IR WEEK 2 Flashcards

1
Q

define statistical power in relation to β error

A

the probability that a test will lead to rejection of the null hypothesis. (probability of attaining statistical significance)

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

List the 4 functions that determine statistical power

A

significance criterion, variance, sample size, effect size

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

Define Variance

A

as variance decreases, the power increases

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

define sample size

A

the larger the sample the greater the statistical power

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

define effect size

A

as effect size increases, then power increases

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

Define the significance criterion

A

as error decreases, power increases. if you lower the alpha level, then you are requiring stronger evidence to determine significance, but means you increase your chances of missing a true effect.

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

Define measurement error

A

the difference between the true value and the observed value

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

define reliability

A

the extent to which a measurement is consistent.

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

Define Validity

A

ensures that a test is measuring what it is intended to measure. implies that measurement is relatively free from error

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

Define accuracy

A

agreement between measured result and actual/true value
(systematic errors affect accuracy)

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

Define precision

A

repeatability or reproducibility of measurement
(consistent value does not imply correct value)

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

Define Systematic error

A

consistent over or under estimation of the true value (predictable)

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

define random error

A

due to chance, unpredictable (human error, simple mistake)

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

define minimal clinically significant difference

A

the smallest difference in a measured variable that signifies an important rather than trivial difference in the patients condition

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

This type of t-test
compares a sample mean to a given population mean
requires a normally distributed population and population mean is known
- the sample standard deviation won’t have a normal distribution (z distribution), because it is not a population in standard deviation

A

one sample t test

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

compares two sample means
requires two normally distributed but independent populations, population mean is unknown

A

students/unpaired t test

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

requires a set of paired observations from a normal population

A

paired t tests

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

List the four assumptions when performing a t test

A
  1. normal/gaussian distribution
  2. randomly sampled
  3. equal variances-
  4. data measured
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

what is a design that indicates one independent variable/factor with three or more variables?

A

one-way ANOVA determines if observed differences among a set of means are statistically significant from each other

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

null hypothesis

A

proposes no statistical significance between a set of observations

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

as error decreases

A

power increases

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

If the probability of committing a type 1 error decreases

A

the probability of committing a type 2 error increases

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

as variances decreases

A

then power increases

24
Q

the larger the sample size

A

the greater the statistical power

25
as effect size increases
power increases
26
as random error decreases
reliability increases
27
first measurement is expected to move closer or regress, toward the group average (mean) on the second measurement
regression toward the mean
28
what affects validity?
systematic error and extreme random error
29
the degree to which the changes in the dependent variable are the result of manipulation of the independent variable
internal validity
30
the degree to which the results of your sample can be inferred to the general population
external validity
31
one sample compared to a population
one sample t test
32
two sample groups
unpaired t test or paired
33
this type of t test compares one set of measurements with a second set of measurements from the same sample.
PAIRED T-test, can be used to compare before and after.
34
design indicates one independent variable/factor WITH 3 or more levels
one way ANOVA
35
other independent variables/factors can be added to the mix. looking for interactions between independent variables.
two way anova
36
Example: measure effects of 3 drugs and 3 diet regimens on blood pressure
can use two way ANOVA
37
Example: measure effects of 3 drugs on blood pressure
one way ANOVA
38
"within subjects" analogous to the paired t tests. measuring things at 2 different times
repeated measures ANOVA
39
includes multiple dependent variables
MANOVA multivariate ANOVA
40
effects of three different medications on diastolic and systolic blood pressure
MANOVA can be used bc there are two dependent variables.
41
this post test looks at a comparison of each group to each group to determine if the specific null for that pair can be rejected
Tukey's post test
42
specifies type 1 error rate for each pairwise contrast, rather than for the "family". is more powerful and more likely to detect significant difference
Newman-Keuls post test
43
uses familywise error rate, therefore as # of comparisons increases, each comparison has to achieve a lower p value to achieve significance
Bonferroni analysis, A PRIORI TEST
44
used to evaluate exploratory data, to evaluate the relationship between two measured variables
correlations
45
a sample size of less than 15 and a correlation below r=0.45, does this demonstrate correlation
it would be considered a weak correlation
46
using relationship between variables as a basis for prediction. draw conclusions about populations based on samples taken from that population
linear regression
47
independent or predictor variable for linear regressions
variable X
48
dependent or criterion variable
Variable Y
49
uses statistical methods to find the "best fit" line (regression line)
linear regression
50
points that do not seem to fit with the rest of the scores; lies outside the obvious cluster of scores
outliers
51
uses linear regression to evaluate new procedures or equipment in clinical setting
comparison of methods
52
perfect method agreement
y=x m=1 b=0 r>0.99
53
r^2 gives the percentage of total change in Y scores that can be explained by the X scores, will range between 0.00 and 1.00
coefficient of determination
54
if you find a correlation of r=0.087 for the regression of blood pressure on age, then r2 = 0.76
76% of the change in blood pressure can be accounted for by knowing the age, the other 24% is due to an unknown or identified variable.
55
linear regression can be used to predict Y when you know X if its within your data set. this is called
interpolation
56
if you attempt to predict Y when you know X and you go beyond your data set, this is called
extrapolation
57