J494 TEST EXAM QUESTIONS AND ANSWERS
LATEST UPDATE
Sample
individuals from a population for research
Sampling frame
a list of individuals from whom the sample is drawn
How often does the census go? Why is it important?
every 10 years, determines government funding
Is the census truly a census?
No, it does not include every person in the US. in a true census, the population and
census are equal
no generalization no error
When you generalize a sample to a population, you have error. If your sample is your
population, there is no error
nonprobability sampling methods
convenient/purposive
convenence sampling method
Using whatever sample is conveniently available, not random
purposive sampling
a biased sampling technique in which only certain kinds of people are included in a
sample, not random
Can you generalize the results of a nonprobability sample to the population?
No
Oneshot study
you dont look at participants over time
longitudinal study. different kinds of longitudinal studies
looking at participants over time, trend, cohort, panel
trend longitudinal study
draws different samples from one population over time
cohort longitudinal study
draws samples from different populations different populations at different times
panel longitudinal study
Surveys the same people over time
Simple random sample
A sample of size n selected from the population in such a way that each possible
sample of size n has an equal chance of being selected.
systematic random sampling
Create a sampling frame, and every nth unit select a participant every nth unit, start at a
random point
What is a skip interval (n)? How do you calculate it?
the number of participants you skip in a systematic random sample, population/sample
size
stratified random sample
, organize participants into strata homogenous, select randomly from strata
cluster sampling
organize participants into heterogenous clusters, select randomly from culsters
coverage error
a sampling error that occurs when sampling frame is flawed or outdated
sampling error
an error that occurs when a sample somehow does not represent the target population
How to reduce coverage error
up to date lists and sampling frame
How to reduce sampling error
draw enough people, have a representative sample
What is a confidence level?
How sure you are that results fall within a certain interval
Standard confidence level for social science/comms?
95%
What is the normal curve?
a symmetrical, bell-shaped curve that describes the distribution of many types of data.
The more random data you have, the more bell shaped it becomes
What is the peak of the bell curve?
The mean, 50th percentile, where most data points lie
How do you increase confidence level?
Increase number of units in the sample
What is a standard deviation?
How far is this data point from the mean?
What percent of data lies +/- one standard deviation of the mean?
68%, 34% on each side
What percent of data lies +/- two standard deviation of the mean?
95%, 34+13.5% on each side
What percent of data lies +/- three standard deviation of the mean?
99%, 34+13.5+2% on each side
What is margain of error?
determines the precision of accuracy of a poll result. actual results lie somewhere
between finding +/- mofe
What happens to mofe if you increase sample size?
decrease
You always have mofe unless
your sample size is your population
Is a poll actually able to predict an election winner if the ranges with mofe
overlap?
No
What is the acceptable sample size and mofe in social science?
384, 5%
Poll
quick version of a survey, 5 to 10 questions
What are some survey design considerations?
Goal, audience, method of delivery, question types and length,
LATEST UPDATE
Sample
individuals from a population for research
Sampling frame
a list of individuals from whom the sample is drawn
How often does the census go? Why is it important?
every 10 years, determines government funding
Is the census truly a census?
No, it does not include every person in the US. in a true census, the population and
census are equal
no generalization no error
When you generalize a sample to a population, you have error. If your sample is your
population, there is no error
nonprobability sampling methods
convenient/purposive
convenence sampling method
Using whatever sample is conveniently available, not random
purposive sampling
a biased sampling technique in which only certain kinds of people are included in a
sample, not random
Can you generalize the results of a nonprobability sample to the population?
No
Oneshot study
you dont look at participants over time
longitudinal study. different kinds of longitudinal studies
looking at participants over time, trend, cohort, panel
trend longitudinal study
draws different samples from one population over time
cohort longitudinal study
draws samples from different populations different populations at different times
panel longitudinal study
Surveys the same people over time
Simple random sample
A sample of size n selected from the population in such a way that each possible
sample of size n has an equal chance of being selected.
systematic random sampling
Create a sampling frame, and every nth unit select a participant every nth unit, start at a
random point
What is a skip interval (n)? How do you calculate it?
the number of participants you skip in a systematic random sample, population/sample
size
stratified random sample
, organize participants into strata homogenous, select randomly from strata
cluster sampling
organize participants into heterogenous clusters, select randomly from culsters
coverage error
a sampling error that occurs when sampling frame is flawed or outdated
sampling error
an error that occurs when a sample somehow does not represent the target population
How to reduce coverage error
up to date lists and sampling frame
How to reduce sampling error
draw enough people, have a representative sample
What is a confidence level?
How sure you are that results fall within a certain interval
Standard confidence level for social science/comms?
95%
What is the normal curve?
a symmetrical, bell-shaped curve that describes the distribution of many types of data.
The more random data you have, the more bell shaped it becomes
What is the peak of the bell curve?
The mean, 50th percentile, where most data points lie
How do you increase confidence level?
Increase number of units in the sample
What is a standard deviation?
How far is this data point from the mean?
What percent of data lies +/- one standard deviation of the mean?
68%, 34% on each side
What percent of data lies +/- two standard deviation of the mean?
95%, 34+13.5% on each side
What percent of data lies +/- three standard deviation of the mean?
99%, 34+13.5+2% on each side
What is margain of error?
determines the precision of accuracy of a poll result. actual results lie somewhere
between finding +/- mofe
What happens to mofe if you increase sample size?
decrease
You always have mofe unless
your sample size is your population
Is a poll actually able to predict an election winner if the ranges with mofe
overlap?
No
What is the acceptable sample size and mofe in social science?
384, 5%
Poll
quick version of a survey, 5 to 10 questions
What are some survey design considerations?
Goal, audience, method of delivery, question types and length,
