Causation and Casual Inference Flashcards
Something that brings about an effect or a result
Cause
Two groups of scientists who look at the factors that influence outcomes
Biomedical scientists and epidemiologists
A ______ is an association between categories of events or characteristics in which an alteration in the frequency and quality of a factor is followed by a change in the other
causal association
Cause must precede effect
Temporality
A change in the cause results in a change in the effect
Directionality
Cause must precede effect
Temporality
The presence of _____ may result in association with no causal link
confounders
What are the steps in the process of casual inference?
- Determine the Validity of the Association
- Determine if the Observed Association is Causal
The research is ____ if the estimate of the effect measure is accurate
internally valid
Internal validity is not due to _____
systematic error
What precedes external validity?
Internal validity
External validity is not due to ____
random error (chanz <3)
First step in establishing the causation of a disease
Descriptive study
Cause-Effect: Causal Association
Effect-Cause: _____
Temporal Ambiguity
Temporal ambiguity is a problem in retrospective studies, and more so in ______
cross-sectional studies
Threats to validity
- Confounding
- Bias
- Chance
A systematic error that results in an invalid or incorrect estimate of the measure of association
Bias
Systematic error can be due to the researcher since it can arise from either from the ______ and/or
______
poor design and/or conduct of the study
Selection Bias: Non-representative sample
Information Bias: Due to ____
inaccurate information
Bias towards null: _____
Bias farther from the null: _____
Underestimation
Overestimation
A distortion in a statistical estimate resulting from procedures used to select subjects and factors that influence participants in the study
Selection bias
Sources of Selection Bias
- Attrition
- Response
- Self-selection
- Use of inappropriate groups in analytic studies
Attrition happens in ______ studies
Prospective
What is the strategy to control attrition?
Maintain high and similar follow-up rates for the study groups
People who volunteer in studies are normally health-conscious, therefore triggering what source of selection bias?
self-selection
What strategies may be employed to account for self-selection bias?
- For occupational research, select an appropriate comparison cohort (same occupational setting)
- Obtain high participation rates among both
cases and controls
What strategy may be employed for the use of inappropriate groups in analytical studies?
Use identical selection criteria for both cases and controls
Selection bias in case-control
Selection of case/control related to exposure status
Selection Bias in Cohort (Retrospective Cohort)
Selection of exposed/unexposed is related to the development of the outcome
Also known as observation bias, measurement error, misclassification bias
Information bias
Sources of information Bias
- Participant (Trait Error)
- Observer/tool (Method Error)
These may happen in trait error of information bias
- Recall bias
- Hawthorne effect
- Social desirability
Strategies for avoiding information bias
- Blind/mask parties as necessary
- Train data collectors
- Triangulate sources of data
Types of Misclassification
- Non-differential
- Differential
Which type of misclassification is preferred?
Non-differential
In non-differential misclassification, bias is always _____ null
towards
Differential misclassification happens when?
methods of data collection are not uniform
Uncontrollable force that seems to have no assignable cause and distorts the true value of an estimate.
Chanz (Chance)
Chance is also called
Random error
Minimizing random error increases ____
Precision
Chance is estimated by ______
Significance level
In box plot, if the line crosses the null value of 1, then what can we infer about its statistical significance?
not statistically significant
What are Hill’s Criteria for Causation?
- Strength of Association
- Temporality
- Consistency
- Theoretical plausibility
- Coherence
- Specificity in the causes
- Biologic Gradient
- Analogy
- Experimental Evidence
