IE Notes (Week 2)
What is Empirical Science?
Based on the principles of Karl Popper (1902-1994).
Main Claim: All statements of an empirical science must be able to fail when confronted with experience.
Terms used must be connected to a reality we can experience.
Statements must consist of descriptions of situations which can be experienced.
Statements must be formulated in a way that they can be falsified.
Tautologies (statements which are always true) are not allowed.
'There are … black swans' statements are not allowed.
“All swans are white.”
This is falsifiable, because observing just one black swan would prove it false.
Knowledge can never be fully verified, only falsified. Hypotheses either hold up to data or they don't.
Knowledge is only temporary.
Empiricism and Pseudo-Science
Empiricism: Empirical studies lacking theoretical foundation.
Badly set up experiments and untested methods/instruments.
Just looking at the data and trying to find patterns without theoretical guidance.5
Epistemological Disputes
Realism:
Objective knowledge exists in principle.
The world can be experienced and measured in an intersubjective, verifiable way (replicability).
Statements can be made about reality.
Constructivism:
Reality is subjectively constructed (invented) and not objectively discovered.
The world is only accessible by observation, and each observation is subject to interpretation.
Generalizations based on specific observations are problematic (Allegory of the Cave).
The Social Construction of Reality (Berger & Luckmann, 1966)
Main Claim: Knowledge is not objective, but sociologically constructed through interaction.
Knowledge is created by accepting what is considered knowledge, reinforced by institutions.
Rational action is subjective and tied to the individual.
Institutions like academies of science and universities legitimize current knowledge.
Subjective perceptions of reality can be as real as physical reality in explaining social action.
It is important to ask people about their wishes, dreams, values, ideas, and goals, not just measure observable phenomena.
Max Weber and the "Werturteilsstreit" (Dispute about Value Judgments)
Max Weber: Scientific research should be neutral and free from moral value judgments.
Research question selection and use of results are not value-free.
Values themselves can be objects of research.
Opponents: Scientists cannot stay outside their object of study; science must be relevant to society.
Dispute taken up again in the 1960s ("Positivismusstreit") between Popper and Adorno/Habermas, leading to a divide among quantitative and qualitative methodologies.
Scientific Revolutions (Thomas Kuhn)
Normal science involves puzzle-solving within the current paradigm, adding new truths to old ones.
Scientific revolutions involve a revision to existing scientific belief or practice, leading to a new paradigm.
The "scientific community" and institutions are important.
Big Data and AI potentially represent the next scientific revolution.
Theory
Theories are ideas in our head, not the empirical world.
Theories are provisional and can be replaced (Popper).
Terminology: model, concept, ideas.
Typology of Theories
Observation of empirical patterns.
Development of adhoc theories.
Middle range theories (e.g., homophily).
Theories of higher complexity.
Research Design and Experimentation Terminology
NVT Approach
The NVT approach helps define and describe important elements of a research design.
N = Number: How many objects of study (1, small N, large N)? What are the units of analysis?
V = Variable: How many (1, 2, more) and which variables are used in the model, theory?
T = Time: How many measurements take place over time? (1, 2, more)
Terminology
Variables (V) denote features of the units (N) in an experiment or study.
Variables must have more than one possible value; otherwise, it's a constant.
Variables are measured at one or several points in time (T).
Additional Important Terms
(Experimental) unit (of observation): Element or amount of experimental material to which a treatment is applied.
Treatment (variable): Distinctive feature, classification, or manipulation applied to experimental units.
Dependent and independent variables: Concept to be explained (dependent)(what you measure), concepts used to explain the phenomenon (independent)(what you change).
Population, Universe: All the units under study (N).
Sample: A selection of units (n).
Internet Voting in Switzerland
Since the first trial in 2003, an increasing number of cantons in Switzerland are offering digital voting channels.
Remote voting is common; in bigger cities, over 90% are postal voters.
The impact of internet voting on the digital divide (favoring the well-educated, economically better-off citizens) is still inconclusive.
Relevant Theories for Internet Voting Study
Digital Divide: Technological options favor the better-off and well-educated, potentially increasing existing inequalities.
Technology Acceptance Model (TAM): User’s perceived usefulness and effort are major explanatory factors for technology acceptance.
Calculus of Voting: Cost of voting is a major factor affecting voter turnout; reforms aim to facilitate voting (remote voting, advance voting, internet voting).
Example: Digital Divide and Internet Voting Study
Theories: Digital Divide, Technology Acceptance Model (TAM), Calculus of Voting.
Method: Partial Least Squares Structural Equation Modelling (PLS SEM).
Data: Opinion Survey with a sample of 1,200 citizens and the relationships between variables.
$$Sex[+] = 0.13