interactions to indicate causality. All icons, regardless of type, have a common range
of attributes that can be set, facilitated by the object-oriented programming used to
construct SESAMME (iPad apps are programmed within the object-oriented software
platform of Objective-C using the software development kit of Xcode).
An example of the default layout for the map area is shown in Figure 2. This shows
the map area and four menus that can be accessed, three of which are hidden (side
menu, icon menu top, icon menu bottom).
ras 00am ad
= ;
— . —~Side menu access
— as
PAKISTAN °,
are pT, chro
baNataDEsM
wwoin
Figure 2. Mapping area screen
Side Menu
The side menu (Figure 3), which is accessed by pressing the top left of the screen,
contains four types of options:
- Map building functions
- Rich picture diagnostics
- File management
- App settings
Load ALL from cloud
‘Show details
By se
Figure 3. The side menu options that appear
Map building functions: The map building functions follow steps in the script used for
rich picture development (Table 2).
Rich picture diagnosis: The diagnostics tool allows basic statistical summaries to be
carried out on the rich pictures. Its intended purpose is to provide rapid assessment of
one or more rich picture maps (e.g. across multiple rich pictures it can address
questions like how many times is an interaction between fish and fishing recorded).
This tool is still under construction but the function (and current development status)
of each of the components within it is as follows:
- Select maps (Figure 4): allows the user to select which maps saved on the iPad are
to be used for diagnostic testing (this is limited to rich pictures saved on the iPad).
This tool is fully functional.
Summary statistics (Figure 5): Displays the bulk statistics for each of the icon
types for each map. This tool is fully functional.
Icon type diagnostics — This tool will allow (this tool is under construction and is
not yet functional) a more in-depth analysis of a specific icon type (Resource,
Activity, Pressure, Decision) across the selected maps. It is planned that this tool
will identify and summarise for each selected rich picture map (using histograms
and tables) which icons are connected to each other, the polarities of these
connections, icon trends and icon states.
pal > asa an 100% me +
< Back
Figure 4. Select map function. Multiple maps can be ‘selected’ to be included in
the diagnostic testing.
rad > 453.0M 100% os
< Back
Figure 5. Summary statistics function. This shows a frequency count of the
different icon types (Resource, Activity, Pressure, Decision) for each map.
File management: SESAMME has a basic file management system that allows rich
picture maps to be saved to, and loaded from, the iPad and cloud storage (Parse). An
early development requirement for SESAMME was that it be connected to cloud
storage so that (1) rich picture maps could be backed up to a secondary storage unit
(the primary storage is the iPad) to provide data security and (2) the rich pictures
developed could be shared (upon upload to the cloud). The process of saving files to
the iPad and cloud storage is outlined within the script.
Settings: This page is, to date, quite rudimentary in that the only setting that can be
adjusted is whether the iPad is connected to the cloud (online) or not (offline). It was
important that the iPad operator could purposely switch between online and offline
modes, particularly in areas where Wi-Fi is intermittent. This situation (intermittent
Wi-Fi) can cause SESAMME to make repeated attempts at connecting to the cloud
database, essentially ‘locking’ up the iPad for several minutes. This issue of
intermittent connection and searching is addressed in the script by requiring that
SESAMME be used offline during the FGD.
Management of Public Procurement of Information System
Yutaka TAKAHASHI
School of Commerce, Senshu University
2-1-1, Higashimita, Tama, Kawasaki, Kanagawa, Japan
takahasi@isc.senshu-u.ac.jp
Abstract
Developed countries’ national governments operates huge and complicated information
systems in order for daily public services. In the United States, public officers can
negotiate about procurement conditions in order to find not only the lowest price but also
high quality. On the other hand, Japan, believed one of the developed countries, prohibits
public officers from making negotiations when they procure any public materials and
services. Today’s information systems are complicated: it is always challenging to express
what to be implemented. Therefore, less communication including negotiations causes
lower satisfaction of users, more expensive prices, and longer delivery time. From this
viewpoints, Japan’s regulation controlling public procurement is not practical.
Nevertheless, they do not plan to change the procedure shortly. This paper shows the
reason that Japan’s law set such a strict procurement rule and explore problems using
system dynamics models.
1. Introduction
Information systems procurement in government (IT public procurement), as
beginning a new development or updating of an existing system, has a wide variety. In
particular, national government’s information system is extremely large in size and
complicated compared to private companies: both software and hardware are supposed
to work not only at headquarters but also nationwide, and required lowest failure rates.
Moreover, the systems’ structures and operations need to obey the low and detailed
guidelines.
This kind of challenging procurement must need consistently good communications
between purchaser and vendors from an early stage of procurement. For example, the
United States of America permits purchase officers a kind of negotiation about price and
other conditions. Complex systems contain various functions, and functions are put in
order of importance. Public officers in the USA can balance price and quality.
On the other hand, public procurement in Japan is quite different: their law requires
Bottom Icon Control Menu
The bottom right hand /con Control Menu is where the library of available icons is
located. These icons are dependent on which icon type has been selected (e.g. if
Resources has been selected then only resource-type icons will be shown — refer to
Figure 6 for an example of icons for each type) and are hidden by default. Originally,
this menu was always shown and feedback from early testing indicated that showing
the icons biased participant responses.
Icon Icon examples
category
Resources @ ( Sd 2.
user-define coral cattle reef fish crops
Activities os >
A> whi |
TA
user-define fishing egg ploughing _ tree clearing
gathering
Pressures 49 P )
user-define climate population economy air pollution
change growth
Decisions
user-define
Figure 6. Examples of icons for each icon category (Resources, Activities,
Pressures, Decisions) in SESAMME
When starting a new map, the default category is ‘RESOURCES’ (as shown in Figure
7). A sub-set of 12 icons from the total available icons is shown at any one time.
Pressing the header of the icon menu (refer Figure 7) will replace this subset of visible
icons with a different sub-set. Note that the first icon is a ‘user define’ icon that is
available for use when it is deemed by the FGD participants that there is not an
appropriate icon in the library. Also, note that the text label attached to any icon,
which is also shown when the icon is dropped into the map view, is editable allowing
the icon to be re-labelled.
vad = 03 AM ¢ 00K mms
ub-set of
e) icons for
and-dropping
Figure 7. The icon menu after the ‘SHOW’ button is pressed. Note that the icon
category header (i.e. the ‘RESOURCES” tab) is a button that enables different
sub-sets of the icons to be presented.
The bottom side menu can also be set to show a library of qualitative trend icons
(Figure 8) when Trends has been selected from the side menu (Figure 3 — map
building functions). These trend icons indicate the past, expected future and desired
future trends for a particular icon.
—YAYVOACEK
no change linear linear exponential exponential goal seek goal seek
increase decrease increase decrease high low
S-curve S-curve goal seek goal seek negative positive oscillating
high low high with low with peak peak
overshoot —_ overshoot
Figure 8. Trend available in SESAMME that can be assigned to an icon
Trends are attached to icons using sub-views so if the parent icon (the reef fish in
Figure 9) is moved the sub-viewed trends will move with it. The trends are sub-
viewed depending on whether they represent past (located upper left of the parent
icon), expected future (upper right) or desired future (extreme upper right) trends.
Note that SESAMME does not require that all trends (past, expected future, desired
future) be assigned to an icon i.e. just one of these trends can be assigned.
Sap
I i
past expected desired
future future
Figure 9. Example of how the trends assigned to an icon appear in SESAMME
Top Icon Control Menu
The top menu is for controlling how the icons are displayed within the map display.
This includes a rubbish bin icon that is used for deleting icons and interactions (arcs)
from the view map area, a note function so that comments can be added to the map
and a set of view toggle buttons that can be used hide and un-hide components of the
rich picture (e.g. trends, labels and/or interactions can be hidden to de-clutter the
map).
The top icon control menu can be used to enter into SESAMME’s edit mode. Edit
mode allows changes to an existing rich picture to be tracked (similar to Track
Changes in Microsoft Word), so that if the rich picture is modified between FGDs, the
changes can be recorded. SESAMME’s edit mode has three modes (Figure 10):
edit -reflects the updated rich picture and is where deletions (e.g. icon removed),
additions (e.g. icon added) and modifications (e.g. polarity changed) are made.
- highlight —changes made to the original rich picture are highlighted.
- original —where the original form of the rich picture (i.e. before edits were made)
is shown.
Figure 10. Example of how an edited map appears in each of the three edit
modes (Original = pre-edited map; Edit = revised form of the map; Highlight =
changes made to the map).
Map Control Menu
The menu bar located along the bottom of the screen contains several functions that
can be applied to the map area. This section briefly describes the use of each function
within this menu.
Map Padlock: A padlock icon located at the bottom left of the screen is used to lock
the map when icons are being dragged onto it, otherwise the map will move when
icons are dragged.
Arc / Interactions: This button allows arcs (interactions) to be added to the rich
picture. Whilst this is enabled, neither the map nor the icons can be dragged (because
SESAMME reserves the touch screen movements to drawing arcs on the map). The
process for creating an arc between two icons is shown in Figure 11. To start an arc,
an icon is touched (a gold rectangle around the icon indicates it is a valid icon! for
developing an arc). Once tethered to a start icon (always contextualised as the ‘cause’
icon in a cause-effect relationship), the arc can be dragged to another valid icon (the
‘effect’ icon). A second gold rectangle around the destination icon indicates that the
connection can be made and is completed by lifting the finger off the screen at this
stage. The colour of the arc changes from red (signifying arc under construction) to
1 The concept of a valid arc is used in the code to differentiate between touching a rich picture
icon (resource, activity, pressure, decision) and a subsidiary icon (e.g. a trend icon - refer Figure
11)
black (are has been successfully constructed). The constructed arc also contains a grey
box at the mid-point, which is used to set the polarity of the arc.
1. A red arc is tethered to a start (cause) icon
2. The arc is dragged to an end (effect) icon
3. A successfully constructed arc changes to
black with a grey ‘polarity’ box appearing at
the mid-point.
Figure 11. The process of creating an arc (interaction) in SESAMME. The gold
rectangles shown for the start (top panel) and the end (middle panel) icons
indicate that they are ‘valid’ icons for tethering the arc to.
State: This button is used to set the current state or condition of an icon. When state is
enabled, touching any Resource, Activity, Pressure or Decision icon in the map view
will sequentially cause it to change colour in the order shown in Figure 12. The colour
scale legend is: red = poor condition, orange = moderate condition, green = good
condition.
Figure 12. Example of the three state settings shown for the reef fish icon and
how this changes the icon’s appearance in the map view.
Exog: This button is used to allow individual Resource, Activity, Pressure or Decision
icons to be specified as exogenous (external to the system) or endogenous (internal to
the system). By default, any icon added to the map view area is endogenous and
becomes automatically geo-located to the map (i.e. if the map is panned or zoomed
then the icon will automatically move with it). Touching an icon within the map area
when the Exog button has been set will change the icon from endogenous to
exogenous (and back again if the same icon is pressed again). If an icon has been set
as exogenous, then its appearance will fade and it will no longer move with the map.
Hence a rapid way of determining which icons are exogenous in a ‘crowded’ rich
picture map is to move the map area and see which icons do not move.
Clear: This button is used to clear the entire map of all icons. Pressing this button
instantiates an alert call-out window that checks this action (‘Clearing all icons. Are
you sure’). If actioned, all icons and interactions are cleared from the map
permanently.
Map: This button enables toggling between regular, satellite and hybrid (regular and
satellite combined) map views.
Undo / Redo: These two buttons are edit buttons that allow sequential changes to be
recovered (undo) or repeated (redo). However, these functions are still undergoing
development and are not functional at this stage.
OUTPUTS FROM THE FGDs
The SESAMME app was recently used in over 166 FGDs in El Nido and Selayar.
Observations and feedback from these FGDs indicated that SESAMME was a good
mechanism for engaging with communities and capturing their mental models related
to socio-ecological problems. Some issues with SESAMME were identified (bugs in
the coding, usability of the app) during the FGDs with most resolved quickly because
of clear communication with the CMTs and the use of the Internet to remotely update
SESAMME.
An example of a rich picture developed from an FGD using SESAMME is presented
in Figure 13. Features of this rich picture are:
¢ Resource, Activity, Pressure and Decision icons;
¢ ‘States’ assigned to Resource icons e.g. the fish icon in the middle of the map;
¢ Trends (past, expected future, desired future) assigned to Resource, Activity
and Pressure icons;
¢ Interactions indicating causal relationships between icons;
¢ Polarities on the interactions indicating the direction of relationship between
connected icons.
+e 9:46 AM 100% mms
Figure 13. Example of rich picture developed using SESAMME for an FGD held
in Selayar, Indonesia. Note that the ‘state’ button is enabled, causing icons that
have been assigned states to be highlighted (e.g. the red fish icon in the middle of
the map is red, indicating poor condition).
The following figures present summaries of some of the data elicited from 17 FGDs
conducted in Selayar, Indonesia, using SESAMME. The focus problem for these
FGDs was ‘Coral reef fisheries decline’.
Figure 14 summarises the percentage of FGDs in which different activities were
identified. For example, ‘Coral Fishing’ was identified in all FGDs. Other forms of
fishing (Poison fishing, Fish bombing) also featured frequently. Conversely, some
Activities (e.g. Hobby fishing, Sea weed farming) were identified at a small number of
FGDs, perhaps reflective of site-specific activities.
Hobby fishing
Sea weed farming
Sambak fishing
Sero fishing
Other destructive fishing
Fish trading
Squib farming
Crab farming
Fish farming
Coconut farming
Cashew farming
Tourism
Setting the marine protected area
Bagang fishing
Compressor fishing
Big vessel fishing
Fish bombing
Poison fishing
Coral fishing
bas
a
Sg
Figure 14. Summary of the Activities reported across the FGDs conducted in
Selayar, Indonesia. The number at the end of the row indicates the % of FGDs in
which an Activity was identified (Adrianto et al., 2015).
Figure 15 summaries the trends (past, expected future, desired future) assigned to
Activities across the FGDs, e.g. 67% of the FGDs perceived that Coral fishing has
been increasing (past trend) and will continue to increase (expected future trend).
However, only a third (33%) of FGDs expressed a desire for fishing to continue to
increase.
ica ead | Percentage
so] | | | | hh
Coral Past. 20 0 0 13 67 ie) 0
fishing Expected 26 0 0 7 67 0 0
Desired 34 0 13 20 33 ie} i)
Poison Past. 29 7 0 29 28 0 7
fishing Expected 36 0 0 36 28 0 0
Desired 43 i) 57 0 0 (o} 0
Fish Past 64 36 0 0 0 0 0
bombing Expected 64 ie) 0 27 9 0 0
Desired 3 0 91 0 i) io) i)
Big vessel Past 13 ie) i) qd 75 ie} i)
fishing Expected 25 QO 0 12 63 ie) ie)
Desired 75 0 13 12 to) ie) 0
ee eccems Past 67 17 0 ie) 16 ie} 0
fishing Expected 33 ie) 0 0 67 0 0
Desired 67 ie) 7 0 16 ie} iv)
Figure 15. Summary of the Activity trends reported by FGDs, Selayar, Indonesia
(Adrianto et al., 2015).
The data extracted from the SESAMME rich pictures can also be used to summarise
the frequency with which interactions between ‘icons’ were elicited. For example,
Figure 16 highlights the Activity — Resource interactions across the FGDs and the
nature of these interactions. This figure highlights that ca. 93% of the FGDs identified
interactions between Coral fishing and Fish (negative polarity) and Fish and Coral
fishing (positive polarity). This particular result highlights that most of the
communities perceive a balancing feedback loop between the Activity (Coral fishing)
and the Resource (Fish).
Nine balancing loops (the blue lines in Figure 16) between Activities and Resources
were identified across the FGDs (Adrianto et al., 2015). These interactions (along
with interactions that did not form loops e.g. the grey and orange lines in Figure 16)
and interactions between other icon types (e.g. Resource — Resource) are being used
to construct causal loop diagrams (CLDs). An example of a preliminary CLD that has
been developed from the FGDs conducted in Selayar is presented in Figure 17. These
CLDs provide a bridge between the rich picture maps elicited using SESAMME and
stock and flow models that will be developed to simulate the effect of feedback loops
on system dynamics.
compressor fishing - coral meen
compressor fishing - fish [ih ees
big vessel fishing - coral
big vessel fishing - fish sees
Poison fishing - sea grass
Poison fishing - sea weed
<
2 Poison fishing - fish
fa [-,0]
g Poison fishi H
¢ ‘ison fishing - COral ps = [0,4]
Fish bombing - sea grass aL]
Fish bombing - fish
Fish bombing - coral oem
Coral fishing -sea grass ——
Coral fishing - coral
Coral fishing - fish
°
5S
nN
8
8
S
é
Ss
2
8
4
3
~
8
©
i}
Ss
s
Percentage
Figure 16. Summary of Activity — Resource interactions reported by FGDs
conducted in Selayar, Indonesia (Adrianto et al., 2015). Note [-,0] means there
was a negative relationship from the Activity to the Resource but no relationship
from the Resource to the Activity. [0, +] means that there was no relationship
from the Activity to the Resource but there was a positive relationship from the
Resource to the Activity. [-,+] means that the Activity had a negative relationship
with the Resource and the Resource had a positive relationship with the Activity
(i.e. feedback loop exists).
no negotiation” between purchaser and vendors. The reason is mainly to prevent
stakeholders from making collusive relationship. Indeed, it used to be many cartels, and
sometimes they were led by a purchaser itself. Therefore, Japan’s law eliminates public
officers’ authority related to public procurement by introducing the Act concerning
Elimination and Prevention of Involvement in Bid Rigging etc., in 2007 (Japan Fair
Trade Commission, 2007).
Then, all conditions related to procurements need to be documented and published
before bidding. Purchasers are encouraged to write more and more detailed specification.
It was believed to make deals transparent, venders’ risk lower, and public service more
efficient.
However, the fact was not so successful. Originally, most public officers are not trained
to deal with information systems development. Therefore, even if purchasers wrote
detailed procurement documents, delivered products are sometimes unsatisfactory.
Besides, detailed specification was sometimes unpractical or out-of-dated, budget is often
big, and purchasers lack incentive to save money. Then, there are huge complicated and
unsatisfactory systems. Even if budget is lowered, prospective vendors would not bid
such deals because of high risk of projects. Then, some big projects’ bidding failed. This
means public operations’ delay or stop.
IT public procurement suffer from the following typical difficulties or problems
(Takahashi, 2014): failure to let vendors understand purchaser’s wishes, collapse of
project due to products’ complexity, misunderstandings of budget and proposal, vendor
lock-in, virtually limiting prospective bidders for bulk procurement, and unethical
activities by vendors exploiting purchasers’ technological weakness.
These problems, because it did not occur much in the procurement of non-IT-related
goods, services nor civil engineering, it can be said unique problems in IT public
procurement. It should be noted that with respect to IT procurement and operations in
the country, the Board of Audit of Japan (2004) conducted a survey, and the Ministry of
Internal Affairs and Communications (2007) shows the guidelines.
Although there are many unsuccessful public procurement cases, Japan’s government
is reluctant to change this situation. This paper shows that the problem is not a matter
of each public officer’s ethics nor ability but caused by problem structure, using system
dynamics. Simulation result explains that encouraging purchasers to make detailed
specifications is rather to make IT public procurement difficult.
2. Problem Structure in IT Public Procurement
Figure 17. Common interactions among Activities, Resources, and Pressures
reported by FGDs held in Selayar, Indonesia (Adrianto et al., 2015)
CONCLUDING REMARKS
Overall, we found the use of an interactive app in conjunction with a script eliciting
rich picture maps from stakeholders to be effective and engaging. This was based on
comments received from the in-country partners who facilitated the FGDs and the
comments from many of the FGD participants. SESAMME is being developed for
iPad and is only currently available through Apple ‘ad hoc’ distribution. However, it
is planned to be available via the Apple App Store after further refinement and testing
has taken place. These refinements include allowing the users to create their own icon
categories and import their own icons into these categories. This will allow
SESAMME to be applied to any problem.
REFERENCES
Adrianto, L., Setianto, N.A., Kusumo, S., Affata, S. (2015). Report on Scoping and
First Round of Focus Group Discussions in Kabupaten Kepulauan Selayar, Indonesia.
Capturing Coral Reef and Related Ecosystem Services draft report. December 2015.
Andersen, D.F., Richardson, G.P. (1997). Scripts for group model building. System
Dynamics Review, 13 (2), 107-129.
Davies, K.K., Fisher, K.T., Dickson, M.E., Thrush, S.F., Le Heron, R. (2015).
Improving ecosystem service frameworks to address wicked problems. Ecology and
Society 20(2), 37.
Hovmand, P.S. (2014). Community based system dynamics. Springer, New York.
Lazarus, R.J. (2009). Super Wicked Problems and Climate Change: Restraining the
Present to Liberate the Future, 94 Cornell L. Rev. 1153.
McAllister, R.R.J., Gordon, LJ., Janssen, M.A., Abel, N. (2006). Pastoralists'
responses to variation of rangeland resources in time and space. Ecol. Appl. 16, 572—
583.
Mendoza, G., Prabhu, R. (2006). Participatory modeling and analysis for sustainable
forest management: overview of soft system dynamics models and applications.
Forest Policy and Economics 9, 179-196.
Murray, G., Neis, B., Johnsen, J.P. (2006). Lessons Learned from Reconstructing
Interactions Between Local Ecological Knowledge, Fisheries Science, and Fisheries
Management in the Commercial Fisheries of Newfoundland and Labrador, Canada.
Human Ecology, 34,(4), 549-571.
Nadkarni, S., Shenoy, P.P. (2004). A causal mapping approach to constructing
Bayesian networks. Decis. Support Syst. 38, 259-281.
Richards, R., Sano, M., Roiko, A., Carter, R.W., Bussey, M., Matthews, J., Smith,
T.F. (2013). Bayesian belief modeling of climate change impacts for informing
regional adaptation options. Environ. Modell. Softw. 44, 113-121.
Senge, P.M., Sterman, J.D. (1992). Systems thinking and organizational learning:
acting locally and thinking globally in the organization of the future. Eur. J.
Oper. Res. 59, 137-150.
In order to succeed in procuring information systems, purchasers are encouraged to
make requirements clear and detailed, which cannot be misunderstood by prospective
vendors. In an idealistic thought, full information can pull the best products without
schedule overrun. If not so idealistic, better products would be expected.
This thought is illustrated as a causal loop diagram in figure 1.
Degree of granularity
of requeirements
‘The argument to make more
detailed requirements go straight to
Vagueness in| total cost decrease.
Gap between
expected and actual
quality Total cost expended
in a procurement
ee
Figure 1. Expectation that detailed description makes the best products.
Reworks'
However, related activities cause side effects. For example, detailed documentation
needs additional time and money (overtime pay, research fee etc.). This relationship is
shown in figure 2. The thick arrow is original thought expressed in figure 1. Side effects
bring about increase of total cost expended in a procurement. This is an undesirable
effect.
Degree of granularity
of requeirements
- Cost of preparation
} for procurement
Vagueness in He TOLD!
f,
Time to publish :
requirements
requirements
| "
Gap between
expected and actual
ality Total cost expended
© Acceptable in a procurement
delivery time "
Ne Reworks:
Figure 2. Side effects.
In particular, shortening delivery time is problematic. Shorter time to bid means
prospective vendors have shorter time not only to make proposals but also to complete
their system development and delivery. If the project failed, payment of a penalty for
breach of contract would occur. Even if not paid the penalty, it is realistic to suppose
increase in rework and maintenance costs due to a rough-and-ready business.
Prospective vendors often shun these risks. In such a situation, no one bids the
procurement bid and cause significant delay of public officers’ tasks. Because of Japan’s
law, raising budget for such unsuccessful procurement is prohibited, public officers need
to correct their plans significantly in order to succeed in the second bid. This makes
public officers reluctant to increase granularity of requirements. This closes a loop in
figure 3.
=_ Degree of granularity
of requeirements
a Cost of preparation
— for procurement
‘Vagueness in
Time to publish :
requirements
/ requirements
Probability to fail
to collect bids |
Gap between x
expected and actual ial ood expenited
- quality ee ae
Acceptable in a procurement
~delivery time Ni. reworts
Figure 3. Appearing the possibility of failure in procurement
Failure of the system procurement of the Patent Office of Japan has been known as a big
thing (Asakawa, 2012). The project over eight years, is obtained not leave only result in
discontinuation. This is the person who created the specifications was transferred
immediately before bidding, after which the specification changed to completely different
from those at the time of bidding, the project has been extremely vague at that point. To
restore the original purchasing officials for settling the situation, specification was also
supposed to development back. But then that this staff was going to provide
inappropriate information to bidders was supposed to leave the revelation to again
project. This is an extreme example, but public officers tend to hope to settle all
procurement as soon as possible. Thus, they must not fail the bid collecting contractors.
Rework in information system related tasks is not simply “do again.” Rework makes
product’s structure complex. This also becomes a seed of other troubles. More reworks,
more additional management and hidden troubles. Figure 4 shows this effect, as a center
feedback loop.
=, Degree of granularity
of requeirements
" Cost of preparation
. 4 for procurement
: im Vagueness in socal
/ Time to publish .
| : requirements
| requirements Y
Probability to fail i" Complexity of
to collect bids | product
id Gap between + ae
Gepected and netial } Total cost expended
ality :
in a procurement
Ze
_ q
Acceptable
delivery time 4 :
Na Reworks
Cost affair is also making feedback loops. In the information business, it is often said
that vague requirement raises project failure risks. CIO Magazine (2003) in exquisite
creation of the RFP in the Ministry of Economy, Trade and Industry has been featured.
Request it is possible to be scrutinized and documented in the ordering party, the party
consciousness increases for the ordering party functions and operations. This, and the
priorities of the functions implemented become clear, carefully and tested during the test
or inspection of the trial, it can be expected to reduce the trouble after start of operation.
On the other hand, as has also been mentioned in the same article, in the government of
personnel rotation is in frequent, it is difficult to skill accumulation and take over, a
large burden on the ordering party. Experience of the "summarizes a hard time" is, thus
a connection tend to attitude that refuse to change the specifications made, there is a
possibility that dismiss the appropriate vendor proposals. In addition, overly detailed
specifications are able to narrow the width of the resulting proposed.
Only positive effect on performance for fineness of specifications is often considered.
However, it is understood that in practice there is a path effects applying a brake on it.
Then, high costs bring about cost cut pressure via improvement the quality of
requirement. This makes two additional feedback loops shown in figure 5.
=, Degree of granularity _
ofrequeirements + ~~ Pressure pn purchaser
to improve direction
@)
‘S} 2 Cost of preparation
— for procurement
i, sci
Time to publish Vagueness in
| requirements requirem NO
Y
Probability to fail f a Complexity of
to collect bids Prot
2 a ’ Gap between Gq
cupeeicy and achval ‘otal cast expended
® Acceptable bared in a procurement
~ delivery time
Reworks
—
Figure 5. Cost-improvement in requirement link makes new feedback loops
Original causal relationship image in figure 1 was single route. However, we can find
five feedback loops even limited around purchaser and vendors. Original intention to
efficient work is not promised because of these multiple loop structure.
3. Simulations
Based on author’s experience at one of committees in national ministry in Japan, each
variable is set. Vensim model can be obtained from conference site
(http‘//conference.systemdynamics.org). Base simulation gives parameters as followings:
average time to bid is 12 weeks, vendor’s system development should be in 48 weeks,
and initial value of degree of granularity of requirement is 10,000. Simulation result is
in the figure 6.
Selected Variables
o 1 2 30 40 SO 60 70 80 0 100
Time (Month)
Degree of granularity of Current
Pressure pn purchaser to Current
Probability to fail to collect bids : Current
Figure 6. Base run result.
Figure 6 shows pressure on purchaser to make more detailed documents are gradually
weakened, and the degree of granularity is once boomed but soon return back to original
status.
If prospective vendors prefer more detailed description, say the twice of base run, the
simulation result is in the figure 7.
Selected Variables
2M
2
0.5
1M
1
0.35
0
0
0.2
o 10 2 30 40 50 60 70 80 90 100
Time (Month)
Degree of granularity Current
Pressure pn purchaser Current
Probability to fail to collect bids : Current
Figure 7. “Vendor prefers more detailed requirement” scenario
The result shown in the figure 7 looks similar to base run in the figure 6. In this point,
we can see that prospective vendors’ attitude have limited effect.
4. Conclusion
Many of the IT public procurement is very large scale and complex. In addition,
services to be realized by it affect the convenience and safety of people's lives. Nowadays,
governments and public officers actively attempt to take advantage of cost savings and
private organizations’ knowledge. However, in order to achieve the intent, and
overlooking the entire environment surrounding the information system after
procurement and operation also examined adverse measures, it is necessary to aim a
good management of balance.
Bibliography
CIO Magazine (2003), "the Ministry of Economy, Trade and Industry" public IT select
“sweethearting is not allowed" CIO Magazine 2003 4F June issue
Japan Fair Trade Commission (2007) “Involvement Prevention Act.”
http‘//www.jtfc.go.jp/en/legislation_gls/aepibr.html
Sterman, J. (2000) "Business dynamics." Irwin McGraw-Hill.
Asakawa Naoki (2012) "Government system failure of the essence" Nikkei Computer
July 19, 2012 issue, pp. 22-39.
Board of Audit of Japan (2004), "the results of the audit on the computer system in each
ministry, etc." http://report.jbaudit.go.jp/org/h17/yousei5/2005-h17-8000-0.htm.
Ministry of Internal Affairs and Communications (2007), "Basic Guidelines of
government procurement of the information system"
http://www.meti.go.jp/policy/it_policy/tyoutatu/kihonshishin.pdf
Takahashi, Y. (2014) “A study of IT public procurement using systems thinking.”
Proceedings of the 69 National Conference of Japan Society of Information and
Management.”
