Chapter 1 Flashcards
What is Sampling Bias?
Sample was not representative of the population
What is Non-response bias?
A lack of response from a significant amount of the population.
Simple Random Sample
-every member of the population has the same chance of being included in the sample (sample is representative of the population)
-members of the sample are chosen independently of each other
Ascertainment (or sampling) bias
some members of the population are less likely to be included than others
What are two examples of nonsimple random sampling?
Random Cluster Sampling and Stratified Random Sampling
What is a random cluster sample?
-divide forest in to smaller geographical sectors
-take a simple random sample of sectors
-count all the butterflies in these sectors and scale appropriately
What is a stratified random sample?
-divide population into groups based on age, income, etc
-collect simple random samples of each group and scale appropriately
What are some of the challenges to random cluster sampling and stratified random sampling?
-outliers propagate too large and there are different ways to make groups
What is anecdotal evidence?
-a small observation or story
What is an observational study?
-the researcher collects more data but they only an observer, they do not manipulate conditions or do an experiment
What is a confounding variable?
-an extra variable that was not accounted for
Which treatment is better?
Treatment A is better for both small stones and large stones because it is important to look at the sample size. Even though doctors favor treatment B for small stones and treatment A for large stones.
What is it called when the doctor’s choice of treatment affected results?
Simpson’s Paradox (also a confounding variable)
What is a randomized control experiment?
-patients should be randomly assigned to either one group that get the new drug or treatment or another group that get the placebo (control)
What is a double-blind experiment?
-the patient shouldn’t know whether they’re getting the drug or the placebo (blind)
-the health professionals also shouldn’t know which patients are getting the drug or the placebo (double-blind)
What are case-control studies?
-take one sample of people with cases and find another sample of people who do not have cases (control)
-the controls must have other factors the same as the cases so there are no confounding variables
What is a historical control?
-compare new procedure to existing data for old procedure
What is a longitudinal (or cohort) study?
-data is gathered for the same subjects over a period of time (often many years)
What are some difference between case-control studies and longitudinal studies?
-In case-control you can get data right when you want it while in longitudinal you ave to wait for data and people can drop out of the study
When you want to study a rare disease which one is better case-control or longitudinal?
Case control because then you do not have to wait to see who will develop the disease
What is usually cheaper and quicker case-control or longitudinal?
case-control
How do they measure illegal drug use?
self reporting; hospital overdose; police arrests; rehab centers; sewer system
How do they measure homeless population?
unsheltered people, police complaints or arrests, divide into sectors and have people count
Why is counting and making figures not always obvious?
-there are decisions to make and these decisions matter
