Review material for Exam 2

Review material for Exam 2
Chapters 12, 13, 14, 15, 16
Warning!
• This review material cannot possibly contain
everything we’ve covered so far.
• Thus, you must study the lecture notes
(available on the class website) to be fully
prepared for the exam.
• Also, looking at the homework problems as
well as class examples/exercises will help.
• Most problems will come from those as well
as from this review, but not all!
Ch 12. Sample Surveys
• What does it mean to survey a sample?
• Sample is a “small” group of individuals, who
are supposed to represent the entire
population.
• To achieve this, a sample must be drawn
randomly from the sampling frame.
• So by surveying a (random) sample, we hope
to get the general opinions of the entire
population.
Types of Random Sampling
1. Simple random sampling (SRS) - All samples of the
same size have an equal chance of being selected
from the entire population.
2. Stratified sampling - The population is sliced into
homogeneous groups (called strata) AND SRS is used
within each randomly chosen stratum.
3. Cluster sampling – The population is sliced into
heterogeneous groups (called clusters) that represent
the population  Perform a census for each
randomly selected cluster.
4. Multistage sampling - Sampling schemes that
combine several methods.
5. Systematic sampling - Drawing a sample by selecting
individuals systematically.
Identify the sampling method.
• #4. … Officials select a team at random, and
drug-testing crew shows up unannounced to
test all 40 players on the team…
• #10. The Gallup Poll interviewed 1007
randomly selected U.S. adults aged 18 and
older, March 23-25, 2007…
Identify the sampling method
(Continued)
• #16 (revised for clarity) Dairy inspectors select
farms randomly and take samples of the milk
from each farm to test for contamination…
• #22 (revised for clarity) …They chose a
random fishing day and randomly selected 5
fishing boats to count the numbers and types
of all the fish on those boats.
Bad (i.e. non-random) sampling
1. Voluntary Response sampling: All are invited
to respond, and anyone who responds is
counted.
 Those who have strong opinions on the
survey questions are more likely to answer!
2. Convenience sampling: we simply include
the individuals who are convenient to survey.
Identify the sampling method
• #8. A question posted on the Lycos Website on 18
June 2000 asked visitors to the site to say
whether marijuana should be legal for medical
purposes.
• #12 (revised for clarity) … The campaign director
for a mayoral candidate selects one block from
each of the city’s election districts. Then the staff
members go there and interview those residents
that are present at that time.
Bias
1. Undercoverage bias: Some portion of the population is not
sampled at all or has a smaller representation in the sample than it
has in the population, due to design issue (NOT because the portion
of people chooses not to answer.)
Example) If a survey is done only through landlines, opinions of those
who do not have landlines cannot be counted. PERIOD. (i.e. NOT by
choice, but by design issue)
2. Non-response bias: Those who don’t respond to a survey may
have different responses than those who do.
3. Response bias: Anything in the survey design that influences
particular responses.
Example) Do you support installing a new vending machine in our floor
so that we won’t have to go all the way to cafeteria to get sodas?
2. Voluntary response bias : from voluntary response
sampling
Identify possible bias
• Sending out questionnaire to parents without
any follow up when they fail to send them
back.
• Sending out questionnaire to students without
postage provided.
• Texting random cell-phone numbers for
questions.
• Asking random people questions on a busy
street AND on a rainy day.
Ch 13. Observational studies and
experiments
• Observational study is a study based on
observation. No manipulation of factors
occurs.
• Observational studies can be categorized as
either prospective or retrospective.
• From the moment of choosing of subjects, you
decide whether the data are collected by
going back or going forward.
• Prospective or retrospective? #10, 26
Questions
• What is the one thing that distinguishes an
experiment from observational study?
• Ans:
• Can you conclude a causal relationship from
an observational study?
• Ans:
Terminologies
• You have an experiment if the factors are
intentionally being manipulated.
• Know the terms “factors”, “levels”, and.
“treatments”
• Example) 1 factor, 3 levels 
• Example) 2 factors with 2 levels in each factor

• Let’s try #14.
More Terminologies
• Know the following terminologies and
relevant examples.
1. Control Group
2. Placebo
3. Blind (Single or Double) – What’s the reason?
4. Blocking (or Blocks)