Unit 1 part 1
What is the difference between a multistage simple random sample and a stratified random sample?
While both simple random sampling methods use multiple stages to obtain a sample, stratified random sampling uses the division of specific groups or strata to obtain its sample, then, takes the SRS within each strata and uses population percentages as weights to get the total population results. Whereas multistage just takes a general sample, then uses several stages to select a specific sample from the larger sample. These are NOT the same processes.
How do you critique a meta-analysis?
Ask:
What studies were conducted?
Can they actually be compared or combined?
Calculate the margin of error for a sample of 625 participants
4%
1 / square root of 625
Is it Ethical to conduct research on primates that you cannot conduct on humans?
Yes, as long as the researchers follow proper guidelines as outlines by the animal version of the Institutional Review Board.
What is the difference between bias and precision?
1. Bias: the extent to which a parameter estimate - statistic – is systematically too high or too low in repeated sampling
2. Precision: the extent to which there is little variability in a parameter estimate from sample to sample . Large variability = decreased precision; small variability = increased precision
What is the formula for the CPI?
Value
CPI = --------------- X 100
Base Value
Which of the following describes a Probability Sample?
A. Every possible sample from the population has a known probability of selection
B. Every sample has an equal chance of selecting participants
C. Sampling frame is equal to the population size
What is Simpson’s Paradox?
The nature of an observed relationship between two variables can change – even reverse - when we take into account other unobserved variables (i.e., lurking variables).
What was so controversial about the first Vietnam draft lottery in 1970?
1. The drawing of those drafted was biased in selection by birthday, such that a majority of people drawn had birthdays clustered around each other—resulting in more deaths around these birthdays.
What are the four errors to watch out for when designing an experiment?
1. Confounding Variables
2. Interacting Variables
3. Placebo, Hawthorne, and Experimenter Effects
4. Lack of ecological validity and consequent generalization
The Standard Deviation is 10 and the Mean is 46. What is the equivalent of -2.5 SD in a normal distribution?
-2.5 SD = - 10 + - 10 + -5 = 21
What is the purpose of the CPI?
CPI measures a consumer's purchasing power from year to year and allows one to make meaningful over-time comparisons
What are the two parts of a confidence statement?
Margin of Error; Level of Confidence
What are AT LEAST two things wrong with CPI?
1. Does not reflect current spending priorities
2. As one item rises, consumers are likely to replace it with another item
3. Does not adjust for changes in quality
4. Does not take advantage of sale prices and online shopping
5. Does not measures prices for rural Americans (our measures are biased)
a. Wealthy
b. Healthy
c. Intelligent
d. Traditional
e. Educated
Describe the complex relationship between a sample and a population, and a statistic and a parameter. *You must use the proper terms.
A sample makes an inference about a population; a statistic is an estimate of a parameter; a sample is related to a statistic; a population is related to a parameter
What are the four major uses of the CPI?
1. Evaluate and determine economic policy
2. Compare prices to different years
3. Adjust to other economic data for inflation
4. Determine salary and price adjustments
4a. Value today = value of base x (CPI today / CPI base)
1. A Margin of Error asks the question: How close is the sample statistic to the population parameter over the sampling process (e.g., +/- 3%)
2. Whereas a confidence interval asks: What percent of all possible samples satisfy the margin of error? (e.g., 95% certainty)
What is the difference between the Margin of Error and a Confidence Interval?
1. A Margin of Error asks the question: How close is the sample statistic to the population parameter over the sampling process (e.g., +/- 3%)
2. Whereas a confidence interval asks: What percent of all possible samples satisfy the margin of error? (e.g., 95% certainty)
What are the advantages of doing an observational study over an experimental study? That is, when is an observational study more appropriate?
1. Good for medical field
1A. Randomization and ethical concerns and financial constraints alleviated
2. Very efficient
3. Reduces Potential Confounding Variables (Can choose controls to do this)
What is the difference between a sampling frame and a population?
While a sampling frame and a population may be equivalent in some cases, a population is the total number of individuals or objects you are evaluating, and a sampling frame is the actual number of individuals or objects you are able to evaluate.
What are the four key components of an experiment?
1. Use random assignment to conditions
2. Control effects of lurking variables through randomization to groups and by comparing several treatments (e.g. using a control group)
3. Replicate experiment by using many subjects per condition
4. Assess statistical significance – Rule out coincidence as explanation