Term
|
Definition
| -appears without exception; has become consolidated by repeated successful testing; often expressed in concise verbal or mathematical statement |
|
|
Term
|
Definition
| -scientifically accepted general principle supported by substantial body of evidence; provides explanation of observed facts/events and basis for future investigation |
|
|
Term
|
Definition
| -a specific statement or question to be evaluated |
|
|
Term
| Does one study ever prove anything? |
|
Definition
|
|
Term
| True or False: Scientifically, we cannot prove anything. |
|
Definition
|
|
Term
| Definition: Null hypothesis |
|
Definition
| -there is no linkage b/n the things we are researching |
|
|
Term
| Definition: Independent variables |
|
Definition
|
|
Term
| Definition: Dependent variables |
|
Definition
|
|
Term
|
Definition
| -existence of difference (association, disease, etc) |
|
|
Term
| What are the 4 possible outcomes of a study? |
|
Definition
1) The study detected a difference when it existed (correct conclusion)
2) Type I error: rejected null hypothesis when it was true (did not detect a difference that was there)
3)Type II error: failed to reject null hypothesis when it was false (detected a difference when there wasn't one)
4) didn't detect a difference when there wasn't one |
|
|
Term
| The probability finding is due to chance and is compared to pre-established _______ to determine significance. |
|
Definition
|
|
Term
|
Definition
| -probability of a type II error |
|
|
Term
|
Definition
| -Power, the probability of rejecting null hypothesis when it is false |
|
|
Term
| How can a wrong conclusion be made in a well-designed study? |
|
Definition
| -by using a small group of subjects to represent the population of interest |
|
|
Term
| What is the significance of the bellcurve? |
|
Definition
| -for many characteristics in the world, there will be an average value and variation in the population (standard deviation) |
|
|
Term
| How should one go about selecting a sample to avoid bias? When is this altered? |
|
Definition
-best way is random selection
-altered is you must block a specific factor to not skew the results (age, gender, breed, location, etc) |
|
|
Term
| How do we extimate the probability of finding a difference this larger, or larger, if the null is true? |
|
Definition
|
|
Term
| How do we determine if findings are significant? |
|
Definition
| -compare calculated P-value to pre-determined alpha threshold (0.05) |
|
|
Term
Explain what happens to the power if the following occurs:
a)higher alpha
b)stronger association
c)larger sample size
d)larger variation among population |
|
Definition
a)increases power
b)increases power
c)increases power
d)decreases power |
|
|
Term
| Are all statistically significant findings clinically relevant? |
|
Definition
| -no, small differences may e found that would not change your clinical judgement |
|
|
Term
| Are all clinically relevant findings statistically significant? |
|
Definition
| -no (low power of study), there are important differences that WOULD alter your clinical judgment that could be missed |
|
|
