General Issues in Research Design: Causation

1
General Issues in Research Design Causation
Validity

• Dr. Guerette

2
Three Elements of Research

• Causation
• Units of Analysis
• Time
• With a primary aim of explanation, cause in
  social science research is probabilistic.

3
Causation in Social Science

• Necessary elements for causality
• 1. Time order
• the cause must precede the effect in time.
• 2. Relationship
• the two variables must be empirically correlated
  with each other or must be logically related.
• 3. Explanation of effect
• the relationship must not be the result of some
  third variable.

4
Types of Causality

• Necessary cause
• is a condition that by and large, must be present
  for the effect to follow.
• Sufficient cause
• is a condition that more or less guarantees the
  effect in question.

5
Validity Causal Inference

• When we are concerned with whether we are correct
  in inferring that a cause produced some effect,
  we are dealing with validity.

6
Types of Validity

• Statistical conclusion validity
• deals with our ability to determine whether a
  change in the suspected cause is statistically
  associated with a change in the suspected effect.
• Internal validity
• deals with whether the association between two
  variables is causal and not the result of the
  effects of one or more other variables.

7
Types of Validity

• External validity
• deals with whether research findings from one
  study can be reproduced in another study often
  under different conditions.
• Construct validity
• deals with how well an observed relationship
  between variables represents the underlying
  causal process of interest.

8
Units of Analysis

• Individuals
• for example inmates, police officers, students,
  or judges.
• Groups
• for example police beats, city blocks or cities.
• Organizations
• for example, political parties, courts or rehab
  treatment facilities.
• Social artifacts
• for example, police reports newspaper editorials
  or TV programs.

9
Errors dealing with Units of Analysis

• The Ecological Fallacy
• the act of making assertions about individuals as
  the unit of analysis based on examination of
  groups or other aggregations.
• Reductionism
• refers to an overly strict limitation on the
  kinds of concepts and variables to be considered
  as causes in explaining a broad range of human
  behavior.

10
In Class ExerciseUnits of analysis
For the following, assign the objects listed with
the unit of analysis

• Individuals
• Groups
• Organizations
• Social artifacts

• Reports of drunk driving accidents
• Cities and counties
• Gangs
• Probationers
• Court cases
• Police agencies
• Burglars
• News reports
• Universities

11
In Class ExerciseUnits of analysis

• Individuals
• Probationers
• Burglars
• Groups
• Cities and counties
• Gangs
• Organizations
• Police agencies
• Universities
• Social artifacts
• Reports of drunk driving accidents
• Court cases
• News reports

12
The Time Dimension

• Cross-sectional Studies
• A technique of using one point in time to gather
  data and analyze that information carefully.
• Longitudinal
• A technique designed to permit observations over
  an extended period of time.

13
The Time Dimension

• Three types of longitudinal studies
• Trend studies
• Look at changes within some general population
  over time.
• Cohort studies
• Look at a specific population (often an age group
  or some other time grouping) over time.
• Panel studies
• Similar to trend and cohort studies, except that
  observations are made on the same set of people
  on two or more occasions.

14
Approximating Longitudinal Studies

• Retrospective studies
• Ask people to recall their pasts as a means of
  approximating observations over time.
• Prospective studies
• Follow people into the future to gauge
  observations over time.

15
In Class Exercise Time in research design
Year of Measurement Year of Measurement Year of Measurement
Year of Birth (cohort) 1980 1990 2000
1960 A-20 B-30 C-40
1950 D-30 E-40 F-50
1940 G-40 H-50 I-60
16
In Class Exercise Time in research design

• 1. Outline (or circle) the cells in the above
  table which would be representative of a
  cross-sectional research design and those of a
  longitudinal design.
• 2. Specify how a panel study, cohort study, and a
  trend study would differ.

17
The Research ProcessThe Traditional Deductive
Model

1. Theory Construction
2. Derivation of theoretical hypotheses
3. Operationalization of Concepts
4. Collection of empirical data
5. Empirical testing of hypothesis

18
Scientific Realism

• Bridges idiographic and nomothetic approaches to
  explanation by seeking to understand how causal
  mechanisms operate in specific contexts.
• This moves away from the traditional social
  science research paradigm.