This site is 100% ad supported. Please add an exception to adblock for this site.

Cards for chap 6 Agresti and finlay text


undefined, object
copy deck
What is a signficance test?
A significance test is a way of statistically gesting a hypothesis by comparing the observed data to the predicted values.
What are the five elements that all statistical tests have?
1. assumptions
2. hypotheses
3. test statistic
4. P-value
5. conclusion
All statistical tests require certain assumptions for the test to be valid. What are the 4 assumptions?
1. type of data (qual. or quan)
2. pop. distribution (normal?)
3. method of sampling (simple random, for instance)
4. sample size.
What are the type types of hypotheses that a significance test considers about the value of a parameter?
1. Null hypothesis (no effect)
2. Alternative hypothesis (some effect)
What is another name given to the alternative hypothesis?
- The research hypothesis.
What is the test statistic?
The test statistic is the statistic calculated from the sample data to test the null hypothesis. It usually refers to a point estimate of the parameter.
What is the P-Value
The P-Value is the probability that, if the null hypothesis were true, that the test statistic would fall in this collection of values.
If one had a very small P-value what could be inferred about the null hypothesis?
A small P-value indicates that the data contradicts the null hypothesis (it can be rejected). Large P-values support the null.
What P-value do most studies require in order to reject the null?
Most require a P-value of P=.05
What would a P-Value of .83 indicate? How about .0002
.83 - keep null hypothesis
.0002 - reject null!
What is the alpha-lvl?
The alpha lvl is a number determined before testing that says at what probability lvl the null hypothesis is rejected.
What is the alpha lvl also called?
The alpha lvl is also called the significance lvl of a test
What are the most common alpha levels?
Alpha = 0.05
Alpha = 0.01
Remember - the smaller the alpha lvl, the stronger the evidence must be to reject the null.
Do A&F think it is better to:
a) report the P-value; or
b) say that a result is statistically significant at a certain alpha lvl.
A & F think it is better to report the p-value because they argue it is artificial to simply call one result significant and the other not.
Explain what Type I and Type II errors are.
Type I - when null is rejected even though it is true;
Type II - when null is not rejected even though it is false.
How are Type I and Type II errors related, if at all?
Type I and Type II errors are inversely related. As one decreases the other increases.
When might the probability of committing a Type II error be rather large?
If the sample size is small there is a greater chance that one would commit a Type Ii error (ie: the null is not rejected though it is false)

Deck Info