Causality Flashcards
Pros of samples
can be studied more quickly than larger populations
Hypothesis testing
based upon what is commonly believed (the status quo)
Alternative hypothesis (H1)
statement that contradicts the null hypothesis
6 steps for evaluating hypotheses
- Formulate the Ho in statistical terms (Ho: Odds ratio = 1)
- Formulate the H1 in statistical terms (H1: Odds Ratio = 1).
- Select the level of significance for the statistical test and sample size (generally 0.05 or conservatively 0.01)
- Select the appropriate test stastic and identify the degrees of freedom and the critical value
- Collect the data and estimate the measur eof association and test statistic
- If the observed measure exceeds the critical value, reject Ho in favor of H1; otherwise do not reject Ho
The null hypothesis is assumed to be correct unless
there is sufficient evidence from the sample data to indicate otherwise
A tentative suggestion that a cerain phenomenon exists
research hypothesis
P value
the probability for evaluating the role of chance; ranges from 0 to 1
Small p value (> or = 0.05)
the result is unlikely to be a product of chance
Confidence intervals
range of reasonable values in which a population parameter lies that is based on a random sample from the population
When sample size increases
the role of chance decreases
To minimize chance
increase sample size
Bias
deviation from the truth
Random error
incorrect result due to chance
Systematic error
incorret result due to bias
Confounding
a third factor that influences the relationship between an expoure and disease outcome
Determinants
causes/factors
Epidemiology
the study of the distribution and determinants of health-related states or events in human populations and the application of this study to the prevention and control of health problems
Identifying what determines (cause) disease allows us to
prevent and control the health problem rather than react to it
Identifying what determines (cause) disease allows us to
prevent and contorl the health problem, rather than react to it
Cause
a variable that precedes a health outcome and is necessary for its occurrence
When is a causative factor “necessary”
if an environmental exposure is required for the outcome to occur
Epidemiology Triangle components
Host, Infectious Agent, Environment, Time
Epidemiology Triangle
a model that characterizes infectious disease causation showing the interaction between agent, host, environment, and time
Agent
causative factors such as a pathogen or chemical
Host
organism and usually a human
Epidemiology Triangle components
Host, Infectious Agent, Environment, Time
Environment
physical, biological, chemical, social, cultural
Each component of the Epidemiology Triangle has
time-related issues
Rothman’s Causal Pies (1976)
Explain the multifactorial nature of causation for many noninfectious disease
Types of causal associations
Direct causal, indirect causal
Factors of causation
Predisposing, Enabling, Precipitating, Reinforcing
Predisposing
factors or conditions already present that cause susceptibility to a health-related state or event without actually causing it
Enabling
precedes a health-related state or event that allow it to be realized
Precipitating
factors essential to the development of a health-related state or event
Reinforcing
factors that aggravate and perpetuate a health-related state or event
Statistical association does not mean what?
a causal association
Causal guidelines
strength of association, consistency of association, temporality, biological plausibility, experimental evidence
Web of causation
graphics, pictorial, or paradigm representations of complex sets of events or conditions
P value < 0.01
99% association
Statistic steps
hypothesis, significance, sample, p-value, decide
Type 1 Error
- to reject a true null hypothesis
- false positive
- alpha
Type 2 Error
to accept a false null hypothesis
false negative
beta
Researcher’s goal
to reject a null hypothesis
A type 1 error occurs when
the reseracher incorrectly rejects a true null hypothesis (false positive)
A type 2 error occurs when
the researcher fails to reject a null hypothesis
