College aantekeningen
ARM lectures and assignments
- Instelling
- Radboud Universiteit Nijmegen (RU)
all lectures without endogeneity and process tracing. All assignments with extra theory
[Meer zien]Voorbeeld 4 van de 46 pagina's
Voorbeeld 4 van de 46 pagina's
In winkelwagengesponsord bericht van onze partner
Verkoper
Volgen
Voorbeeld van de inhoud
ARM lecture 1: Introduction and causal mechanismsCritically judge the quality of analysis in articles, especially when it comes to causality ( X Y).
Make your own research methodologically sound.
- Theory & inference
Theory = Logically interrelated propositions. A systematic explanation. System of ideas. About
empirical reality, about something that is happening. Trying to explain general principles. Saying
something general about empirical reality.
Inference: the goal of social science is to infer beyond the immediate data to something broader that
is not directly observed (generalisations). Can have qualitative and quantitative inference.
Inductive and deductive inference:
- Start from observation to theory (inductive). Theory generating
- Theory testing (deductive). Test theories with observations.
We need both and we need a descriptive set. Can describe the social world or try to explain it (causal
questions).
Descriptive and causal inference:
- Focus on explanation (XY)
- But: description necessary first step (show what is happening before they attempt to explain
why it is happening).
Causation as robust dependence
- While correlation does not imply causation, causation must in some way or other imply
association. (Covariance, association and causation implies the same thing. x becomes higher
or lower, y becomes higher or lower). Why values of y are dependent on values of x.
- Robust: correlation remains after controlling for possible other explanations (confounders)?
Does initial correlation of what remains of it robust to inclusion of confounding variables. Is
relationship of interest (XY) dependent of other variables.
Income trust
Education
Spurious association or Omitted variable bias. Effect is caused by an omitted variable. Find an effect if
we do not control for education, but the effect is actually caused by education and when controlling
for it the initial causal relationship of interest disappears.
,Critique 1:
- Never no if you controlled enough. No strict causal test; confounders always possible.
Alternative is consequential manipulation (experimental logic).
- Based on counterfactual reasoning. Can only be sure of causal effect if we observed y in the
presence or absence of x. Everything is identical except your variable X, no confounding
variables. Achieved by random assignment. If randomization succeeds, I have approached the
counterfactual.
- Effects of causes: causes must be manipulable. But what to do with structural characteristics
that are not easily manipulated? Cannot manipulate everything. Also not ethical with human
beings.
- Problem of agency: humans have agency. In an experiment only the researcher has agency,
subject do not. But this is not how it works in the real world.
Critique 2: what is the mechanism? Why is X connected to Y? alternative: causation as generative
process (mechanism-based approach).
X----------Y
Problem 1: Some relationships are impossible or very hard to study with regression models. Need to
specify what variables are control variables or intervening variables. Can be with control variables
that some effect you thought was causal is not causal when controlling for it. You need to specify the
causal mechanism. Cannot put blindly controls in your model, need to know how it effects the causal
mechanism.
Problem 2: multiple paths cancel out. Nul effect (no effect between x and y ) can be how multiple
paths cancel each other out. Need to specify the causal mechanism.
Causal mechanisms:
A causal mechanism is a theoretical account that
- Shows how specific outcomes or empirical regularities come about.
- Shows the structure (cogs and wheels) of the causal process.
- Logical connection between x and y.
- Shows causal pathways.
Can should we empirically study causal mechanisms?
- At minimum: theoretical specification
- If possible: empirically observable implications.
Theory informs what we think is happening in the real world. Need to make causal process we think is
happening in the real world apparent by observations and theorizing about what the implications are.
,Mechanism require some form of methodological individualism. Approach to causality:
- Social phenomena should be explained as resulting from the behavior of individual actors
- These behaviours should be explained as motivated by intentional states of the actors (i.e.,
there should be some theory of action)
The strong versions treat social phenomena as either:
- Irrelevant
- Or strictly endogenous
o Social phenomena are strictly understood as the intended and unintended
consequences of actions of individual actors
o They are only the explanandum, not the explanans
- Very economic, rational view. Reductionism, reduces everything to individuals. No society or
structure (macro-level structures).
Weak version treat social phenomena as causal influence, constrained by norms and institutions
(social structures) as:
- Of fundamental
Structure plays a role in individual actions, and individual actions can collectively influence structure.
There are structures and agents within the structures. Theory of action is always micro-level. But the
individuals or actors can also be organisations or governments. Colemans bathtub is a structural
individualism methodological approach.
In any causal process there are 3 causal mechanisms in Colemans Bathtub:
- Situational mechanisms. How people can act is constrained by the context. (macro to micro)
- Action-formation mechanisms (rational choice): if people have certain believes, attitudes,
plans, social situations, economic situations than how do individuals act (micro, individual
level).
- Transformational mechanisms: how these individual acts influence the macro. (micro to
macro).
If you study macro phenomena, how does micro phenomena influence these? There are always micro
foundations for macro-relations.
This way of explanation requires an account of how actors act:
- Theory of action
But does not stipulate any specific theory of action. Could be:
- A rational choice theory
- A social identity theory
- A behavioural theory
- Etc.
, Putnam ethnic conflict theory for thesis?
ARM Lecture 2
Conspiracy theorists like macro-level correlations. Do not take into account confounding variables.
Misses micro level foundations (not going down to individual levels, no micro-level explanations).
Ecological fallacy Robinson, make micro-level statement based on macro-level data.
Do not know if same individuals have same properties we observe in aggregated data.
Aggregated (ecological) data: data that is aggregated over a group of individuals, characteristics of
individuals, but only have 1 number describing the whole group. But in essence this is about
individual characteristics. People make statements with aggregated data to make statements about
what individuals do or what characteristics of individuals are.
Descriptive properties of individuals vs. descriptive properties of groups.
Ecological fallacy: misses information about the individual level. You would have to observe each
individual.
Marginal frequencies vs. cell frequencies. You only have proportions only based on aggregated data.
Argument of Robinson, you need cell frequencies and only those cells can tell you about the micro-
level. Aggregated data can change with individual properties, but aggregated data stays the same.
Aggregated correlation is not the same as individual correlation, but you can say some probabilistic
statements with aggregated data. Can show some interesting patterns which can influence more
research. More migrants in very right-wing voting areas. Migrants vote right-wing or is it a reaction?
Aggregated data can show interesting patterns that, with the influence of theory, can be the basis for
more research and causal mechanisms.
If you just look at aggregated data, you can say for example that in areas with more black people
there is more illiteracy. But just the numbers of people and illiteracy does not show you who is
actually illiterate. It could be that in these areas black people are actually literate and more white
people are illiterate, but just from the above statement you could make different assumptions.
Look at assignment 1.
- Causal inference at macro and micro level
Often a trade-off between theoretical appropriateness and data-availability
Problems with macro data:
- Small n
- Potential misfit theory-data
Voordelen van het kopen van samenvattingen bij Stuvia op een rij:
Verzekerd van kwaliteit door reviews
Stuvia-klanten hebben meer dan 700.000 samenvattingen beoordeeld. Zo weet je zeker dat je de beste documenten koopt!
Snel en makkelijk kopen
Je betaalt supersnel en eenmalig met iDeal, creditcard of Stuvia-tegoed voor de samenvatting. Zonder lidmaatschap.
Focus op de essentie
Samenvattingen worden geschreven voor en door anderen. Daarom zijn de samenvattingen altijd betrouwbaar en actueel. Zo kom je snel tot de kern!
Veelgestelde vragen
Wat krijg ik als ik dit document koop?
Je krijgt een PDF, die direct beschikbaar is na je aankoop. Het gekochte document is altijd, overal en oneindig toegankelijk via je profiel.
Tevredenheidsgarantie: hoe werkt dat?
Onze tevredenheidsgarantie zorgt ervoor dat je altijd een studiedocument vindt dat goed bij je past. Je vult een formulier in en onze klantenservice regelt de rest.
Van wie koop ik deze samenvatting?
Stuvia is een marktplaats, je koop dit document dus niet van ons, maar van verkoper stanseerden. Stuvia faciliteert de betaling aan de verkoper.
Zit ik meteen vast aan een abonnement?
Nee, je koopt alleen deze samenvatting voor €6,63. Je zit daarna nergens aan vast.
Is Stuvia te vertrouwen?
4,6 sterren op Google & Trustpilot (+1000 reviews)
Afgelopen 30 dagen zijn er 79223 samenvattingen verkocht
Opgericht in 2010, al 14 jaar dé plek om samenvattingen te kopen