Critical thinking about causality Flashcards
What four premises guide the purpose and organisation of this article regarding causal inference and developmental psychology?
- causal inference is essential to accomplishing the goals of developmental psychologists
- In many analyses psychologists are attempting causality but doing so badly
- Assumptions should be identified explicitly and checked empirically and conceptually
- Once introduced to the broader issues, developmental psychologists will recognise the central importance of causal inference and embrace the methods
What is often the reason for psychologists doing causality so poorly?
often employ implicit or even implausible assumptions. These assumptions are both statistical (such as linearity) and conceptual
(such as no unobserved confounding.)
What are missing in the methods of causal inference in developmental psychology?
An overarching framework in which to embed comparisons of alternative approaches
What is meant by the ignorability assumption?
assumption of no unobserved confounding
When acknowledging that causal inference cannot be made in an article, what two dissatisfying and possibly misleading directions do many articles tack in?
The authors hold causal inference as unattainable, which is not practically useful. Authors can also rely on longitudinal studies tio make a leap from correlational to causal. Unfortunately, they often apply tools that have limitations for performing causal inference, such as linear regression, and make implausible assumptions about the nature of the association of interest
Give 3 reasons as to why causal inference is needed
- to improve life causal inference is needed
- causal thinking is unavoidable
- Even if scientists can distinguish causal and association results, many lay people cannot
Is the observed effect or association usually stronger? Explain
When one manipulates one variable in order to influence another, the (causal) impact on the effect will often be smaller than the association observed in data. What has happened is that part of that relationship was explained by other, third factors (confounders or omitted variables) that were unaffected by the manipulation.
What two conceptual tools are especially helpful in moving from associations to causal relationships?
- Directed Acyclic graph (DAG)
- Potential outcomes framework
What requirement of moving from association to causality is DAG particularly useful with?
Ruling out potential confounders and covariates
What is meant by the causal markov assumption?
The absence of a path implies the absence of a relationship
What requirements are there for a dag, as suggested by the name?
All paths must have directions, the effects should not be cyclical
Should a variable be placed in a DAG if it is unobserved, unmeasured or otherwise unavailable?
Yuh
How is a DAG structurally stable?
An intervention on one component of the model does not alter the broader structure. Intervening on a variable may change how it relates to other variables statistically
What other two preferences does the DAG assume?
The DAG also assumes a preference for simplicity and probabilistic stability. Simplicity means that models that represent data with fewer linkages are preferred to the more complex. Stability refers to the robustness of a set of relationships across a range of possible magnitudes.
What three traditional criteria must be satisfied in order to imply causality?
Priority: Change x precedes change y
Consistency: Change x varies systematically with change y
Exclusivity: no alternative explanation for relationship
What problems arise with this criteria?
Priority can can post hoc ergo propter hoc reasoning errors
Two variables can be consistently correlated but no causal effect
One variable can cause another but this mightn’t be the only with the dependent variable occurs (exclusivity)
