Chapter 11 - Experimental Designs Flashcards
Establishing a Causal Relation
- Association
– Independent and dependent variables covary - Temporal precedence
– Presumed cause (IV) precedes presumed effect (DV) - Isolation
– Presumed cause (IV) must be able to affect the presumed effect
(DV) all by itself (NOT due to another factor – i.e., a nuisance,
confounding, contaminating, or extraneous variable)
– Does NOT mean that there is only one cause for any given effect
(DV) – think “a” cause, not “the” cause
Extent of Researcher Interference
Control
Extent of Researcher Interference
Control
Control is ____
– Important for establishing causality
– Designs vary in degree & forms of control
Factors of cause-probing Experiment
– Manipulation of IV(s)
– Random assignment of units to conditions
* Control & experimental
– Measure effects on DV(s)
– Control potential
confounds
– Clarification:
* Sometimes researchers will do a pre-test and post-test on the DV
* Pre-testing not a defining feature of an experiment
Factors of cause-probing Quasi-experiment
– Very similar to experiment
– Except: No random assignment
* May use intact or selfselected groups instead
– More likely to measure potential confounds
* Statistical control
– Similar to both experimental & relational designs
What was the Hypothesis of the Popcorn Experiment video?
Orville popcorn makes movies more enjoyable.
List the below Features of the Popcorn experiment video
– Manipulation (IV/treatment/intervention)?
– DV/outcome?
– Assignment to conditions?
– Other considerations (e.g., nuisance variables)?
– Manipulation (IV/treatment/intervention)?
——The Popcorn
– DV/outcome?
—— Perceived enjoyment (Rated by patrons)
– Assignment to conditions?
—— Quasi-Experiment
– Other considerations (e.g., nuisance variables)?
—— Lifting of the curtain, Temperature variance,
Internal Validity
The extent to which causal inferences (e.g., A causes B) are justified by a research design.
External Validity
The extent to which the results from a study (e.g.,
a causal relation between A and B) can be applied
to other settings, people, or events.
Threats to Validity
History
Maturation
Testing
Selection bias
Mortality (attrition)
Statistical regression (Regression to the mean)
Instrumentation
General observation
Threats to Validity with History
Events occurring concurrently with the treatment could cause the observed effect on the DV.
Threats to Validity with Maturation
Naturally occurring changes over time could
be confused with a treatment effect.
Threats to Validity with Testing
Exposure to a test can affect score on subsequent
exposures to that test, and this could be confused
with a treatment effect.
* Main (pre-test affects post-test)
* Interactive (pre-test affects reaction to the treatment)
Threats to Validity with Selection bias
Systematic differences between conditions in
respondent characteristics that could cause any
observed effect.
Threats to Validity with Mortality (attrition)
Loss of respondents to treatment or to
measurement can produce spurious effects if the
attrition is systematically correlated with
conditions (members of one group are more likely
to drop out).
Threats to Validity with Statistical regression
When units are selected for their extreme scores,
they will often have less extreme scores on
repeated measurements or on other variables.
Threats to Validity with Instrumentation
The nature of a measure may change over time or
conditions in a way that could be confused with a
treatment effect.
Threats to Validity with General observation
“Factors Affecting the Validity of Experiments” or “Threats to Validity”
* Inclusion of this material is a nice feature of an introductory
research methods text!
* More complete lists of threats to validity are available
elsewhere.
Experimental Designs
Basic