Transcript Treatment - s3.amazonaws.com

Designing Experiments
Purpose for experiments – to study
the response of one variable to the
changes in other variables.
Experimental Units (Subjects)
Individuals on which experiment is
performed
Treatment
Applied experimental conditions
e.g. aspirin
Factor
Explanatory variable of an
experiment
Designing Experiments
Example: Researchers studying the
absorption of a drug into the
bloodstream inject the drug into 25
people. 30 minutes after the injection,
they monitor the concentration of the
drug in a subject’s blood.
Treatment =
Individuals =
# of factors =
Response variable =
Designing Experiments
Comparative experiment
Units  Treatment  Response
designed in such a way as to
eliminate or minimize
confounding variables
Designing Experiments
Confounding variables can introduce bias into
the results of experiments
Their influence can be minimized by specific
experimental design
Randomization of individuals into two groups.
Control group – group of individuals that
receive sham treatments
Placebo group (Placebo – dummy
treatment)
½ of individuals
Experimental group – group of individual
that receive the treatment
½ of individuals
Designing Experiments
Assignment into either experimental or control
groups can be done via double-blinding
where neither the subjects nor the personnel
who interact with the subjects know who
receives treatment and who receives placebo
Designing Experiments
Randomized Comparative Experiment
Let impersonal chance assign subjects to
groups – Randomization
Completely Randomized Designs
All experimental units are allocated at
random among all treatments
Group 1
Treatment
Random
Assignment
Response – compare
Results from two groups
Group 2
Placebo
Designing Experiments
Example: Eye cataracts are responsible for over 40%
of blindness around the world. Can drinking tea
regularly slow the growth of cataracts? We can’t
experiment on people, so we use rats as subjects.
Researchers injected 18 young rats with a
substance that causes cataracts. One group of
the rats also received black tea extract; a second
group receive green tea extract; and a third got a
placebo. The response variable was the growth of
cataracts over the next six weeks. Yes, both tea
extracts did slow cataract growth.
Outline the design of this experiment.
Designing Experiments
Random Allocation of Subjects
There will always be some
difference outcomes due to chance
variation among the subjects
Designing Experiments
Observed difference in the
response of an experiment must be
large enough to assure that it did
not arise just by chance
• When comparing response from each
group (placebo and experimental) after
the experiment is finished, the
difference between the two groups must
be large enough so that scientists can be
sure that it is arising due to effect of the
treatment, and not by chance
An effect that is large is called:
Statistically Significant
Designing Experiments
Principles of Experimental Design
1.
Control for confounding variables by
comparing various treatments
2.
Use randomization in choosing subjects
for various experiments
3.
Gather enough information (use enough
subjects) in each group to reduce
chance variation in the results.
MATCHED-PAIRS DESIGNS
Compares two treatments
Example of block design
2 ways:
1.
Chose pairs of subjects closely matched
One of the treatments is randomly assigned to
one of the subjects
2.
use only one individual
give both treatments at random
Example: Coke taste test: Subjects were given
taste of Coke and Pepsi and reported their
preference. Whether Coke or Pepsi were
offered first was chosen randomly
BLOCK DESIGN
BLOCK
group of experimental units or subjects that
are known before the experiment to be similar
in some way that is expected to affect the
response to the treatments.
Random assignment of units to treatments is
carried out separately within each block.
Another form of control of lurking variables
Draw separate conclusions about each block
Example: Women and men respond differently to
advertising. An experiment to compare the
effectiveness of three television commercials
for the same product will want to look
separately at the reactions of men and
women, as well as assess the overall
response to the adds. A completely
randomized design considers all subjects,
both men and women, as a single pool. The
randomization assigns subjects to three
treatment groups without regard to their sex.
This ignores the differences between men and
women. A better design considers women
and men separately. Randomly assign the
women to three groups, one to view each
commercial. Then separately assign the men
at random to three groups.