Transcript Section 5.2 - Humble ISD

Section 5.2
Designing Experiments
• Observational Study - Observes
individuals and measures variables of
interest but DOES NOT attempt to
influence the responses.
• Experiment – DELIBERATLY imposes
some treatment on individuals in order to
observe their responses.
• Experimental Units – individuals on which
the experiment is done.
– If the experimental unit is a human being,
then they are called subjects.
– Treatment – experimental condition applied to
units
• Purpose – of experiment is to reveal the
response of one variable to changes in other
variables.
– Must distinguish between response and explanatory
variables.
• Factors – the explanatory variable in an
experiment.
• In theory experiments can give good evidence
for causation.
– Allow us to study specific factors we are interested in
while controlling lurking variables.
3 Principles of Experimental Design
1) Control – needed to counter effects of
lurking variables, etc. Simplest form of
control is comparison. Experiments
should control 2 or more treatments in
order to avoid confounding the effect of a
treatment with other influence
3 Principles of Experimental Design
2) Randomization – subjects assigned to
treatments by pure chance (table of
random digits).
- Creates groups that are similar except for
chance variation.
- ***
3 Principles of Experimental Design
3) Replication – do experiment on many
subjects to reduce chance variation in the
results.
• Treatment – a specific experimental
condition applied to experimental units.
• Placebo – a dummy treatment that can
have no physical effect
• Control Group – receives dummy
treatment
– Helps control lurking variables.
5.31
• The ability to grow in shade may help pines
found in the dry forests of Arizona to resist
drought. How well do these pines grow in
shade? Investigators planted pine seedlings in a
greenhouse in either full light or light reduced to
5% of normal by shade cloth. At the end of the
study, they dried the young trees and weighed
them.
• Experimental units?
• Factors?
• Treatments?
• Response Variables?
5.34
• Sickle-cell disease is an inherited disorder of the
red blood cells that in the United States affects
mostly blacks. It can cause severe pain and
many complications. Can the drug hydroxyurea
reduce the severe pain caused by sickle-cell
disease? A study by the National Institutes of
Health gave the drug to 150 sickle-cell sufferers
and a placebo (dummy medication) to another
150. The researchers then counted the
episodes of pain reported by each subject.
• Experimental units?
• Factors?
• Treatments?
• Response Variables?
Comparative Experiments
• Experiments should compare treatments
rather than attempt to assess a single
treatment in isolation
– Control and placebo groups
– Laboratory experiments often simple design
with single treatment.
– EX gastric freezing p 292, 5.11
Completely Randomized
Experiments
• All experimental units are allocated at
random among the treatments
• Example 5.12, p295
• 32,33,35,36,37,39,40
Logic of Experimental Design
• Randomization – produces similar groups
• Comparative Design – ensures
experimental treatments operate equally
on all groups
• Results – differences in response variable
must be due to effects of treatment
• A strong association in data from well organized
data experiments does imply causation. (cause
– effect relationship)
• Statistically significant observation – an
observed result too unusual to be an outcome
determined by pure chance (more in ch 10)
• Randomized Comparative Experiments are best
means of gaining knowledge about the effects of
explanatory variables on a response
• Examine every experiment with a critical eye,
watch for bias
• Double – blind experiment – neither the
subjects nor the people who have contact
with them know which treatment a subject
receives.
• The most serious weakness of
experiments is lack of realism
• Lack of realism can limit the ability to apply
the conclusions of an experiment to the
settings of greater interest
• Block design – is a group of experimental
units or subjects that are similar in ways
that are expected to affect the response to
the treatments.
– Used to minimize variation.
– Similar to stratified designs
• Matched Pairs Design – a common form of
blocking for comparing two treatments.
• 38 in class
• 37,39,40, 44, 50, 52, 54, 57 homework
Section 5.3 – basics of simulation
• Simulation – the imitation of chance
behavior, based on a model that
accurately reflects the experiment under
consideration
– an effective tool for finding likelihoods of
complex results once we have a trustworthy
model.
– Gives us good estimates of probabilities
Steps of Simulation
1)
2)
3)
4)
State problem or describe experiment
State the assumption
Assign digits to represent outcomes
Simulate many repetitions