Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
34 Cards in this Set
- Front
- Back
|
What do observational studies do? |
They observe |
|
|
What does a retrospective study do? |
It examines data from past observations |
|
|
What is a lurking variable? |
A variable that is not readily noticeable |
|
|
How do retrospective studies get errors? |
Fro imperfectly collected historical data |
|
|
What is Prospective Study? |
A study which involves identifying subjects in advance to study |
|
|
Can Observational studies make causal relationships? |
No |
|
|
How to establish a causal relationship? |
Through experiments |
|
|
What does an experiment require? |
Random assignment of subjects to treatements |
|
|
How do experiments study variables? |
Through the relationship between multiple variables |
|
|
What is a Factor? |
An explanatory variable that can be manipulated |
|
|
Why is an experiment different from other studies? |
Active and deliberate manipulation of factors to control details of possible treatements |
|
|
What does the experimenter observe for experiments? |
The response variable compares to responses to different groups |
|
|
What are Experimental units or Subjects? |
Individuals that are experimented on |
|
|
What is a Level? |
The specific values used for a Factor |
|
|
What is the Treatment? |
The combination of specific levels from factors |
|
|
What are the Four Principles of Design? |
Control, Randomize, Replicate, Blocking |
|
|
What is the Control principle of design? |
Make conditions as similar as possible, reduce variation to protect response variable |
|
|
What is the Randomize principle of design? |
Randomization equalizes effects of unknown or uncontrollable variation, need randomness |
|
|
What is the Replicate principle of design? |
Outcome of single experiment anecdote, replicate with controlled sources of variation at different levels essential |
|
|
What is the Blocking principle of design? |
Compromise between random/control, can recognize similarities in individuals, group them, then randomize within groups, reduce variability due to differences among blocks |
|
|
When is a result Statistically Significant? |
When there is a big difference between different treatments |
|
|
What does a Sample Survey do? |
Tries to estimate pop parameters, sample needs to represent whole pop |
|
|
What do experiments do? |
Assess effect of treatments |
|
|
How to treat result of experiment? |
Do not generalize until experiment repeated many times on different groups |
|
|
When do biases cancel out? |
When all biases point the same way |
|
|
What is a Control group? |
The group given a baseline treatment, is another level of the factor |
|
|
What is Blinding? |
When individuals of one class are blinded, experiment is single blinded |
|
|
What is Double-Blinding? |
When both experimenters and subjects are blinded |
|
|
What does blinding do? |
Disguises treatment info that would bias |
|
|
What is a Placebo? |
Fake treatment that hides whether subject is control or not |
|
|
What is Randomized Block Design? |
When randomization occurs within blocks |
|
|
What is Confounding? |
When levels of a factor are associated with levels of another factor |
|
|
How to tell if Lurking variable? |
Lurking variable usually associated with factors that make it appear as if one causes another |
|
|
How to tell if Confounding variable? |
May have had effect on response, cannot tell what cause effect, our factor or confounding |
