Homework Questions?
Designing Sample finishing Day
19
"Convenience Sample/ Voluntary Response Sample/ Sampling
Frame/ Nonresponse bias"--relationships.
Simple Random Sample from Table B. Sources of bias even
if sample designed OK. Inference to the population.
More kinds of probability samples:
We will focus on the mathematics of the SRS,
the most basic. In practice, more sophisticated sampling methods
may be preferred. The math needed to analyze their effects is beyond
our course.
Optional: Here
are
some other ways to design a probability sample.
Literary Digest poll,
narrative
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
Ch. 3.2 Designing Experiments
Observational Study
vs. Experiment day
19
Different jargon; different traditions.
Do something to:
"Experimental Units" = "Subjects"
Treatment: Specific experimental condition.
Factor: Explanatory Variable we manipulate.
Levels: Specific values of a factor that
we set.
Response variable(s)
E.g. 2 headache medications, in combination?
A two-factor experiment, each with 3 levels. 9 possible treatments.
Factor A: Aspirin: levels None, 500 mg,
1000 mg
Factor B: Caffeine: levels None, 50 mg, 100 mg
Response variable: reported pain relief
| Aspirin | ||||
| None | 500 mg | 1000 mg | ||
| None | Treatment 1 | Treatment 2 | Treatment 3 | |
| Caffeine | 50 mg | Treatment 4 | Treatment 5 | Treatment 6 |
| 100 mg | Treatment 7 | Treatment 8 | Treatment 9 |
Lurking variables: Control--how?
Nothing except experimental treatment should differentially affect response.
Compare responses under several treatments,
look at differences.
Placebo effect: a positive response to a "sham" medical treatment--if
you believe it will work, it very likely will. (Tinkerbell?)
A medical treatment must be shown to be better than
a placebo (at least) to be approved by the FDA.
(Cf. "Claritin,"
Sunday NYTimes magazine, March 11, '01)
placebo="I
shall please" (Latin)
To control for the placebo effect, All treatments should "look alike".
Treatment 1 above should be a pill with no medicine--a "placebo".
(Some experiments even try to duplicate side effects of actual medication.)
"Control group" Group that gets the "baseline"--"null"--
"none" or "placebo" level of the factor. Should be "just like" the
group(s) that get the "treatment" ("real" levels of the factor).
So Treatment 1 above will go to the "control group", the other 8 will go
to "experimental" or "treatment groups."
Murky language here: "Experimental vs. control" or "Treatment
vs. control" is different usage from "Treatments", one of which is the
"control"="none"/"placebo".
*Sometimes the Control is the current
"best practice" treatment, rather than none.
How to get groups "just like" one another? Randomize
who
goes into which group. (Usually our batch of experimental units
is not a random sample from
the population of all individuals--volunteers, etc.)
Randomized comparative experiment : Diagrams
of design, Moore pp. 140-141
Completely randomized: all exp. units allocated at random among
the treatments.
E.g. does acupuncture work for PMS? Response: report of symptoms.
One factor, 3 Levels: None (music?),
Acupuncture
(wrong places), Acupuncture (right places). 3 treatments.
(control(s)?)
30 subjects with PMS: Randomize,
10 each treatment. Administer treatments. Compare symptoms.
(Do diagram)
Picking groups with random number table: Pick "sample" of size 10 from the 30 for first treatment. Pick another "sample" of size 10 for 2nd treatment, from the remainder. The 10 remaining get the 3rd treatment.
Why 10 each, not just 1 each? Replication of the experimental treatments on many unitsallows "averaging out" chance variation in the units. (Don't confuse the replication needed within one experiment, with "replication" of the whole experiment in a different time and place to confirm its results.)
Principles of designing an experiment (p. 143) See above
Start here Friday:
Lack of realism: Do sociology, psychology experiments
generalize to "real life?"
--Subjects are not a random sample from the population. (Most
psychology "facts" were based on studies of Ivy League males, before 1970's.)
--Ethical questions...Milgram
Placebo effect article:
Usually an experiment treats the placebo effect as a confounding
variable, and is designed so placebo effect will work equally
on all groups. There is no attempt to measure the
placebo effect. ("All" drug studies.)
PMS/acupuncture:
Acupuncture (wrong) vs. Acupuncture (right).
Sometimes an experiment deliberately tries to measure the
placebo effect (as in the article).
Acupuncture (wrong) vs.
Music.
| Sievers home | Math151-Fall04/Dayf20.htm | 3:40pm | 10/13/04 |