Statistical test selection, Interpretation of the evidence

Question 1

Use of ___ and ____ tests for testing hypotheses about

Answer 1

z
t
single sample means

Question 2

Exist numerous statistical tests to

Answer 2

analyze statistical significance of data

Question 3

Most research will use

Answer 3

a combination of descriptive and inferential statistics

Question 4

Characteristics of the data determine the

Answer 4

appropriate inferential statistics

Question 5

Selection of the most appropriate test is determined by: ______ of the measurement to ______ data

Answer 5

scale
obtain

Question 6

Selection of the most appropriate test is determined by: number of

Answer 6

groups

Question 7

Selection of the most appropriate test is determined by: if measurements are on

Answer 7

independent subjects or related samples

Question 8

Selection of the most appropriate test is determined by: _________ involved in statistical test

Answer 8

assumptions

Question 9

STUDY CHARTS IN SLIDE DECK FOR SELECTION OF APPROPRIATE TESTS

Answer 9

STUDY CHARTS IN SLIDE DECK FOR SELECTION OF APPROPRIATE TESTS

Question 10

Tests for nominal and ordinal data =

Answer 10

non-parametric or distribution free

Question 11

Non-parametric require

Answer 11

few or no assumptions about population distributions

Question 12

Tests for analyzing interval or ratio data =

Answer 12

parametric

Question 13

Parametric requires

Answer 13

assumptions for populations from which samples are drawn

Question 14

Interpretation of the evidence: important distinction between

Answer 14

statistical significance and clinical significance

Question 15

In statistical significance the results are

Answer 15

because of chance or reflect real population trends

Question 16

Interpretation of the evidence: researcher assembles ________, calculates __________ and determines _________

Answer 16

data
value of statistic
if results statistically significant

Question 17

Effect size expresses

Answer 17

the size of the change or degree of association that can be attributed to a health interventions

Question 18

Even if the results are statistically significant, effect sizes may

Answer 18

not be clinically significant or theoretically interesting

Question 19

Statistical power analysis is useful to know

Answer 19

how likely a miss (type II error) is to occur

Question 20

Statistical power of a statistical analysis is defined as

Answer 20

Power = 1 - beta (probability of a miss)

Question 21

Contemporary health research requires a statistical power analysis of at least

Answer 21

0.8

Question 22

5 factors affecting power

Answer 22

size of difference between means
significance level
sample size
variance
other factors (normality, statistical procedure, one-or two-tailed)

Question 23

Factors affecting power: small sample, analysis will have

Answer 23

low power

Question 24

The best defense against low power is

Answer 24

a good sized sample

Question 25

What enables us to conduct power analyses for estimating the minimum sample size for detecting if an effect is really true

Answer 25

the evidence from previous research and the results of pilot studies

Question 26

How to interpret findings: clinical and statistical significance

Answer 26

clear, strong evidence for treatment effect

Question 27

How to interpret findings: no statistical significance, yes clinical significance

Answer 27

inconclusive; suggests findings might be meaningful, need further research

Question 28

How to interpret findings: yes statistical significance, no clinical significance

Answer 28

clear, strong evidence for lack of treatment effect

Question 29

How to interpret findings: no statistical significance, no clinical significance

Answer 29

inconclusive; need for further research to determine meaningfulness of the findings

Question 30

Relationship between effect and sample size: Large effect size, high sample size

Answer 30

both statistical and clinical significance are likely to be demonstrated

Question 31

Relationship between effect and sample size: large effect size, low sample size

Answer 31

statistical significance would be likely, but the results might not indicate clinically applicable outcomes

Question 32

Relationship between effect and sample size: small effect size, high sample size

Answer 32

statistical significance might be demonstrated, but clinical significance is unlikely

Question 33

Relationship between effect and sample size: small effect size, low sample size

Answer 33

neither statistical nor clinical significance is likely; statistically significant results might result in a type I error

Question 34

Clinical decision making

Answer 34

clinician decision procedures for making diagnosis on the basis of uncertain information are similar to the scientist's hypothesis-testing procedure

Question 35

Like researcher, clinician faces decision with regard to

Answer 35

type I (false alarm) or type II (miss) error

Question 36

What type of error is generally more acceptable for clinician

Answer 36

type I error (false alarm) more acceptable than type II (miss)