Term
| descriptive statisticsalpha and the number of tails. If alpha = .05 and we are using a two-tailed test, then the critical value is 1.96 (from the Z tables). |
|
Definition
| the science of describing distributions of samples or populations |
|
|
Term
|
Definition
| the science of using sample statistics to make inferences or decisions about population parameters |
|
|
Term
| How do we show a statement is most likely false? (5 steps) |
|
Definition
1. Create a statement (null hypothesis)
2. Create a distribution of outcomes given that hypothesis
3. Determine outcomes we consider unlikely
4. Evaluate if our experimental result is one of those unlikely outcomes
5. two possible outcomes |
|
|
Term
| what are the two possible outcomes to see if a statement is most likely false? |
|
Definition
1. reject null hypothesis
2. fail to reject null |
|
|
Term
| What does it show if we reject our null? |
|
Definition
| we have shown evidence that it is likely to be untrue |
|
|
Term
| How likely is it that we are wrong when we reject our null? |
|
Definition
| The size of the rejection region - alpha |
|
|
Term
if null is true in reality..
Probability of error, and no error |
|
Definition
Error = alpha
no error = 1 - alpha |
|
|
Term
| if null is not true, error if we fail to reject, and correct if we reject |
|
Definition
probability error = beta
probability correct = 1 - beta |
|
|
Term
| beta is much harder to calculate than α because it depends on how distributions overlap which is affected by... |
|
Definition
effect size
sample size
stand dev
alpha
tails of test |
|
|
Term
| two tail - non directional |
|
Definition
no prior evidence
H0: u=a
H1: u NOT =a |
|
|
Term
|
Definition
prior evidence u>. a
H0: u lessthen/= a
H1: mu greatr than a |
|
|
Term
|
Definition
prior evidence u less than a
Ho: u greater than or = to a
H1: u less than a |
|
|
Term
| What happens in TYPE 1 errors? |
|
Definition
| Other researchers use this false result as information, Fail to replicate, Waste time |
|
|
Term
| What happens in Type II error? |
|
Definition
| No research report is created, Not much loss to the scientific community, Loss to researcher, Other researchers can still find the effect |
|
|
Term
type I error?
-occurs with what probability |
|
Definition
When you reject your null and in reality it is true
-occurs with probability of alpha |
|
|
Term
type II error?
-occurs with what probability |
|
Definition
When you accept/fail to reject your null, and in reality it is false
-occurs with probability of beta |
|
|
Term
| When do you reject your null? |
|
Definition
| if the probability of the sample result is <= to alpha (significance) level |
|
|
Term
| When do you accept your null? |
|
Definition
| if probability of sample result is above alpha |
|
|
Term
| hypothesis testing calculates |
|
Definition
| the probability that the differences are due to random variability (i.e., the first option or the Null Hypothesis) |
|
|
Term
|
Definition
| a small probability that the null hypothesis will be rejected when random variability is the cause of the differences in the conditions |
|
|
Term
|
Definition
|
|
Term
|
Definition
| alpha, effect size, sample size, std dev, tails of test |
|
|
Term
| comparison distributions are used.. |
|
Definition
| are used to compute the probability of the obtained sample result under the assumption that the null hypothesis is true |
|
|
Term
| when are scores related? (dependent) |
|
Definition
-when same person is on both conditions
-when scores are matched on a meaningful variable (steroids and weightlifters, GPA and SAT scores) |
|
|
Term
| obstacles to using dependent |
|
Definition
-if impossible to participate in both conditions (normal vs parkinsons, men vs women)
-if participation in one condition affects other conditions (conjunction fallacy, solving a reasoning problem) |
|
|
Term
| why are dependent tests preferred? |
|
Definition
| preferred because they are better at detecting differences due to a reduction in variability |
|
|
Term
| when are independent tests used? |
|
Definition
-when the subjects in one group are not related to the subjects in the other group
-when there is a possibility that participating in one condition can affect the measurement of the other condition |
|
|
Term
| why are dependent tests more powerful than independent |
|
Definition
| because they eliminate the variability between the pairs of scores |
|
|
Term
| why are power calculations used? |
|
Definition
| used to find the probability that your experiment will detect an effect (or difference) in the world, if it exists |
|
|
Term
|
Definition
| he probability of selecting a mean from Mean Distribution 2 and having that mean be above the critical value for distinguishing it from Mean Distribution 1 |
|
|
Term
| 3 important factors in determining area of power |
|
Definition
-where the critical value is
-how far the distributions are away from each other
-how skinny the distributions are |
|
|
Term
| effect size is determined by |
|
Definition
| the distance between the two distributions |
|
|
Term
| the critical value is effected by |
|
Definition
| alpha and the number of tails. If alpha=.05 and we are using a two-tailed test, then the critical value is 1.96 (from the Z tables). |
|
|
Term
| the width of the mean distributions is affected by |
|
Definition
| the standard deviation of the score distribution and the sample size which determines how much skinnier the mean distribution is than the score distribution |
|
|
