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;
13 Cards in this Set
- Front
- Back
Type 1 error (Decision Errors)
|
Reject null, when null is true
|
|
Alpha level (Decision Errors)
|
probability of making a type 1 error, held at .05
|
|
Type 2 error (Decision Errors)
|
retaining null when null is false
|
|
Beta level (Decision Errors)
|
probability of making a type 2 error
|
|
Power
|
correctly rejecting the null hypothesis
1- Beta (probability of rejecting the null when the null is false) calculated through being given cohen's d and the sample size of either an independent or dependent samples t-test. |
|
Effects of increasing and decreasing alpha
|
increased: beta becomes smaller, and power increases
decreased: beta becomes larger, and power decreases "Decreasing alpha to .01 decreased the power to detect a mean difference of .2 standard deviations with a sample size of 30 for an Independent Samples t-test. The power is now approximately 2.13%" |
|
cohen's d
|
absolute distance between means in terms of standard deviations.
|
|
effect size/true mean difference
|
increases: power increases
because it decreases beta which makes it less likely to make a type 2 error |
|
effect size (small, medium, large)
|
cohens d = .2 means an 85% overlap in the null and alt hypothesis graph
cohens d=.5 means a .67 percent overlap cohens d=.8 means a 53 percent overlap The larger cohen's d (distance between means), the less overlap. |
|
Changing sample size
|
has an effect on the sampling distribution of the mean
(increasing sample) size: less variance in scores, means smaller standard error, which (increases power) |
|
how to make a certain power possible
|
use power to find the non centrality parameter and then plug in the numbers to the non centrality parameter equation
|
|
standard desired power
|
.8
|
|
Over lap (cohen's d/null and alt. graphs)
|
high overlap between the null and alternative hypothesis means the means are closer together and more similar than different.
high overlap = low significance of effect low overlap = high significance of effect because the graphs are farther apart, meaning cohen's d is telling us that the means are too far apart to be from the same or similar distributions. |