To deal with increasing complexity and connectivity of socio-technical systems it
becomes unlikely for individuals to be able to oversee all possible changes. These
systems are riddled with a plurality of actors with differing interests, disciplines,
institutions and ecological l
...
To deal with increasing complexity and connectivity of socio-technical systems it
becomes unlikely for individuals to be able to oversee all possible changes. These
systems are riddled with a plurality of actors with differing interests, disciplines,
institutions and ecological limitations. Examples of systems like these are energy
and gas grids. If one wants to tackle problems on these systems one would ideally
understand possible results of changing things in these systems as a change in one
part of the system can lead to results in other subsystems. If a tree falls down on
an energy pole, for example, chemical plants can stop functioning. This in turn can
cause orders to be late, influencing a whole production chain.
To this end one would ideally one would ideally apply systems thinking: ”Systems
thinking is a set of synergistic analytic skills used to improve the capability of
identifying and understanding systems, predicting their behaviors, and devising
modifications to them in order to produce desired effects. These skills work together
as a system.”
However, because the earlier pluralities it unlikely for a single individual or organisation to have all the required information for this. Processes are required where
information is collected and co-created with multiple parties in such as system to
enable the creation of comprehensive solutions for these problems. What is required
is social learning. Social learning (SL)is learning that happens by people participating
in so called communities of practice. A community of practice can be seen as a group
of people with converging interests and skills. An example would be a grid operator,
which have their own sub-communities (E.G. cable technician or systems manager).
They can be smaller groups, E.G. a family, and participation is often not mutually
exclusive. Rather than being part of a community of practice one could be seen
as being part of the landscape of practice, consisting out of multiple communities.
By partaking in these communities people gain experiences by both learning and
expanding on a communities’ knowledge.
The aim of this thesis is to gain insights into: 1. mechanics shaping and steering social learning 2. how to measure social learning processes for in vivo experimentation 3. design mechanics for participatory modelling processes and Social learning in
general to improve development of such processes
To do this a theory has been made on the mechanics and behaviour of individuals in
a SL process. This has been done in chapters 4, 5 and 6. To do this a PM perspective
has been used as this give a structure on the actions someone can take and requires
shared information to be structured. Additionally it is seen as a useful tool for
tackling socio-technical problems.
The theory thus focussed on the main action of PM, namely sending information,
receiving and processing information and deciding upon the model. To develop
the theory knowledge from multiple disciplines is needed. To this end a theory has
been developed using supersynthesis. This is a research method where multiple
theories are combined to make a new on to explain something. The aim is not to
supersed the combined theories, but to explain something new. The fields that have
been the focus are communication science, helping understand how information
is processed and sent, and social psychology, helping understand why and when
individuals take certain actions. The theories have been synthesised in two rounds of
conceptualisation, with the first focussing on conceptualising every possible action
and mechanic. The ones that were deemed most interesting or useful have been
conceptualised more in depth. To this an additional conceptualisation of knowledge
and information is made. The final theory is as follows:
Knowledge takes the shape of a knowledge graph. In this graph fields of expertise
are called topics. One can think about the weather as such a topic. These topics
consist out of information items, think cloudiness or temperature, and links between
these items indicating their relation, think cloudiness leads to lower temperatures.
These items may have links to items from other topics. For example cloudiness is
related to sun hours and yields of Photovoltaics from th topic of Photovoltaics.
It is assumed that a complete knowledge graph exists. Each individual knows part
of this graph, signifying knowledge or expertise in the topic. The larger part of
a topic they know the more expertise they have in that topic. This includes both
information items as their links (relationships). All these items, links and topics
also have a perceived relevance for people. This is based on interests or affiliation.
Affiliation means relationship with a group of people, for example, meteorologists.
Something is also seen as more relevant if it is discussed often, attributed to common
knowledge effects. In a social learning process there are several individuals. Each
round they are able to share information. Information is seen as information items
and/or their links. A topic as a whole can also be discussed, but this is not seen as
actual information for learning. What they share is based on the amount of energy
they have and are willing to spend on sharing information. This is dependent on
perceived relevance of the item they are considering to share, their expertise on the
related topic and their tiredness. It is assumed that the energy one has decays over
rounds.
Shared energy is received by others and they start to process it. Here something
is integrated or learned if they are able or willing to invest enough energy in
the processing. This is dependent on their expertise of the related topic, their
perceived relevance of what is shared and their attitude with regards to the sender.
If expertise, relevance and attitudes are high enough someone will process and
integrate information. Processing energy is also assumed to decay each round.
In addition to information on knowledge, individuals can also share relational
information. These are details like hobbies and other personal details. These are
processed as either positive or negative and influence attitudes.
Processing of shared information may also happen during breaks or downtime.
Here one has more energy to spend and attitudes are less relevant. Total recall of
information is assumed in the whole model (I.E. people do not forget anything).
To allow for further reflection on the theory and to act as a proof of concept of
the theory the theory has been translated into an agent based simulation model.
This model has been analysed in a sensitivity experiment using LHS and extremely
randomised forest in addition to a variety of plotting techniques in R.
Additionally two experiments are designed, inspired by real cases. These are used
to reflect on the theory and find less noticeable quirks from the ABM.
Based on the theory and ABM the following things have been concluded:
-While communication science theories and social psychology theories have
been used for theory development, they are not a be all end all. One can apply
other fields if one wants. This specific combination, however works especially
well for an individual perspective.-The theory can be used to reflect on SL by practitioners as it tells why and
how people can act. Furthermore the idea of the knowledge graph can be connected to landscapes of practice, with topics relating to a community and
links between items of these topics to those of other relating places where
boundaries interact
-Matters like conflict and increasing conflict, coalitions, personal inhibitions and
norms are some of the values that would make sense to include in the theory.
This would make the theory less usable for simulation modelling, however
and would add a lot of behaviour that is not directly related to the learning
process. To implement these additional behaviour could be conceptualised and
added, making the theory more complete but less comprehensive. The most
important addition that could be made according to me would be an extension
on the actions influencing attitudes and the actual definition of a process result
(a participatory built model or a plan).
-This new theory is valuable as this individual based perspective has not been
taken before, inviting to reflection on practice.-The knowledge graph could be used as a means for building new theories that
are comparable. Additionally it is a way to explicitly learning.
-The combination between social learning and Participatory modelling has not
been made this explicit before. It would allow participatory modellers to reflect
on their practice.-The ideas of energy for sharing and processing are quite influential in the
ABM. They are interesting as they give clear reasons why learning may not
happen or happen suboptimally. For learning to happen information need to
be shared. If people lack the energy or the willingness to spend energy sharing
will not happen. If they do not have the energy to process this they will also
fail to learn. This highlights the need for keeping energy levels in mind when
designing these processes. It is assumed that these energy levels decay linearly.
While arguably too simplistic still it does show how intensive processes or
boring processes can fail.
The following design mechanics are proposed:1. Usage of a knowledge graph to keep track of what is learned by researchers 2. Usage of knowledge graph to steer the process order that makes learning more
likely (topics that closely relate to all participants first and expand that towards
specific participants later down the line). 3. Use set structures, conceptual modelling, drawings and other tools to make
information sharing and processing easier and less intensive. This would
make the process spend less energy if the used tools are chosen well (I.E. a
conceptual drawing of what is said or what someone wants to explain using
causal diagramming is probably better to explain ideas than doing so via live
programming of a simulation model). 4. Use actions like summarizing what has been said to slow down the process if
it becomes to quick, leading to a processing energy deficit. 5. Use means like using an agenda to ensure the speed of the process does not
become to slow, leading to boredom and potential energy decays
