Flashcards in RESEARCH EXAM 3--ch 12 & 13 Deck (136):

1

## statistical analysis examples (2)

### descriptive statistics and inferential statistics

2

##
o Used to DESCRIBE and synthesize data

### descriptive statistics

3

##
o Used to make inferences/objective decisions about the population based on parameters using sample data

### inferential statistics

4

## descriptive indexes examples (2)

###
parameter

statistic

5

##
o A descriptor for a population

(The average or percentage of age of menses for American females)

### parameter

6

##
o A descriptor for a sample, a descriptive index

o the average age of menses for female professors at MSU

### statistic

7

##
ϖ A systematic arrangement of numeric values on a variable from lowest to highest, and a count of the number of times (and/or percentage) each value was obtained or has occurred

### freq. distributions

8

## freq. distributions can be presented in which 2 ways

###
a table

graphically (frequency polygons)

9

## freq distributions can be described in terms of:

###
shape

central tendency

variability

10

##
− When folded over the two halves of a frequency polygon would be superimposed

MIRROR IMAGES OF EACH OTHER

### symmetric shape

11

##
− Peak is in the center and one tail is longer than the other

### skewed (asymmetric)

12

##
long tail drifts off to the right

### − Positive skew

13

##
− long tail drifts off to the left

### negative skew

14

## # of peaks

### modality

15

## 1 peak

### unimodal

16

## 2 peaks

### biomodal

17

## 2+ peaks =

### multimodal

18

## normal distribution = what shape

### bell-shaped curve

19

## characteristics of normal distribution (bell shaped curve)

###
♣ Symmetrical

♣ Unimodal (1 peak)

♣ Not very peaked, not too flat

20

## normal distribution (bell shaped curve) are an important distribution in _____ statistics

### inferential

21

##
• Indexes of “typicalness” of a set of scores that comes from CENTER of the distribution

### central tendency

22

## Researchers avoid using the term “average” because there are three indexes of central tendency:

###
• the mode, the median, and the mean

23

## most frequent

### mode

24

##
the point in a distribution (middle) above which and below which 50% of cases fall

###
median

(NOMINAL MEASURES)

25

##
• equals the sum of all scores divided by the total number of scores

###
mean

(SKEWED DISTRIBUTION)

26

##
o Useful mainly as gross descriptor, especially of nominal (e.g., gender) measures

### MODE

27

##
o Useful mainly as descriptor of typical value when distribution is skewed (household value)

### median

28

##
♣ Preferred when a distribution is highly skewed

### median

29

##
o Most stable (best) and widely used indicator of CENTRAL TENDENCY

### mean

30

##
o the Mean Can do a lot of further analysis such as calculating _____ statistics

### inferential statistics

31

##
• The degree to which scores in a distribution are spread out or dispersed – how scattered numbers are

### variability

32

##
− Used to describe one variable at a time

### univariate

33

##
o Little variability

###
homogeneity

TALLER AND SKINNIER TABLE

34

## great variability

###
heterogeneity

WIDER AND SHORTER TABLE

35

##
• Indexes of variability describes how different the scores were such as homo or heater variability by using what 2 things:

### range and SD

36

##
o Highest value minus lowest value

### range

37

##
o Average amount of deviation (variability) of values from the mean

### standard deviation (SD)

38

##
o _____ the number in the standard deviation, the more variable the sample’s variables were

### Bigger

39

##
o tells use how much, on average, the scores deviate from the mean

### SD

40

##
o In a normal distribution, 95% of the scores fall within

### 2 SDs of the mean

41

##
ϖ Used for describing the relationship between TWO variables

don't answer research questions

###
Bivariate Descriptive Statistics

42

## 2 common approaches for bivariate descriptive statistics

###
crosstabs and

correlation coefficients

43

## crosstabs =

### contingency tables

44

##
• A two-dimensional frequency distribution; frequencies of two variables are

### cross-tabulated

45

##
at intersection of rows and columns display counts and percentages

### • “Cells”

46

## crosstab variables are usually ____ or _____

### nominal or ordinal

47

##
o whether there is a relationship between smoking and sex)

### crosstabs

48

##
• Indicate direction and magnitude of relationship between two variables

###
o CORRELATION COEFFICIENTS

49

## The most widely used correlation coefficient is

### Pearson’s r

50

## Pearson’s r is used when

###
• both variables are interval or ratio-level measures

51

## • Correlation coefficients can range from

### -1.00 to +1.00

52

## the higher the correlation coefficients the ____ the relationship

### stronger

53

##
− Negative (inverse) relationship ranges from?

### (0.00 to -1.00)

54

##
♣ One variable increases in value as the other decreases

e.g. amt of exercise and weight

### Negative (inverse) relationship

55

## positive relationship ranges from

### 0.00 to +1.00

56

##
♣ Both variables increase

o E.g., calorie consumption and weight

### positive relationship

57

## is -0.45 or +0.40 stronger?

###
-0.45

− Because -.45 is closer to one & you disregard the (+) & (-)

58

##
• Usually, think of an r of ___ as high; perfect correlations are very, very rare

### .70

59

##
frequently used indexes:

###
absolute risk

absolute risk reduction (ARR)

odds ratio (OR)

60

##
the proportion of people with an adverse outcome relative to those without it (e.g., the odds of…)

widely reported risk index

### odds

61

##
ϖ Used to make objective decisions about population parameters using sample data

### inferential statistics

62

##
ϖ inferential statistics is based on:

### laws of probability

63

## ϖ Its standard deviation is called the

###
standard error of the mean (SEM)

64

##
o you Can estimate the SEM from data from a single sample, using two pieces of information:

###
SD for the sample

sample size

65

##
indicate the upper and lower confidence limits

ϖ a wide range of values for the population value and the probability of being right

ϖ Reflect how much risk researchers are willing to take of being wrong

###
Confidence Intervals

66

## o CI of 95% reflects that researchers accept the risk they will be wrong

###
5 times out of 100

67

##
ϖ Uses objective criteria for deciding whether research hypotheses should be accepted as true or rejected false

### hypothesis testing

68

## states that there is no relationship between the independent and dependent variables

### null hypothesis

69

##
o The way researchers seek to accomplish rejection of the null hypothesis

###
ϖ Statistical tests

70

##
ϖ statistical test Involve statistical decision making to either:

###
o Accept the null hypothesis

o Reject the null hypothesis

71

##
o Rejection of a null hypothesis when it should not be rejected

A false-positive result

### type 1 error

72

##
o Risk of Type I error is controlled by the

### level of significance (alpha)

73

## the minimal acceptable alpha level is

### 0.05

74

##
o Failure to reject a null hypothesis when it should be rejected

• A false-negative result

### type 2 error

75

## the risk of a type 2 error

### beta

76

##
• is the ability of a statistical test to detect true relationships and is the complement of beta

### power

77

## power =

### 1 - B

78

## how to fix a type 2 error

### increase the sample size

79

##
o Use study data to compute a test statistic

###
ϖ Tests of Statistical Significance

80

## value that indicates that the null is improbable

### statistically significant

81

##
• Mean that any observed difference or relationship could have been the result of a chance fluctuation

(bias, etc.)

### nonsignificant results

82

## Overview of Hypothesis Testing Procedures steps:

###
ϖ Select an appropriate test statistic

ϖ Establish significance criterion

ϖ Compute test statistic

ϖ Calculate degrees of freedom (df)

ϖ Compare the computed test statistic to the table value

ϖ Decide to accept or reject null hypothesis

83

##
o Involves estimation of a parameter

o Assumes variables are normally distributed in the population

### parametric statistics

84

## parametric statistics: which part of NOIR

### interval/ratio scale

85

## ϖ Nonparametric Statistics: which part of NOIR

### nominal/ordinal

86

##
o Tests the significance of differences between TWO group means

### t-Tests

87

## independent t-test test what?

###
between subjects

(men vs. women)

88

## dependent (paired) groups t-tests test what?

###
within subjects

(before and after surgery)

89

##
o Tests the mean group differences of THREE or more groups

### ANOVA

90

##
o Tests the difference in proportions in categories within a contingency table

### chi-squared

91

## chi-squared tests between which 2 frequencies?

### observed and expected frequencies

92

## a descriptive and an inferential statistic

### pearson's r

93

##
o Tests that the relationship between two variables is not zero

### correlation coefficients

94

## Estimates of the magnitude of effects of an independent variable on the dependent variable

###
effect size indexes

(POWER ANALYSIS)

95

##
ϖ Statistical procedures for analyzing relationships among THREE or more variables simultaneously

###
Multivariate Statistical Analysis

96

##
• Used to predict a dependent variable based on two or more independent (predictor) variables

### multiple regression

97

##
o Continuous

o Interval- or ratio-level variables

###
• Dependent/Outcome variables

98

##
o Either interval or ratio level variables

OR dichotomous

###
• Independent/Predictor variables

99

##
• Used to control confounding variables statistically

### ANCOVA

100

##
= the confounding variables being controlled

### • Covariates

101

## depdent variables in ANCOVA

### continuous--ratio or interval

102

## independent variables in ANCOVA

### nominal (group status)

103

## covariates in ANCOVA

### continuous or dichotomous

104

##
ϖ Analyzes relationships between a nominal-level dependent variable (outcome) and 2+ independent variables

###
Logistic Regression

105

##
ϖ The ______is calculated after first removing (statistically controlling) the effects of confounding variables

### OR

106

##
ϖ typically summarize sample characteristics

### Descriptive statistics

107

##
ϖ Text gives you the following information:

###
o Which test was used

o Value of the calculated statistic

o Degrees of freedom

o Level of statistical significance

108

##
o Tucker tested the difference in the proportion of smokers versus nonsmokers who had ever tried an illegal drug

### chi squared

109

##
o Chase tested the difference in the mean birth weights of infants whose mothers either had or had not participated in a special prenatal education program

### t-tests

110

##
o Hutchings compared mean preoperative anxiety levels in three groups of patients with different types of cancer

### ANOVA

111

## Discussion Section of the Research Report has what content:

###
An interpretation of the results

• Clinical and research implications

• Study limitations and ramifications for the believability of the results

112

## Remember that researchers are seldom totally

### objective

113

##
• The statistical results of a study, in and of themselves, do not communicate much meaning.

• Statistical results must be ______ to be of use to clinicians and other researchers.

### interpreted

114

##
￼Interpretive Task

• Involves addressing six considerations:

###
credibility

precision

magnitude

meaning

generalizability

implications

115

## • Interpreting research results involves making a series of

### inferences

116

## • We infer from study results to

### "truth in the real world"

117

## Approach the task of interpretation with what type of mindset

### critical and even skeptical

118

##
Test the ―null hypothesis that the results are ____ against the ―research hypothesis that they are _____

###
null wrong

hypothesis right

119

## • The better the _____, the more credible the results

### proxies

120

##
Reporting guidelines have been developed so that readers can better evaluate methodologic decisions and outcomes.

include a flow chart for documenting participant flow in a study.

### CONSORT guidelines

121

## What alternative and potentially competing hypotheses could explain the results?

### internal validity

122

## • Do the specified eligibility criteria adequately capture the population construct?

### construct validity

123

## Was a power analysis done?

### statistical conclusion validity

124

##
To whom would it be safe to generalize the results

￼in this study?

### external validity

125

## Types of Biases

###
• Expectation Bias • Hawthorne Effect • Selection Bias

• Attrition Bias

• History Bias

• Extreme Response Bias • Type II error

126

## Multiple measures of the same outcome

### triangulation

127

## Mixed methods can corroborate evidence that can lead to heightened _____ of the data

### credibility

128

## results should be interpreted in light of ?

###
precision of the estimates

and

magnitude of the effects

129

## (often communicated through confidence intervals)

### precision

130

## effect sizes

### magnitude

131

## precision =

### confidence intervals

132

## • Confidence intervals indicate the ________ of the evidence of quantities of direct interest, such as treatment benefit

### accuracy

133

## magnitude =

### effect sizes

134

## tell us how much of a change implementing the intervention will give. Is it a large change or a relatively small change? This can affect whether we adopt the EBP change or not.

### effect sizes

135

## correlation does not prove

### causation

136