Term
|
Definition
| Categorical data, represent frequency distribution among different categories |
|
|
Term
|
Definition
| most common method to represent trends over time |
|
|
Term
|
Definition
| contains 67% of the values |
|
|
Term
|
Definition
Used to compare 2 sets of continuously
distributed data
Means of the 2 sets of data must be KNOWN |
|
|
Term
|
Definition
Used to compare two different sets of
continuously distributed data with unknown means |
|
|
Term
|
Definition
| Both normal distributions. Z-distribution is tighter; t-distribution is broader. T-tests require greater difference than z to be statistically significant |
|
|
Term
|
Definition
|
|
Term
|
Definition
Test used when the Chi-Square is
unstable due to small numbers.
(Chi-Squared needs to have at least “1”
in each cell and <20% of cells with <5.) |
|
|
Term
|
Definition
| Analogous to t-tests, but with >2 samples. Assesses multiple independent populations. Used in laboratory studies when there are multiple measures |
|
|
Term
|
Definition
Test used when the t-test cells are <5.
(If cells are 0, or 1, there is a major
problem in calculating the results)
(F-distribution is not normally
distributed) |
|
|
Term
|
Definition
Test used when data are paired.
Otherwise, analogous to the Chi-Square |
|
|
