Different paths to knowledge

The course Scientific methods - Different paths to knowledge introduces the learner to the fundamentals of scientific work. It summarizes key developments in science and introduces vital concepts and considerations for academic inquiry. It also engages with thoughts on the future of science in view of the challenges of our time. Each chapter includes some general thoughts on the respective topic as well as relevant Wiki entries.

Definition & History of Methods

Epochs of scientific methods The initial lecture presents a rough overview of the history of science on a shoestring, focussing both on philosophy as well as - more specifically - philosophy of science. Starting with the ancients we focus on the earliest preconditions that paved the way towards the historical development of scientific methods.

Critique of the historical development and our status quo We have to recognise that modern science is a system that provides a singular and non-holistic worldview, and is widely built on oppression and inequalities. Consequently, the scientific system per se is flawed as its foundations are morally questionable, and often lack the necessary link between the empirical and the ethical consequences of science. Critical theory and ethics are hence the necessary precondition we need to engage with as researchers continuously.

Interaction of scientific methods with philosophy and society Science is challenged: while our scientific knowledge is ever-increasing, this knowledge is often kept in silos, which neither interact with other silos nor with society at large. Scientific disciplines should not only orientate their wider focus and daily interaction more strongly towards society, but also need to reintegrate the humanities in order to enable researchers to consider the ethical conduct and consequences of their research. Ideally, responsibility as a researcher should be rooted in ethics, and directed at society.

• History of Methods

Methods in science

A great diversity of methods exists in science, and no overview that is independent from a disciplinary bias exists to date. One can get closer to this by building on the core design principles and a new ontology of of scientific methods.

Quantitative vs. qualitative The main differentiation within the methodological canon is often between quantitative and qualitative methods. This difference is often the root cause of the deep entrenchment between different disciplines and subdisciplines. Numbers do count, but they can only tell you so much. There is a clear difference between the knowledge that is created by either qualitative or qualitative methods, hence the question of better or worse is not relevant. Instead it is more important to ask which knowledge would help to create the most necessary knowledge under the given circumstances.

Inductive vs. deductive Some branches of science try to verify or falsify hypotheses, while other branches of science are open towards the knowledge being created primarily from the data. Hence the difference between a method that derives theory from data, or one that tests a theory with data, is often exclusive to specific branches of science. To this end, out of the larger availability of data and the already existing knowledge we built on so far, there is a third way called abductive reasoning. This approach links the strengths of both induction and deduction and is certainly much closer to the way how much of modern research is actually conducted.

Scales Certain scientific methods can transcend spatial and temporal scales, while others are rather exclusive to a specific partial or temporal scale. While again this does not make one method better than another, it is certainly relevant since certain disciplines almost focus exclusively on specific parts of scales. For instance, psychology or population ecology are mostly preoccupied with the individual, while macro-economics widely work on a global scale. Regarding time there is an ever increasing wealth of past information, and a growing interest in knowledge about the future. This presents a shift from a time when most research focused on the presence.

• Ontology

Critical Theory & Bias

Critical theory The rise of empiricism and many other developments of society created critical theory, which questioned the scientific paradigm, the governance of systems as well as democracy, and ultimately the norms and truths not only of society, but more importantly of science. This paved the road to a new form of scientific practice, which can be deeply normative, and ultimately even transformative.

The pragmatism of bias Critical theory raised the alarm to question empirical inquiry, leading to an emerging recognition of bias across many different branches of science. With a bias being, broadly speaking, a tendency for or against a specific construct (cultural group, social group etc.), various different forms of bias may flaw our recognition, analysis or interpretation, and many forms of bias are often deeply contextual, highlighting the presence or dominance of constructed groups or knowledge.

Limitations in science Rooted in critical theory, and with a clear recognition of bias, science(s) need to transform into a reflexive, inclusive and solution-oriented domain that creates knowledge jointly with and in service of society. The current scientific paradigms are hence strongly questioned, reflecting the need for new societal paradigms.

• ADD Bias entries

Experiment & Hypothesis

The scientific method? The testing of a hypothesis was a breakthrough in scientific thinking. Another breakthrough came with Francis Bacon who proclaimed the importance of observation to derive conclusions. This paved the road for repeated observation under conditions of manipulation, which in consequence led to carefully planned systematic methodological designs.

Forming hypotheses Often based on previous knowledge or distinct higher laws or assumptions, the formulation of hypotheses became an important step towards systematic inquiry and carefully designed experiments that still constitute the baseline of modern medicine, psychology, ecology and many other fields. Understanding the formulation of hypotheses and how they can be falsified or confirmed is central for large parts of science. Hypotheses can be tested, are ideally parsimonious - thus build an existing knowledge - and the results should be reproducible and transferable.

Limitations of hypothesis These criteria of hypotheses showcase that despite allowing for a systematic and - some would say - ‘causal’ form of knowledge, hypotheses are rigid at best, and offer a rather static worldview at their worst. Theories explored in hypothesis testing should be able to match the structures of experiments. Therefore, the underlying data is constructed, which limits the possibilities of this knowledge production approach.

• ADD Experiment entries

Causality & Correlation

Defining causality Building on the criteria from David Hume, we define causality through temporal links ("if this - then this"), as well as through similarities and dissimilarities. If A and B cause C, then there must be some characteristic that makes A and B similar, and this similarity causes C. If A causes C, but B does not cause C, then there must be a dissimilarity between A and B. Causal links can be clearly defined, and it is our responsibility as scientists to build on this understanding, and understand its limitations. Understanding correlations

Correlations statistically test the relation between two continuous variables. A relation that - following probability - is not a coincidence but from a statistical standpoint meaningful, can be called a significant correlation.

The difference between causality and correlation With increasing statistical analysis being conducted, we sometimes may find significant correlations that are non-causal. Disentangling causal form correlative relations is deeply normative and needs to acknowledge that we often build science based on deeply constructed ontologies.

• ADD Causality & Correlation entries

Emergence of agency

Agency, complexity and emergence System thinking gained prominence during the last century, allowing to take interactions and interdependencies into account when investigating a system. To this end, a ‘system’ stands for something that is in the focus of the investigation, such as a catchment area, a business enterprise, a social institution, or a population of wasps. Such systems show signs of complexity, which means that the interactions of smaller entities in the system may be unpredictable or chaotic, which makes systems more than the sum of their parts. Under this definition, solving problems within a system is often the greatest challenge: while the system dynamics may be understandable, the solutions for an emerging problem are not instantly available. Consequently, solutions for complex problems emerge, and this demands a new line of thinking about systems.

• ADD TD entry

Data and methods

A new age of data The increasing availability of data offers many new possibilities, and the emergence of new ways of getting data through the internet poses not only opportunities, but also - among others - ethical challenges. All the while, ever more diversity of data becomes available to people, leading to an even larger wealth of qualitative data. Again, this poses many ethical questions, and imposes an all new agenda onto many methodological approaches, including data security, picture rights, normative interpretations and even culture wars.

The limitations of technology While technology has surely changed the way we live our lives, it has certainly also changed the way we do science. We should however recognise that technology cannot produce knowledge, and therefore can only be a means to an end, but not an end in itself.

A way forward for methods Methods need to acknowledge the increasing diversity but also the new challenges that emerge from the exponential growth of science. Interactions between different disciplines are strongly increasing, and the core goal for science will be how to facilitate this through suitable communication and additional resources.

• ADD statistics entries

Mixed Methods

Why combine methods? The new challenges we face demand the production of new knowledge to approximate solutions. Scientific collaboration is therefore necessary in order to open up domains that were previously sealed and oftentimes arrogant.

How to combine methods Methods are often characterized by a specific language, which is why a lot of time needs to be invested into understanding each other. In addition, experts in one method are often deeply invested in their specific focus, which is why interdisciplinary collaboration is mainly built on trust.

How not to combine methods Methods are not like cooking recipes that anyone with the right recipe can unlock. Instead, methods evolve, just as the harmonisation of methods evolves. A mixed methods approach is thus not like a mere recipe, but should be approached as something new, and only time will tell how we may become more experienced and systematic in the combination of different methods.

• ADD triangulation entry

Methods of transdisciplinary research

Shifting paradigms in science For a long time, scientific disciplines existed alongside each other and separated from the rest of society, using the world as an object of inquiry and gathering data all the while. However, the state of our world does not allow for such separation anymore. We need a new collaboration within science and between science and society in order to find solutions to urgent challenges.

A new contract between science and society Transdisciplinary research is built on the premise of scientific and non-scientific actors from diverse backgrounds jointly framing, understanding and solving societally relevant problems. Transdisciplinarity is reflexive, participatory and holistic and able to develop mutually beneficial approaches to complex and normative issues.

Learning from and with each other Despite the rather recent emergence of transdisciplinary research, a lot has been developed in terms of appropriate research modes, methods and surroundings. Examples of this are Visioning, Scenario Planning and Living Labs. More is yet to come, but the foundations have been set.

• ADD TD entry

---