Module 6 Flashcards
Definition of variable
Any attribute that describes a person, place, phenomenon, event or idea. Varies in value.
What are examples of a variable?
Age
Height
Weight
Income
Education
Sex
Race
Dz status
Diet
Anything that can be studied.
Definition of exposure
An independent variable that might cause or prevent a dz or death, a possible determinant.
Definition of risk factor
Any exposure that increases a person’s chance of developing a dz or dying.
Definition of protective factor
Any exposure that decreases a person’s chance of developing dz or of dying
Definition of outcome
A dependent variable that is the effect or consequence of the risk and/or protective factor(s). In epi, this is usually the dz of interest or death.
Nominal variable
A type of variable that consists of unordered categories or classes.
Examples of nominal variables
Sex
Race/ethnicity,
Sexual orientation
Marital status
Any variable that can have a yes or no value
Ordinal variable
Variables with ordered categories but with no consistent distance between them
Examples of ordinal variables
Stages of cancer
Levels of education
Values on a Liekert scale (excellent, good, fair poor)
Discrete variables
Integers that differ by fixed amounts, with no intermediate values possible
Examples of discrete variables
Number of children a parent has
Number of times tails comes up when a coin is flipped
Continuous variables
Measurable quantities, not restricted to integer values, with an infinite number of intermediate values. These variables can take on any values between their minimum and maximum values.
Examples of continuous variables
Time
Height
Weight
Cholesterol
BP level
Study design definition
The set of methods and procedures used by the researchers in collecting and analyzing data to test or generate hypotheses.
What are the two major types of study designs in epi?
Experimental
Observational
Components of observational studies
Investigator does not have control over the exposure factor
Investigator is unable to assign participants randomly to study conditions
Investigator simply “observes” and lets nature take it course
Components of experimental studies
Investigator controls who is exposed to a factor of interest and how much they are exposed to
Investigator assigns participants randomly to study groups
Investigator intervenes
What can ecologic studies also go by?
Aggregate and correlate studies
Is an ecologic study an observational or experimental study?
Observational
What is the unit of analysis in an ecologic study?
Group. This means we compare averages or rates among a sample of groups or units of aggregated data
What correlations are obtained in ecologic studies?
Between exposure rates and dz rates among different groups or populations
Definition of unit of observation
The level on which data are collected. It is the type of unit you are gathering data about. It is the smallest unit on which the investigators have data.
Unit of analysis definition
The level on which data are analyzed and conclusions made.
Examples of units of analysis that are groups
A country
A state
A county
A city
A neighborhood
A school
Components of a descriptive ecologic study
May have only an outcome or an exposure variable, but not both.
Looks for patterns in the distribution of averages or rates
Person, place, or time with no other exposure.
No “a priori” hypothesis
Components of an analytic ecologic study
Has an outcome and an exposure variable
Tests an a priori hypothesis that a specific exposure helps to cause or prevent a specific dz
What are the three main types of ecologic studies?
Ecologic comparison study aka multigroup study
Ecologic trend study aka time trend
Mixed ecologic study
What does an ecologic comparison/multi-group study do?
Examines exposure rates and dz rates among different groups over the same time period
What does a time trend study/ecologic trend study do?
Examines changes in exposure and changes in dz within the same community, country, or geographic area over time. Units that are compared are time periods.
What does a mixed ecologic study do?
Combines aspects of both the ecologic comparison and ecologic trend studies
Observes the relationship between the average change in dz incidence in several groups over time
Examples of analysis techniques in ecological studies
Usually continuous data
Scatter plots
Correlation coefficients
Linear regression
Which axis is associated with what component of the study?
X is exposure and Y is outcome
Definition of correlation
A measure of linear association between two ordered lists, or two continuous variables
Definition of correlation coefficient
A measure of how nearly a scatter plot falls on a straight line. Always between -1 and 1.
When is a correlation positive?
When the coefficient is from 0 to +1.
When the coefficient is exactly +1, this is a perfect positive; as one variable goes up, so does the other.
The closer it gets to 0, the weaker it is. The closer it gets to +1, the stronger the positive association is.
What does a correlation coefficient of 0 indicate?
No association or relationship between the two variables
When is a correlation negative or inverse?
When the coefficient is from 0 to -1.
A correlation coefficient of -1 is a perfect negative: as one variable goes up, the other comes down.
The closer it is to 0, the weaker it is. The closer it is to -1, the stronger the inverse it is.
Advantage of ecologic studies
Quick
Inexpensive
Utilize data that is already available
Good approach for generating hypotheses when a dz is of unknown etiology
Useful whenever it is desirable to make inferences about the group, and not the individual, such as in evaluating a population intervention or a policy
More appropriate than observational studies of individuals when intra-group (within the group) variability is small but inter-group variability is large.
For which exposures are ecologic studies best?
Social and environmental exposures
What are disadvantages of ecologic studies?
Imprecise measurement of exposure and outcome
Difficult to adjust for other factors in the study
Ecologic fallacy
Ecologic fallacy definition
When incorrect inferences about the individual are made from group level data.
