Ecological Models - PowerPoint PPT Presentation

1 / 21
About This Presentation
Title:

Ecological Models

Description:

... norms, cultural identification, and definition of self in daily life. ... Brenner s conclusion that Americans are more likely to die as a result of recession ... – PowerPoint PPT presentation

Number of Views:30
Avg rating:3.0/5.0
Slides: 22
Provided by: benjami3
Category:

less

Transcript and Presenter's Notes

Title: Ecological Models


1
Ecological Models
2
Why Do We Care About Multiple Levels?
It helps differentiate the causes of the cases
from the causes of the causes proximate vs.
distal causes fundamental causes Consequently,
it helps better specify the total effect of
important social and economic risk
factors consider education and income consider
education and job control Therefore, it takes the
dummy out of dummy variables! Finally, as public
health practitioners , it helps us identify areas
for social action inside and outside of health
care
3
What are the Major Levels?
The Main Determinants of Health (Source
Dahlgren G. and Whitehead M (1991)
4
Defining Characteristics of Society and Health
Population Health Equity (VS. Individual
Health)
Social Action - Redress the Social Gradient in
Health IN ADDITION TO Health Care Health
Promotion
Social Context - Place Matters Geographic
Context Historical Context Cultural Context
Life Course Context
Sociobiologic Translation General Susceptibility
5

The Social Context
  • The social context weaves together the social
    conditions of a society through processes of
    social stratification, institutional
    legitimization of behaviors and behavioral norms,
    cultural identification, and definition of self
    in daily life.

Source Amick and Lavis 2000
6
Place Matters Within Any Given Household Income
Range, Neighborhood Quality Affects Health
QUALITY RESIDENTIAL NEIGHBORHOODS
CHD INCIDENCE/PER 1,000 PERSON-YEARS
Household Income
Source Diez Roux, NEJM, 12 July 2001
7

The Ecological Fallacy
  • Inferring individual-level relationships from
    relationships observed at an aggregate level.
  • This is because the correlation between ecologic
    variables is often markedly different from the
    corresponding individual correlation within the
    same population
  • Robinson, 1950, Ecological Correlations and the
    behavior of individuals. Am Soc Rev.

8
The Ecological Fallacy What is the Erroneous
Conclusion?
r1.00
100
Over 50,000
0
0
100
Vote Republican
9
Why is it the Erroneous Conclusion?
Lets Look at A Sample of 100 Voters in Each
Community
Community A
Community B
Community C
10
Why is it the Erroneous Conclusion?
Lets Combine the Community Samples
r-.33
11

The Ecological Fallacy Some Examples
  • Durkheims conclusion that protestants are more
    likely to commit suicide

Wolffs conclusion that there is a Rosetto
Effect leading to greater longevity
Brenners conclusion that Americans are more
likely to die as a result of recession
12

Is Wilkinson Committing An Ecological Fallacy?
13
Fallacies of the Ecological Fallacy
Individual-level models are more perfectly
specified than ecologic-level models
Ecological correlations are always substitutes
for individual-level correlations
Group-level variables do not cause disease
Source Schwartz, AJPH, 1994, p 819
14
The Individualistic Fallacy of Atomist Fallacy
Drawing conclusions about groups when only
individual-level data exist
To count the percentage of individuals who agree
with particular statements on trusting
neighbors and to take this as an indicator of
social capital is to commit the individualistic
fallacy. An area can have very weak social
capital, even if people believe they can trust
one another
To count the percentage of individuals who agree
with particular statements on democracy and to
take this as an indicator of the degree a
political system is democratic is to commit the
individualistic fallacy. An political system can
have an authoritarian regime even if most
citizens share democratic values
15
Important to Attend to What the Appropriate Unit
of Analysis is to Answer Question Individuals
The most typical unit of analysis for public
health research is individual human beings.   We
tend to describe and explain social groups and
interactions by aggregating and manipulating the
descriptions of individuals.   When individuals
are studies, they are characterized in terms of
certain dispositions and behaviors (e.g.,
hostility, committing suicide, mental health,
feelings of happiness) and are compared in terms
of properties (e.g., class, religion, age, sex,
etc.).   Descriptive studies using individuals as
the units of analysis typically aim to describe
the population that comprises those
individuals.   Explanatory studies using
individuals aim to discover the social dynamics
operating within that population.
16
Important to Attend to What the Appropriate Unit
of Analysis is to Answer Question Groups
Groups consist of people who interact with one
another according to established patterns,
identify themselves as group members, and are
identified as group members by others.   If
collective information about a particular group
is used on a statistical summary basis simply to
describe the makeup of individuals within the
group, then the unit of analysis is the
individual.   If aggregate information is assumed
to measure properties of the collectivity as a
whole, and is used in a comparison of different
groups, then the unit of analysis is the
group.   It is also possible to analyze
individuals while using aggregate data to
characterize the groups to which the individuals
belong.   An individual can be described as
living in a county with a high or low average
income, or as attending a racially integrated or
racially segregated school.   In these examples,
the individual is described in terms of
characteristics of the group, while
characteristics of the groups are formed by
aggregating information about individuals.   Compa
ring individuals who are in different groups
permits analysis of the effects of those groups
on the individual. However, in this case, the
individual would be the unit of analysis. 
17
Contextual vs. Compositional Effects
Compositional Explanation Areas include
different types of peoples and it is the
composition of the area that leads to observed
differences in health between areas. Poor People
Die Earlier than Rich People, so it is not
surprising that areas with lots of poor people
have a lower average life expectancy. Life
expectancy is a property of individuals not the
areas
Contextual Explanation Features of the social
and physical environment influence the health of
those exposed to them
Caution These two types of explanations are
highly intertwined. Not everyone or every group
has the same choices in life
18
Are Studies of Social Capital Using Contextual or
Compositional Explanations?
19
Place Matters Within Any Given Household Income
Range, Neighborhood Quality Affects Health
QUALITY RESIDENTIAL NEIGHBORHOODS
CHD INCIDENCE/PER 1,000 PERSON-YEARS
Household Income
Source Diez Roux, NEJM, 12 July 2001
20
Area Explanations Offered by Professor Sally
McIntyre
Physical features of the environment shared by
all residents
Availability of a healthy environment at work,
home and play
Services provided, publicly or privately, to
support people in their daily lives
Sociocultural features of the neighborhood
The reputation of an area
21
Things to Remember
Correlation is not cause!
Things that happen in the same place dont
necessarily happen to the same people
The unit of analysis is the primary determinant
of the types of conclusions to be drawn
We are working a world of multi-level models so
you need to construct causal chains
Write a Comment
User Comments (0)
About PowerShow.com