Section 4.1: The Beauty of Sampling
Sample Survey: When a subgroup of a large population is simply questioned on a set of topics. (A type of observational study).
Margin of Error: The Measure of accuracy in a sample survey.
The sample proportion and the population proportion with a certain opinion or trait differ by less than the margin of error more than 95% of the time.
M.o.E:
*This is a Conservative Margin of Error (i.e. the actual margin of error is never greater than this estimate).
Confidence Intervals: For about 95% of properly conducted sample surveys, the interval,
(sample prop. –M.o.E., sample prop. + M.o.E)
will contain the actual population proportion.
Ex. In order to analyze the true ‘normal’ body temperature of United States citizens a simple random sample was taken of 106 people. The sample mean was 98.2 Degrees. What was the Margin of Error for this study? Construct a 95% confidence interval for the true body temperature of U.S. citizens.
Advantages of a Sample Survey:
- A census isn’t always possible (Especially if the population is large).
- Quick to Implement. Results are immediate and can be released before they are out-of-date.
- Results can be more accurate because researchers can focus on surveying the small sample properly.
Bias:
- Selection Bias: When the method used for selecting the participants produces a sample that does not represent the population of interest.
Ex. People at an airport are surveyed to determine their attitudes about the price of current flights.
- Nonresponse Bias: When a representative sample is chosen for a survey, but a subset cannot be contacted or does not respond.
Ex. A survey conducted by stopping people outside of the Tivoli during the lunch hour would omit people who work during the day and take night classes.
- Response Bias: When participants lie or respond differently than they actually feel.
Ex. How many times have you consumed alcohol within the last 2 weeks?
(Is your response affected by the fact that your professor is in the room? People you don’t know are in the room? People you do know are in the room?)
Section 4.2: Simple Random Sampling and Randomization
**Don’t allow humans to play a part in the random number generator**
**Tables of Random Numbers, Calculator Programs, Computer Programs, Statistical Software**
Section 4.3: Other Sampling Methods
Stratified Random Sampling: We use this method when we are concerned about differences among subgroups, or strata, within a population. First identify the subgroups and then draw a simple random sample within each subgroup. The total sample consists of all the samples from the individual subgroups.
Ex. Political Elections
Strata: Men, Women
Strata: 18-22, 23-30, 31-40, 41-50, 51-60, over 60
Strata: Seniors, Juniors, Sophomores, Freshman
Cluster Sampling: Population unites are divided into groups, called clusters, but rather than sampling within each group we select a random sample of clusters and measure only those clusters.
Ex. To measure customer satisfaction of RTD customers the department of transportation randomly samples a set of bus rides and distributes a survey to everyone on those bus rides. Each bus is a cluster.
**Caution must be taken during analysis because similarities will inevitably arise amongst clusters.**
Systematic Sampling: Use a simple system to choose the sample, such as selecting every 10th or every 50th member of a population.
Ex. In order to attain a sample of Auraria Campus students, a researcher went to the registrars office during the first day of classes and surveyed every 5th person in line after starting with the 2nd individual in line.
Random-Digit Dialing: Provides a simple random sample of all households in the U.S. that have homes.
**Once a number has been chosen to be included in the simple random sample, multiple attempts must be made to reach the individual at that number**
**This method may tend to include an over-representation of females.**
Multi-Stage Sampling: When a combination of sampling methods are used.
Ex. First randomly select a sample of counties from all 50 states, then randomly select cities and towns in those counties, then randomly select residential blocks in each city or town, then randomly select households in each block, then randomly select someone from each household.
Section 4.4 Difficulties and Disasters in Sampling
Sampling Frame: The list of all units from which the sample is selected.
**Beware of using the wrong sampling frame**
Ex. Predicting election results from a list of registered voters.
**Not Reaching the Individuals Selected**
-Can you ever be sure that a survey conducted by mail was completed by the intended unit?
-Telephone surveys tend to reach more women.
News-media Complicate Statistics with ‘Quickie Polls’
Ex. The United States captured the leader of a terrorist organization at 1:00 in the morning. T.V. and news journalists conduct quick polls to get the ‘public opinion’ for the morning paper/television show.
Problems:
- Questions are quickly put together, not tested, and poorly presented.
- How do you get a simple random sample between 1:00 and 5:00 in the morning.
- Convenience Sampling is often utilized.
Non-response or Volunteer Response:
If 75% of intended survey participants respond…the survey was incredibly successful.
Ex. The Nightline Opinion Poll
Convenience Sampling or Haphazard Sampling: Sampling is done in such a way that is convenient to the researcher and disregards simple random sampling.
Ex. Yelling out your window to the first five people that pass by “What is your opinion on the death penalty?”
Ex. Surveying students eating lunch in the food center of the Tivoli. Surveys will tend to approach people whom they think will support the results of their study or whom look friendly and easily approachable.
Section 4.5: How To Ask Survey Questions
Beware Of:
- Deliberate Bias: Using the wording of questions to achieve the desired response.
Ex.
Do you agree that the president has unfair health care policies for the elderly?
Do you agree that the president has made great strides for our country with his health care policies for the elderly?
Ex. Have you quit smoking in the last 6 months?
- Unintentional Bias: A researcher unintentionally words a question in such a way that the subject is confused or misunderstands the question.
Ex. In the last 6 months have you used any drugs?
- Desire to Please: People tend to respond in a manner which will please the person who is asking the question.
Ex. Someone come up to you who is wearing a Register to Vote t-shirt. They ask you if you are registered to vote. What do you say?
- Asking the Uninformed:
Who would you rather see as the next president of Israel:
Moshe Kastev, Ariel Sharon, or Elij Wolyam
- Unnecessary Complexity:
“Do you agree with the University’s new credit policy, since it ensures that all UCD students have a well-rounded education.”
- Ordering of Questions
What did you eat for dinner last night?
Would you consider yourself a healthy individual?
- Confidentiality and Anonymity: Who is going to find out the answer to your responses.
**(Often follow-up surveys are a necessity…so confidentiality can be ensured but anonymity is less-likely.)
What is Being Measured?
- Some concepts are hard to define.
Ex. How polite was your cashier yesterday?
- Open or Closed Questions: Should Choices be Given?
Ex.
What are you currently paying too much money for?
What are you currently paying too much money for?
Rent:
Heat:
Food:
Entertainment:
Child Care:
Other: