Transcript 5.2 Day 2: Designing Experiments

5.2 Day 2: Designing
Experiments
Block Design



In general, men and women may react
differently to different medications or be able
to lift different amounts of weight or on
average do a different number of push ups.
When comparing the number of push ups
that a group of fitness students can do at the
end of a training camp, it would make more
sense to separate men and women into
separate comparison groups.
This type of separation is called blocking.
Block Design



A block is a group of experimental units or
subjects that are known before the
experiment to be similar in some way that is
expected to systematically affect the
response of the treatments.
In a block design, the random assignment of
units to treatments is carried out separately
within each block.
Blocks are another form of control.
Ex: Comparing Cancer
It’s important that you
Therapies
understand that blocking is not




always due to gender.
Subjects could be blocked
based on whether or not they
exercise. Blocking is used to
reduce variability. Blocking
has nothing to do with random
assignment-for example, one
does not randomly assign
subjects to gender!
The progress of a type of cancer differs in
women and men.
A clinical experiment will compare three
different treatments.
Men and women will first be separated into
blocks.
Then each block will be randomly assigned to
the three different treatments.
Outline of block design for
cancer therapies experiment
Importance of Blocking


Blocks allow us to draw separate conclusions
about each block, for example, about men
and women in the cancer study.
Blocking also allows more precise overall
conclusions because the systematic
differences between men and women can be
removed when we study the overall effects of
the three therapies.
Blocking vs. Randomization



Blocking is used to control for the variables
you know about that might influence the
response.
Randomization is used to control for the
variables you do not know about.
Use the mantra: control what you can, block
what you can’t control, and randomize the
rest.
Matched Pairs Design

Completely randomized designs are the
simplest statistical designs that clearly
demonstrate the principals of CRR.

However, completely randomized designs are
often inferior to more elaborate statistical
designs.

Using a matched pairs design, where
subjects are matched in various ways can
produce more precise results.
Matched Pairs Design
Matched pairs
are an
example
of
 The subjects are matched in pairs
and only
two treatments are compared block design.



For example, an experiment to compare two
advertisements for the same product might
use pairs of subjects with the same age, sex,
and income.
It is not always easy to match subjects.
One common variation of the matched pairs
design imposes both treatments on the same
subjects, so that each subject serves as his
or her own control.
Ex: Cell Phones and Driving



In this experiment, the effects of driving while
talking on a cell phone are to be observed.
There are two treatments: driving in a
simulator and driving in a simulator while
talking on a hands-free cell phone.
The response variable is the time the driver
takes to apply the brake when the car in front
brakes suddenly.

40 students subjects are assigned at random,
20 students to each treatment


Since subjects differ in driving skill and
reaction times, experimenters used a
matched pairs design in which all subjects
drove both with and without using the cell
phone.
They compared each individual’s reaction
time with and without using the cell phone.

The proper procedure would
require that all subjects first
be trained in using the
simulator, that the order in
which a subject drives with
and without the phone be
random, and that the two
drives be on separate days to
reduce carryover effects.
The reason that subjects are separated into
two groups, those who drive first without a
cell phone and those who drive first with a
cell phone is to reduce the possibility that
talking on a cell phone would be confounded
with driving a simulator for the first time.
Is the placebo effect
pseudoscience?

Fourteen healthy men were
given a saltwater injection
that caused pain to their
jaws. They were then
injected with a placebo and
told it was a pain killer.
Researchers monitored
their brain activity during
the process. Each man’s
brain released more natural
painkilling endorphins after
the placebos were
administered.
Double Blind Experiment


In a double-blind experiment, neither the
subjects nor the people who have contact
with them know which treatment a subject
received.
In the case of a medical study, neither the
doctor nor the patient would know whether or
not the patient was taking a placebo. This
helps eliminate unconscious bias in the way
the patient is treated.
Lack of Realism

Lack of realism is the most serious potential
weakness of experiments.

Ex: A study compares two television
advertisements by showing TV programs to
student subjects. The students know it’s “just an
experiment.” We can’t be sure that the results
apply to everyday television viewers. Many
behavioral science experiments use as subjects
students who know they are subjects in an
experiment. That’s not a realistic setting.
Ex: The Third Brake Light
When the experiment was




first conducted, most cars
did not have the third brake
light, so it caught the eye of
following drivers. Now that
all cars have them, they no
longer capture attention.
Do high centered brake lights, which have
been required on all cars sold in the U.S.
since 1986, really reduce collisions?
When randomized comparative experiments
were conducted prior to 1986, collisions were
reduced by as much as 50%.
After 1986, requiring the third light only led to
a 5% drop.
What happened?
Ex: Placebo cigarettes?




A study of the effects of marijuana recruited
young men who used marijuana.
Some were randomly assigned to smoke
marijuana cigarettes, while others were given
placebo cigarettes.
This failed: the control group recognized that
their cigarettes were phony and complained
loudly.
It may be quite common for blindness to fail
because the subjects can tell which treatment
they are receiving.