Transcript 5_sampling2
Sampling (conclusion) &
Experimental Research Design
Readings: Baxter and Babbie, 2004,
Chapters 7 & 9
Issues in Non-probability sampling
Bias?
Is the sample representative?
Types of sampling problems:
Alpha:
find a trend in the sample that does not
exist in the population
Beta: do not find a trend in the sample that
exists in the population
Principles of Probability Sampling
each member of the population an equal chance of
being chosen within specified parameters
Advantages
ideal
for statistical purposes
Disadvantages
hard
to achieve in practice
requires an accurate list (sampling frame or operational
definition) of the whole population
expensive
Types of Probability Sampling
1. Simple Random Sample
With
replacement
Without replacement: link
2. Systematic Sample (every “n”th person) With Random Start
Urban
studies example)
3. Stratified Sampling:
Sampling
Disproportionately and Weighting
4. Cluster Sampling
Examples of sampling issues &
techniques
Survey about football (soccer) market
Rural poverty project and sampling issues
Postpone: Techniques for
Assessing Probability Sampling
We will discuss these in connection with
Chapter 11 material:
Standard deviation
Sampling error
Sampling distribution
Central limit theorem
Confidence intervals (margin of error)
Introduction to Experimental Design
Recall discussion of
experiments in lecture on
Research Ethics
Milgram experiment (on
obedience)
Stanford prison experiment
about how prisons as
institutions communicate
roles and shape actions
(still photo from video on
right showing research
subjects dressed as prison
guard & prisoners)
Trends in Experimental Social
Research
types of subjects & reporting style (naming
vs. anonymity)
deception & risk
debriefing
Single & double Blind Experiments
Neuman (2000: 239)
Key Notions / Terms
Treatment, stimulus, manipulation (independent
variable)
observable outcome (dependent variable)
Experimental Group
Control group
pretest (measurement before treatment)
posttest (measurement after treatment)
Random Assignment
Neuman (2000: 226)
Comparison with Random Sampling
Neuman (2000: 226)
How to Randomly Assign
Neuman (2000: 227)
Experimental Design Notation
O= observation
X= treatment
R= random
assignment
Some Common Types of Design
Three common types of
experimental design: Classical
pretest-post test –
Total population randomly divided into
two samples;
control
sample
experimental sample.
Only the experimental sample is exposed
to the manipulated variable.
compares pretest results with the post test
results for both samples.
divergence between the two samples is
assumed to be a result of the experiment.
Solomon four group design –
The population is randomly divided into four
samples.
Two of the groups are experimental samples.
Two groups experience no experimental
manipulation of variables.
Two groups receive a pretest and a post test.
Two groups receive only a post test.
improvement over the classical design because it
controls for the effect of the pretest.
Factorial design –
similar to a classical design except
additional samples are used.
Each group is exposed to a different
experimental manipulation.
Factorial
Design
Validity Issues
internal validity: elimination of plausible
alternative explanations
external validity: ability to generalize
(outside the experiment)
Internal Validity Threats
selection bias: groups not equivalent
history: unrelated event affects exp.
maturation: separate process causes effects
testing: ex. Pretest effects
More Internal Validity Threats
instrumentation: measure changes
mortality/attrition
statistical regression : ex. Violent films
contamination
compensatory behaviour
experimenter expectancy
External Validity Threats
realism
reactivity:
Hawthorne
effect
novelty effect
placebo effect
Laboratory vs. Field experiments
lab.- more control , higher internal validity
field- more natural, higher external validity
Recall : New Ethical Norms
protection of subjects
debates about deception