Experiments and Generalized Causal Inference General Terminology

1
Experiments and Generalized Causal
Inference(General Terminology)

• Presented By Lizzy Newsome and Darlene
  Villafranca

2
Overview of Chapter 1

• The nature of causation that experiments test
• Specialized terminology
• The problem of how to generalize causal
  connections from individual experiments
• Situating the experiment within a larger
  literature on the nature of science

3
Randomized Experiments

• Experiments where participants are randomly
  assigned to the experimental groups or
  conditions.
• Often referred to as true experiments.
• A study in which an independent variable is
  deliberately manipulated and a dependent variable
  is assessed.

4
Reasons for Using Random Assignment

• Provides a safeguard against biased assignment of
  sampling units to the different treatment groups.
• Distributes the characteristics of the sampling
  units over the different conditions to prevent
  biased outcomes.
• Permits the use of statistical analyses that
  require certain data characteristics.

5
Cause, Effect and Causal Relationships

• Cause a variable that produces an effect or
  result (give example)
• Inus Condition from philosopher J.L. Mackie
  (1984), the idea that a cause is an insufficient
  but non-redundant part of an unnecessary but
  sufficient condition for bringing about an effect
  (fire/match example)

6
(No Transcript)
7
Cause, Effect and Causal Relationships

• Counterfactual-something that is contrary to the
  fact .
• Effect-the difference between what did happen and
  what wouldve happened.
• There are more complicated causal relationships,
  whereby the same variable can be both a cause and
  an effect therefore creating a reciprocal
  relationship between two variables that cause
  each other.

8
Cause, Effect and Causal Relationships

• According to John Stuart Mill,
• a causal relationship exists if
• (1) the cause preceded the effect
• (2) the cause was related to the effect,
• (3) we can find no plausible alternative
  explanation for the effect other than the cause.
• These three characteristics mirror what happens
  in experiments in which
• (1) we manipulate the presumed cause and observe
  an outcome afterward
• (2) we see whether variation in the cause is
  related to variation in the effect
• (3) we use various methods during the experiment
  to reduce plausibility of other explanations for
  the effect, along with ancillary methods to
  explore the plausibility of those we cannot rule
  out (most of this book is about is about methods
  for doing this).

9
Causation, Correlation, and Confounds

• Correlation does not prove causation - which
  variable came first?? (ex. education or income)
• Correlation - A measure of the strength of
  relationship between two variables.
• Confounds an extraneous variable that covaries
  with the variable of interest. Sometimes
  relationships may not be causal at all, but
  rather confounded
• Causation concerns the time order relationship
  between two or more objects such that if a
  specific antecendent condition occurs the same
  consequent must always follow.

10
(No Transcript)
11
(No Transcript)
12

•   Psychologists gather data in order to
  describe, understand, predict, and control
  behavior. The scientific method refers to an
  approach that can be used to discover accurate
  information. It includes these steps understand
  the problem, collect data, draw conclusions, and
  revise research conclusions.

13
(No Transcript)
14
Basic Terminology

• The extent to which a test measures what it is
  intended to measure is called validity.
• A specific statement about behavior or mental
  processes that is testable through research is a
  hypothesis. Empirical means the use of working
  hypotheses which are capable of being disproved
  using observation or experiment.

15
Basic Terminology
The Ladder of Inference

• Inference is the act or process of drawing a
  conclusion based solely on what one already
  knows. Theories are logically self-consistent
  models or frameworks describing the behavior of a
  certain natural or social phenomenon. They are
  broad explanations and predictions concerning
  phenomena of interest.

16
(No Transcript)
17
Causal Relationship

•    To say that A is a sufficient condition for B
  is to say precisely the converse that A cannot
  occur without B, or whenever A occurs, B occurs.
  Ex. That there is a fire is sufficient for there
  being smoke.

Interpret the data on this graphIs there a
causal relationship?
18
(No Transcript)
19
Generzalization

•  In conditioning, the tendency for a
  conditioned response to be evoked by stimuli that
  are similar to the stimulus to which the response
  was conditioned is a generalization. The greater
  the similarity among the stimuli, the greater the
  probability of generalization.A statistically
  significant difference between two groups
  experiencing different manipulations of the
  independent variable is called the experimental
  effect.

20
Correlations

• A correlational study observes or measures two
  or more variables to find relationships between
  them. Such studies can identify lawful
  relationships but cannot determine whether change
  in one variable is the cause of change in
  another.A statistical technique for determining
  the degree of association between two or more
  variables is referred to as correlation.

21
(No Transcript)
22
Most experiments are highly local but have
general aspirations

• Correlational design
• Passive observational design
• Nonexperimental design
• Cause effect are identified
• Measured
• No random assignment
• No pretest control groups to construct useful
  counterfactual inferences

23
Most experiments are highly local but have
general aspiration

• Conducted in a restricted range of settings
• Convenient sample
• One type of treatment
• Several measures with theoretical assumptions

24

•   Placebo refers to a bogus treatment that has
  the appearance of being genuine.

25
Basic Terminology

•   A condition in a scientific study that is
  manipulated (assigned different values by a
  researcher) so that the effects of the
  manipulation may be observed is called an
  independent variable. A measure of an assumed
  effect of an independent variable is called the
  dependent variable.

26
Basic Terminology

•    Assignment of participants to experimental and
  control groups by chance is called random
  assignment. Random assignment reduces the
  likelihood that the results are due to
  preexisting systematic differences between the
  groups.A group that does not receive the
  treatment effect in an experiment is referred to
  as the control group or sometimes as the
  comparison group.

27
(No Transcript)
28
Popper and Kuhns Ideas

• Popper proposed a set of methodological
  rules called Falsificationism. Falsificationism
  is the idea that science advances by unjustified,
  exaggerated guesses followed by unstinting
  criticism. Only hypotheses capable of clashing
  with observation reports are allowed to count as
  scientific. Falsifiable theories enhance our
  control over error while expanding the richness
  of what we can say about the world. Kuhn is
  most famous for his book The Structure of
  Scientific Revolutions in which he presented the
  idea that science does not evolve gradually
  toward truth, but instead undergoes periodic
  revolutions which he calls "paradigm shifts."

29
Basic Terminology

• A generalized concept, such as anxiety or
  gravity, is a construct.An experimental design
  that is used to establish the relationship
  between two variables without the ability to
  infer causal relationships is a correlational
  design.Population refers to all members of a
  well-defined group of organisms, events, or
  things.

30
(No Transcript)
31
Cronbach

• Cronbach is most famous for the development of
  Cronbach's alpha, a method for determining the
  reliability of educational and psychological
  tests. His work on test reliability reached an
  acme with the creation of generalizability
  theory, a statistical model for identifying and
  quantifying the sources of measurement error.

• Cronbachs UTOS
• Uunits (people)
• Ttreatments
• Oobservations (outcomes)
• Ssetting
• refers to the instances on which data are
  collected

32
Construct Validity

• Problem
• How to go from the particular units, treatments,
  observations, and setting on which data are
  collected to the high order constructs these
  instances represent.
• Add own interpretations
• Change original thinking of the research
• How can we generalize from a sample of instances
  and data to particular target constructs?

33
External Validity

• External validity is a term used in
  scientific research. It signifies the extent to
  which the results of a study can be applied to
  circumstances outside the specific setting in
  which the research was carried out. In other
  words, it addresses the question "Can this
  research be applied to 'the real world'?"
• Choosing a sample so that each member of the
  population has an equal chance of being included
  in the sample is called random selection.

34
Principals

• Surface Similarity
• Ruling Out Irrelevancies
• Making Discriminations
• Interpolation and Extrapolation
• Casual Explanation