An overview of the key concepts/topics from the ARM masters course of Political Science. The concepts are in alphabetical order and key information/ a summary is stated per concept based on the lectures and articles.
Model without explanatory variables, only Y So the current value of Y is predicted by the past
value of Y multiplied by B1 estimate of B1 shows you the dynamics of Y, how the value of Y
depends on past values of Y (Yt=B0 + B1*Yt-1 + Et)
Solves the serial correlation problem
Can strengthen causal inferences there might be an effect of Y on X instead of the other way
around, so the previous Y can have an effect on X (X is government spending, Y is support. Example:
change in support Yt-1 can influence the government spending X in the next year). So yt-1 related to
both Xt and Yt. This strengthens your inference with this model you control for Yt-1 so that you
can really see if X still has a significant influence on Y
Extending the model: you can also include a second lag beside the first lag + adding X variables
Lecture 6
AR (autoregressive) model (=lagged dependent variable model) solves the problem of serial
correlation
Three reasons: control for serial correlation, theoretical expectations, deal with reverse causality
Causal heterogeneity
Lecture 3
Effect of micro-level X variable (slope) identical across groups (slope = effect of X on Y, so the
direction/shape of the line is etc.)
But this might not always be the case = causal heterogeneity the effect of micro level variable X
can vary across groups
Can be modeled by including random coefficients (and estimating their variance)
Gives us the mean effect X on Y across groups
Allows us to establish the variability of the effect of X on Y across groups (variance of
the coefficient, so variance of the slope)
We can even try to explain the variance of the coefficient with a macro-level X variable
Cross-level interactions
1
,Heterogeneity= the effect of X is different across groups (because of a certain influence/variable)
Variance of a certain variable is the variance of the effect that that has across groups (so you have
individual variance, group variance and this variable variance)
Causal heterogeneity thus refers to the effect of a micro-level variable varying across macro-level
contexts.
Cross-level interactions are a way to explain such causal heterogeneity
Causality/causal mechanisms
Lecture 2
Criteria:
Covariation
Time order (so first applying sth and then seeing what happens)
No alternative explanations
Mechanism (what is between X and Y, how does it work?)
Causation must imply association, so correlation/covariance/association is a necessary condition for
causation
Causal mechanism = Theoretical account that shows how specific outcomes/events or empirical
regularities come about; shows the structure of the causal process; shows the logical connection
between x and y; is an answer to the ‘why’ question pathway through which X might affect Y
Mechanisms require methodological individualism (all three authors argue this) statistical
techniques only show relation, not how these are produced through the action of individuals
Social phenomena should be explained as resulting from the behavior of individual actors
Goldthorpe
Approaches to causation
Causation as robust dependence
X is a ‘genuine’ cause of Y in so far as the dependence of Y on X can be shown to be
robust = cannot be eliminated through one or more other variables being introduced
Granger causes: causation in terms of predictive power: causation of x on y if the
values of x still add to the ability to predict y when taking into account all other
information.
Causation as robust dependence has appropriateness when the main aim of research
is prediction, especially prediction in the real world rather than a lab.
Causation as consequential manipulation
Causes can only be those factors that could serve as ‘treatments’ in experiments:
causes must in some sense be manipulable.
2
, Genuine causation if a causal factor X is manipulated and then (given appropriate
controls) a systematic effect is produced on the response variable Y.
Causation as consequential manipulation is concerned with the effects of causes (Y is
an effect of X) while causation as robust dependence is concerned with the causes of
effects (X is a cause of Y).
The idea of causation as consequential manipulation fits to research that is
undertaken through experimental methods where the central concern is the
consequences of performing particular acts.
Causation as generative process
Tying the concept of causation to some process existing in time and space, that
actually generates the causal effect of X on Y and in doing so produces the statistical
relationship that is in evidence.
This approach is used to complement and/or correct the previous two approaches
Causation as generative process is the best basis. It focuses on the causes of effects,
analysis begins with the effects for which causal explanation is sought. It applies to a
deeper more microscopic level. It is not expected to achieve once and for all
verification (it is falsification)
When the ultimate aim is causal explanation, it cannot be arrived through statistical methodology
alone: a subject-matter input is required in the form of background knowledge and theory.
Gerring
Challenges of testing causal mechanisms
The difficulty of testing empirical mechanisms begins with the difficulty of articulating all the
possible (theoretically plausible) causal mechanisms for a given X–Y relationship.
The problem of omitted factors deserves special notice. If no list is ever demonstrably
complete (there are always novel and heretofore unimagined ways in which X may influence
a distal Y), then no testing procedure is definitive.
A third challenge is that mechanisms under review in social science frequently consist of
vague and abstract concepts that are rather difficult to operationalize.
A fourth challenge is that plausible mechanisms connecting X and a distal Y are usually
multiple and may be difficult to tease apart from one another.
A fifth challenge is the variety of possible interrelationships that may be found among
various causal mechanisms that help to explain X’s relationship to Y.
A final challenge is that the same causal mechanism may have opposite effects on an
outcome depending on the trigger or the context of an event.
What we need is intelligent discussion of plausible causal mechanisms, which should be subjected to
empirical testing to the extent that it is feasible.
Lecture 11
(Process tracing)
Understanding of causation: mechanistic (mechanism that leads from X to Y, elements that produce
an outcome), asymmetric (presence of X has consequences), deterministic
3
, Causal mechanism = a model of (underlying , unobserved)processes that bind causes and outcomes
together
Causal asymmetry: implications of presence/absence of X
Ways to prove (=substantiate) causal mechanisms:
1. interventions: change in X, see what it does to Y
2. natural experiments : wait for something to happen in real world to see the effect on Y
3. hypothetical counterfactuals: what would have happened if…
Causal mechanisms: three main elements
1. entities: actors who engage in activities
2. activities: producers of change
3. observable/empirical manifestations: observable elements of activities
4. arrows: logical connections between the elements, the causal reasoning that binds them
together 5 context: context in which the causal mechanisms work temporal
dimension (when does it work), spatial dimension (where does it work), social dimension
(in what social settings does it work)
we find causal mechanisms at meso level, we study them on micro level (so meso level theorizing,
micro level empirical tracing)
Coleman’s bathtub
Lecture 2
Macro: country level phenomenon (ex. amount of seat of a party)
Micro: individual level phenomenon but also other types of actors such as organizations this level
is where the agency is
4
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 daphnehermsen. Stuvia faciliteert de betaling aan de verkoper.
Zit ik meteen vast aan een abonnement?
Nee, je koopt alleen deze samenvatting voor €5,49. Je zit daarna nergens aan vast.
