Statistical modelling Flashcards
4 different paths + possible variables for relationship of education on income
Why is the third one correct?
causal effect = hard to get, u enter a hypothetical world –> what if I gave you more education?
Why confounding?
It could be sth else then education, it could be intelligence –> messes with causality
Why first and not most popular one?
Purely observational data, I cannot just give you more education, you take it –> Why do you take it and others not? That’s the problem, you self select bc of certain characteristics
Influence/Variance not a problem as long as there is no confounding, we just want the causal effect we don’t have to explain everything, very often even harmful
Variables that could confound the relation between education & income?
Should we control for all those variables?
No –> kitchen sink appraoch leads to big measurement errors =The more complex a model, the more likely that any measurement error will compound, and cause the outputs to be wildly inaccurate (cross-sectional data only takes you so far when it comes to causal effects)
Solution by design - finding randomness to emulate experiments
Solution by design - finding randomness to emulate experiments II
Solution by design - finding randomness to emulate experiments III
rid the estimates of genetical confounding
Statistical model for: Is education achieved or inhereted?
c) can help us to compare i.e. different structures in different countries
d) how much education should we give? is there over-education?
e) i.e. when is the right tracking age (later tracking weakens that link), G9, etc.
Nuance - a virtue?
How do observations versus experiments relate to causal effects?
What is the strict view on explanatory variables as the potential “causes”?
experimentally manipulable explanatory variable, at least in principle, particularly sticky point because, in social science, many explanatory variables are intrinsically not subject to direct manipulation, even in principle
–> i.e. gender cannot be considered a cause of income, even if it can be shown (perhaps after controlling for other determinants of income) that men and women systematically differ in their incomes, because an individual’s gender cannot be changed.
