week 6 in class

Question 1

RCT Methodological Considerations

Answer 1

hypothesis - research question should be restated in statistical language.

Example: “ is there a difference in GPA by sex?” - t-test type of question

is there a relationship between GPA and income level? - correlation type of question.

Question 2

population

Answer 2

description of who was studied

Question 3

interventions/instruments

Answer 3

measurement reliability - like kappas/ICCs

Question 4

inferential statistical testing

Answer 4

used to answer questions concerning comparisons or relationships

Question 5

why is descriptive statistics important?

Answer 5

baseline comparability and generalizability

Question 6

to answer inferential statistical questions

Answer 6

decide whether observed difference between samples is due to “chance” or represents a “real” difference between populations

Question 7

null hypothesis

Answer 7

a difference does not exist between two samples being compared

Question 8

hypothesis

Answer 8

a difference does exist between two samples being compared

Question 9

high p value =

Answer 9

high probability of chance difference

Question 10

low p value =

Answer 10

low probability of chance difference

Question 11

type I error-

Answer 11

alpha

Question 11

type II error -

Answer 11

beta

Question 12

ratio =

Answer 12

between group means / variability within groups

Question 13

the bigger the ratio -

Answer 13

the smaller the p value

Question 14

statistical power -

Answer 14

is the probability that a test will lead to rejection of the null hypothesis

Question 15

power -

Answer 15

avoids missing a positive finding when it actually exists

Question 16

to use a parametric test -

Answer 16

measured variables must be normally distributed
samples must be drawn at random
variances in the samples being compared must be equal
data must be measured on interval or ratio scales

Question 17

nonparametric tests used when -

Answer 17

data is not normally distributed (skewed)
used when variance between samples is not equal
developed to operate on data expressed in the nominal and ordinal scales

Question 18

independent samples -

Answer 18

when two or more groups consist of completely different individuals

Question 19

dependent samples -

Answer 19

when the same individuals are tested more than one time

Question 20

t-tests

Answer 20

used to compare the means of 2 groups

Question 21

t statistic =

Answer 21

difference between group means / variability within groups

Question 22

parametric tests -

Answer 22

ANOVA and repeated measures ANOVA
For comparing three or more samples or groups

Question 23

f ratio =

Answer 23

treatment/ error variance

the larger the f-ratio the greater the difference between the group means relative to the variability within groups.

the bigger the f ratio the smaller the p value for the comparisons

Question 24

parametric tests

Answer 24

one-way ANOVA

Question 25

two-way ANOVA

Answer 25

2 independent factors

Question 26

turkey test -

Answer 26

decreased risk of type I error

Question 27

newman-keuls test -

Answer 27

more risk of type I error

Question 28

bonferroni t-test -

Answer 28

greatly decreased risk of type I error

Question 29

scheffes comparisons -

Answer 29

decreased risk of type I error, but less power

Question 30

data normally distributed -

Answer 30

then parametric tests are used

the t-test, analysis of variance, person correlation

Question 31

data not normally distributed -

Answer 31

then non-parametric tests are used

the Mann-Whitney U test, spearman correlation