Causal Inference Flashcards
___ ___ is the process where we use statistical methods to characterize associations between variables involved.
Statistical association
Statistical association could either be ___ or ___.
Positive or negative
In statistical association, we are proving the statistical d___ between the two v___.
- Dependence
- Variables
After statistical association, we will determine the c___ ___.
Causal relationship
___ ___ is the process of ascribing causal relationships to associations between variables.
Causal inference
What are the two types of association?
Causal and non-causal
Under causal association, we also have d___ and i___ causation.
Direct and indirect causation
A ___ is a factor that will play a potential role in producing an outcome.
Cause
___ is an identifiable relationship between exposure and disease.
Association
___ is presence of mechanism that leads from exposure to disease.
Cause
___ is an alteration in the frequency or quality of one event followed by a change in the other.
Causal
___ is when an association is a result of the relationship of both factor and disease with a third variable.
Non-causal
In ___ causation, the alteration in factor A is directly related to the change in factor B.
Direct causation
In ___ causation, there is another factor associated with the change of the outcome.
Indirect causation
What is the first step in the process of causal inference?
Determine the validity of association
By determining the ___ of the association, you rule out chance, bias, and confounding as explanation of observed association.
Validity
Step 2: Detemine if the observed association is ___.
Causal
In step 2, you consider the ___ of evidence taken from a number of sources.
Totality
What are the two types of validity?
Internal and external validity
___ validity is the validity within the study.
Internal validity
___ validity is the validity beyond the study.
External validity
In IV, estimate of effect measure is ___.
Accurate
In EV, estimate of effect measure is generalizable to b___ ___.
Generalizable to bigger population.
IV is not due to ___ error.
Systematic
EV is not due to ___ error.
Random
The goal of Epidemiologic studies is to estimate the ___ of the p___ with l___ ___.
- Value
- Parameter
- Little error
Random errors come from ___ errors.
Sampling errors
The ___ ___ is the difference between population value and estimate value.
Random errors
___ errors are biases and confounding.
Systematic errors
Systematic errors are the ___ in the estimation of the magnitude of association between E and D.
Distortion
What are the three types of biases in systematic errors?
- Selection
- Information
- Confounding
___ bias is choosing non-representative sample.
Selection
___ bias is inaccurate information from sample.
Information
In Information bias, the misclassification has d___ and n___.
Differential and non-differential
Non-differential (random) occurs when errors are in ___ proportion in groups being compared.
Similar
Differential (non-random) occurs when rates of errors ___ in the groups being compared.
Differ
What are the sources of misclassification?
- Instrument
- Subjects
- Observers
___ is mixing the effect of exposure on the disease with that of the third factor.
Confounding
The third variable that is related to the exposure and outcome is called the ___.
Confounder
The confounder is also known as a ___ ___ in the development of disease.
Risk factor
When a confounder is present, the ___ can’t be the cause of the outcome.
Exposure
What are the methods of controlling confounding?
- Design
- Analysis
R___ is to distribute confounders r___ between s___ ___.
- Randomization
- Randomly
- Study groups
R___ is to restrict the entry to study of ___.
- Restriction
- Confounders
M___ aims for e___ ___ of confounders.
- Matching
- Equal distribution
In S___ analysis, confounders are distributed evenly in each ___.
- Stratified
- Stratum
M___ analysis requires many statistical tests to come up with good analysis.
Multivariate
When determining if observed association is causal, we use ___ ___.
Hill’s criteria
In ___ ___ ___, the higher the risk ratio, the more likely it is to be causal.
Strength of association
___ is the temporal relationship; the exposure preceeds the disease.
Temporality
___ is about the consistent findings across designs, populations, investigators.
Consistency
___ ___ does not contradict the natural history or biology of disease.
Theoretical plausibility
