Chapter 4 (Sections 3 4) PowerPoint.ppt

Download Report

Transcript Chapter 4 (Sections 3 4) PowerPoint.ppt

GATHERING DATA
Chapter 4
4.3 What are Good and Poor Ways to Experiment?
Elements of an Experiment




www.clinplus.com
Experimental units:
Subjects
Treatment: Conditions
imposed on subjects
Explanatory variable:
Defines groups and
treatments
Response variable:
Outcome
Experiments



Impose treatments on
subjects to observe
responses
Goal: compare effects of
treatments on response
Randomized experiments –
subjects randomly
assigned to treatments
www.privatemdlabs.com
Placebo effect
Placebo – fake
treatment; sugar
pill
Placebo effect –
improving not from
real treatment but
from belief that he
or she should
improve
www.scientificamerican.com
3 Components of a Good Experiment



pages.cs.wisc.edu
Control or Comparison Group
Randomization
Replication
Principle 1: Control or Comparison Group



Helps analyze
effectiveness of
primary treatment
Placebo removes
lurking variables
Control group gets
placebo
 Clinical
trials may
compare new
treatment with existing
www.redicecreations.com
Control or Comparison Group


www.gpaulmckinneyjr.com
Experiments should
compare treatments
rather than effect of
single treatment
Example: 400
volunteers asked to quit
smoking with some
taking and some not
taking antidepressant
Principle 2: Randomization
1.
2.
colorectal.surgery.ucsf.edu
Eliminates bias
from researcher
assigning
subjects
Balances groups
on known and
lurking variables
Principle 3: Replication
1.
2.
Reduces
difference due
to ordinary
variation or
chance
Increases
chance that
results show
true difference
www.printercopierblog.com
Blinding the Experiment


gridskipper.com
Blind – subjects unaware
of which treatment used
Double-Blind Experiment
- Neither subjects nor
investigators know which
treatment
 Controls bias from
respondent and
experimenter
Statistical Significance
Statistically Significant
Difference –
Observed
difference is larger
than expected from
chance
Generalizing Results


static.howstuffworks.com

Goal of
Experimentation –
Analyze association
between treatment
and response for
entire population
Generalize only to
population
represented by study
Page 180 #34, 40
4.4 Other Ways to Conduct Experimental and
Observational Studies
Sample Surveys: Random Sampling Designs

codetechnology.files.wordpress.com
Alternative to
experiments
1. Simple Random
Sampling
2. Cluster Random
Sampling
3. Stratified
Random
Sampling
Cluster Random Sample
1.
2.
3.
Divide population
into large
number of
clusters, such as
city blocks
Select simple
random sample
of clusters
Use all subjects in
clusters as sample
Cluster Random Sample
reactorfire.files.wordpress.com
Advantages
 Sampling
frame
unavailable
 Cost
Disadvantage
 Need larger
sample size
for same
reliability
Stratified Random Sample
1.
2.
3.
Divide the
population into
groups, strata
Select SRS from
each strata
Combine samples
from each for total
sample
www.nedarc.org
Stratified Random Sample
www.smh.com.au
Advantage
 Ensures stratum
representation
Disadvantage
 Need sampling
frame and to
which stratum
each subject
belongs
Comparing Random Sampling Methods
Types of Observational Studies
Sample Survey:
current
2. Retrospective
Study: past
3. Prospective Study:
future
Cause not proven,
but studies can
support beliefs
1.

Retrospective Case-Control Study
Retrospective
Studying sunlight
exposure and multiple
sclerosis connection…
Cases – have MS
Controls or don’t
Explanatory variable –
low sun or not
www.lisisoft.com
Prospective Case-Control Study
Prospective
Studying
effects of
vegetarian
diet on heart
disease…
Cases – have heart disease
Controls or don’t
Explanatory variable –
vegetarian or not
www.dubaiinternetmarketing.com
Multifactor Experiments


Single experiment
analyzes two or more
factors
Learn more since
combinations may
affect response
biobreak.files.wordpress.com
Matched Pairs Design



www.marshallhammondwedding.info
Subjects are somehow
matched
 Husband/wife, two plots in
same field, etc.
 Same individual – crossover
design
Randomly assign or randomize
order of treatments
Reduces effects of lurking
variables
Randomized Block Design


Block – subjects with common characteristics
Randomized Block Design, RBD – within each block,
randomly assign to treatments
Examples

Page 191, #44, 46, 48, 50