Psych 301, Lecture 8

Front 1

Inferential statistics

Back 1

used to infer what the effect is for the whole population of insomniacs

Front 2

Descriptive statistics

Back 2

describe our data (these 10 people)

Front 3

Sampling Error

Back 3

sample doesn't represent the population completely. sample has unique characteristics

Front 4

Sampling error's relationship to probability

Back 4

how much of the probability is due to sampling error? or is it really something else?

Front 5

Primary goal of hypothesis testing

Back 5

use a sample to test hypothesis about the relationship between variables in the population

Answers the question: how likely is it that our results occurred by chance?

Front 6

Alternative Hypothesis (H1)

Back 6

when people take our new drug, they will sleep more

Front 7

Null Hypothesis (Ho)

Back 7

logical counterpart to the alternative hypothesis
*if Ho is false, H1 is true
*If Ho is true, H1 is false
**Ho and H1 are mutually exclusive

Front 8

Hypothesis Testing

Back 8

If H1 is nondirectional, Ho will be nondirectional
If H1 is directional, Ho will be directional

Front 9

Hypothesis testing

Back 9

We always evaluate the reasonableness of the null hypothesis
*Critical point: we don't directly evaluate the alternative hypothesis

Front 10

Hypothesis testing

Back 10

tell us how likely it is that our results are due to chance; reject or retain the null hypothesis

Front 11

alpha level

Back 11

can be any probability (most common .05 and .01)
-if the probability of our results occurring by chance are less than alpha, then we reject the null
-if the prob. of our results occurring by chance exceed alpha, then we can't reject the null

Front 12

Sign test

Back 12

-an inferential statistic, ignores the magnitude of difference scores and considers only their direction and sign

*how many pluses and minuses?

Front 13

Errors

Back 13

in reality, Ho is either true or false, we either reject or retain the Ho

Front 14

Type 1 error

Back 14

a decision to reject the null hypothesis when the null is true

Front 15

Type 2 error

Back 15

decision to retain the null hypothesis when the null is false

Front 16

setting alpha to .05

Back 16

willing to limit the probability of rejecting the null hypothesis when it's true to 5 times in 100, limits the probability of making a type 1 error to .05

Front 17

directional hypothesis

Back 17

more specific, indicates which group will score higher/lower

ex) when people take our new drug they will sleep more

Front 18

nondirectional hypothesis

Back 18

ex) when people take our new drug, it affects their sleep

Front 19

directional hypothesis

Back 19

1 tailed test

Front 20

nondirectional hypothesis

Back 20

2 tailed test
-one group can be less than OR greater than the other group
-you divide alpha in to 2 equal parts of the distribution (1 part for each of the tails)

Front 21

Nondirectional hypothesis

Back 21

evaluate the probability of observing the obtained results or any more extreme result in both directions (both tails)

Front 22

Rule of thumb

Back 22

the evaluation should always be 2 tailed unless the experimenter will retain Ho when results are extreme in the direction opposite to the predicted direction