Brainscape's Knowledge GenomeTM


Top Flashcards (Certified)
All Flashcards

Epidemiology > Exam 3 > Flashcards

Exam 3 Flashcards

(93 cards)

Start Studying
1
Q

With a null hypothesis, what are the researchers looking to accomplish?

A

Either accept or don’t accept the null hypothesis based on statistical analysis

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

If a hypothesis states there is no relationship between variables/no difference between groups what is the OR equal to?

A

1

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Define superiority

A

Drug is better than placebo

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Define non-inferiority

A

Not much worse than no/standard treatment

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Define equivalency

A

Same as placebo/standard treatment

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What does variance estimate?

A

Estimates the variability in the sample around the sample mean

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What does a standard deviation close to 0 tell us?

A

If all the of the observed values in a sample are close to the sample mean

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

If the standard deviation is large, what does that tell us?

A

Observed values vary widely around sample mean

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What does it mean when the standard deviation is equal to 0?

A

Values in sample are identical

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Why do we use interquartile range (IQR)?

A

When a data set has outliers we use the difference between 1st and 3rd quartiles

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

How do we know if a data is normally distributed?

A

Graphical presentation is bell shaped because the mean and median are equal/near equal to each other and the data is evenly spread

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What % of the data lies 1 SD from mean

A

68%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What % of the data lies 2 SD away from the mean

A

95%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What % of the data lies 3 SD away from the mean

A

99.7%

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Any time the median is different from the mean describes:

A

Skewness

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

T/F: distribution is perfectly symmetrical when mean, median, and mode are all the same

A

True

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

If data is positively skewed, what is going on with the mean and median?

A

Mean is higher than the median value; “tail pointing to the right”

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

If the data is negatively skewed, what is going on with the mean and median?

A

Mean is lower than the median value; “tail is pointing to the left”

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What does kurtosis look at?

A

How much does the observation cluster around the mean

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

Positive kurtosis means:

A

More cluster

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

Negative kurtosis means:

A

Less cluster

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

Normal distribution kurtosis is equal to:

A

0

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

What does correlation aim to test?

A

If there is an ASSOCIATION/RELATIONSHIP between 2 elements variables

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

What does a regression aim to do?

A

Predicting some outcome; dependent variable. Looks at several variables to predict if there is a relationship/association. (OR)

Asses the relationship between an outcome variable and one or more risk factors or confounding variables

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
What does a survival comparison aim to do?
Looks at a lack of an outcome occurring or preventing an event; time to event
26
A correlation value (r) of 0 tells us:
No relationship
27
A correlation value of (r) of 1 tells us:
Perfect relationship
28
What is the correlation test name for nominal data?
Contingency coefficient
29
What is the correlation test name for ordinal data?
Spearman correlation
30
What is the correlation test name for interval data?
Pearson correlation
31
Linear relationship between 2 paired continuous variables
Pearson correlation
32
Relationship between 2 variables with 3rd variable removed to control for confounding
Pearson correlation ran as a partial correlation
33
Kappa statistic does what?
(Correlation test) The kappa statistic is frequently used to test interrater reliability. The question of consistency, or agreement among the individuals collecting data immediately arises due to the variability among human observers. Well-designed research studies must therefore include procedures that measure agreement among the various data collectors. Study designs typically involve training the data collectors, and measuring the extent to which they record the same scores for the same phenomena.
34
Kappa value of +1 tells us:
Observers perfectly classifies everyone exactly the same way (good agreement)
35
Kappa value of 0 tells us:
No relationship
36
Kappa value of -1 tells us:
Observers classify everyone exactly the opp of each other (poor agreement)
37
A p>0.05 for a pearson correlation tells us:
No LINEAR relationship; still non-linear correlations present
38
Measures relationship between variables by predicting the DV knowing the value of IVs describes:
Regression tests
39
What is the regression test for nominal data?
Logistic regression
40
What is the regression test for ordinal data?
Multinominal logistic regression
41
What is the regression test for interval data?
Linear regression
42
For logistic regression we use this for
If the outcome is dichotomous; yes/no, died/lived
43
We use linear regression in order to:
Prediction of an interval outcome/DV by utilizing multiple IVs
44
Derive an equation used to predict scores on one variable based on scores on one or more variables describes what statistical test
Multinominal logistic regression
45
Until have an event of interest ex hart attack, cancer remission, death; or proportion of events over time we would use what statistical test?
Survival tests
46
T/F: all data types can be represented by a kaplan-meier curve?
True
47
What is the survival test name for nominal data?
Log rank test
48
What is the survival test name for ordinal data?
Cox-proportional hazards test
49
What is the survival test name for interval data?
Kaplan-Meier test
50
When do we use a parametric test?
On normally distributed outcome mean or difference in means
51
What are the 4 parametric tests?
ANOVA Chi-Squared Correlation Regression
52
When do we use a non-parametric test?
Small sample size and outcome does no follow a normal distribution; eg: When the outcome is an ordinal variable or rank When there are definite outliers When the outcome has clear limits of detection
53
What are the 3 non-parametric tests?
Mann-Whitney Wilcoxon Signed Rank Kruskal Wallis
54
Why do we use the Bonferroni test of Inequality (Bonferroni correction)?
It adjusts the p value so to prevent data from incorrectly appearing to be statistically significant. Protects against type I error when conducting multiple analyses on the same DV. **in 3 or more comparisons, determines which groups are different!**
55
Which tests compares the median values between groups?
Mann-Whitney and Kruskal-Wallis; Wilcoxon Signed Rank and Friedman (Ie ordinal data, independent and paired)
56
What are the 3 post-hoc tests for ordinal data group comparisons?
Student-Newman-Keul, Dunn, and Dunnet
57
Why do we run post-hoc test for 3 or more group comparison in ordinal and interval data?
If tests come back statistically significant, we want to see where all the specific differences are
58
In what post-hoc test must the groups be equal in size?
Student-Newman-Kuel, Dunnett,Tukey/Scheffe
59
What post-hoc test allows for groups of unequal size?
Dunn test
60
What tests compares the means of all groups?
Student t-test and paired-t test; ANOVA and ANCOVA
61
What statistical test compares the means derived from unpaired samples; “difference between 2 conditions”
Student t-test
62
What statistical test compares the means of all groups agains a single DV?
ANOVA
63
What test looks to see if the observed frequency is statistically different from the expected value (determining whether 2 categorical variables are associated with one another); when the outcome is discrete (one, two, and more than 2 independent comparison groups)
Chi-squared
64
What tests looks at frequency data/difference in proportions
Chi-Square of Independence
65
T/F: interval data needs statistically equal variances?
True
66
How do we test to see if interval data has statistically equal variances?
Levene’s test
67
If the Levene’s test is statistically significant, what does that tell us?
There are differences->go to ordinal sheet
68
What are the required assumptions for interval/ratio data: proper selection for a parametric test?
1. Normally distributed 2. Equal variances (Levene’s test) 3. Randomly-derived and independent
69
How do we handle interval/ratio data that is not normally distributed?
1. Transform data to standardized value: z-score/log transformation 2. Use non-parametric test 3 use ordinal level stats test
70
What does it mean to have a alpha error?
Type I error when we reject the null hypothesis but null hypothesis is true; there is no difference but we are claiming a difference
71
What does it mean to have a beta error?
Type II error we are accepting the null hypothesis when we should reject it; not claiming there is a difference when there is
72
When we increase sample size we also increase:
Power
73
Power allows us to detect what?
Detect a difference if one exists
74
If we have a small sample size, what type of error are we more willing to make?
Type II/Beta
75
When do we need a smaller amount of people for our sample size
When we have bigger the difference
76
When do we need large amount of people for our sample size?
Closer the difference (small/minute differences)
77
When there is more variation, it is more difficult to denote:
A difference
78
P value tells us:
Probability of rejecting null hypothesis was incorrect (type I error)
79
Gives you inference on its magnitude and range of magnitude
Confidence interval
80
For what type of data must the CI cannot be below or above 1 (both have to be below or above 1)?
Ratio; ie OR/RR/HR
81
What type of data must the CI must be both pos or both neg (cannot be neg to pos #)?
Absolute differences
82
NCT stand for and do?
National clinical trials number and reduces publication bias
83
DOI stand for and do?
Digital object identifier and is a digital location for an article
84
what is the CONSORT checklist?
Interventional studies for clinical trials
85
What is the PRISMA checklist?
Interventional studies for systematic reviews and meta-analysis
86
What is the STROBE checklist?
Observational studies
87
What is the STROBE-ME checklist?
Observational studies for molecular epidemiology
88
What is the STREGA checklist?
Genetic association studies (observational)
89
What is the TREND checklist?
Non-randomized designs of behavioral and public health interventions
90
What is the REMARK checklist?
Tumor marker prognostic studies
91
What is the GRIPS checklist?
Genetic risk prediction studies
92
What is the STARD checklist?
Diagnostic studies
93
What is the QUADAS-2 checklist?
Systematic reviews of multiple diagnostic accuracy studies (2nd ed)

Epidemiology (1 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2025 Bold Learning Solutions. Terms and Conditions