Stats Flashcards
P value definition
chance that the null hypothesis was rejected in error
odds that the Null was correct and the results were due to chance
probability of type 1 error
Type I error
big OOPS
thought something was significant and it was NOT
rejected the null in error
probability of type 1 error is the P values
study was overestimated
Type II error
Rejected the null hypothesis in error
Study was underestimated
Validity;
Internal vs External Validity
if an instrument or test actually measures what it’s supposed to measure;
internal - accuracy
external - generalizability
(cannot have EV without IV)
Reliability
the consistency or repeatability of scores
(may or may not be valid)
Incidence
of NEWLY diagnosed cases of a disease in a given period of time
Prevalence
TOTAL number of cases of disease existing in a population at a given time
Standard deviation
measure of the spread of individual values around the mean
Standard error
describes how accurate the SAMPLE MEAN is compared to the “TRUE” POPULATION MEAN
Confidence interval
measure of reliability of result
range of values in which you are confident that the true population result will be found
% within:
+/- 1 SD
+/- 2 SD
+/- 3 SD
+/- 1 SD = 68.2% [ +/- 34.1 ]
+/- 2 SD= 95.4% [ +/- (34.1 + 13.6) ]
+/- 3 SD = 99.8% [ +/- (34 + 13.6 + 2.2)
(outside 3 SD = 0.2%; +/- 0.1%)
Power
how to increase
of participants
probability that a test will correctly reject the null
= 1- beta
= 1- type II error rate
increases with:
higher effect expected
precision of measurement
increased type I error rate (alpha)
(inversely related to standard deviation)
Test for:
continuous variables
compare means of 2 groups
t-test
*Tea is MEANt for TWO
Test for:
compare means of 3 or more groups
ANOVA
(analysis of variance)
Test for:
categorical, discreet data
compare 2 or more %ages or proportions
Chi-square
*not mean values
NNT equation
1/ARR
(inverse of Absolute Risk Reduction)
e.g. if ARR is 25%; NNT = 4 = (1/0.25)
Perinatal mortality
deaths between 22wks gestation and 7 days of life
Neonatal mortality
deaths in the first 28 days of life
Infant mortality
deaths that occur within the first year of life
Most significant population accounting for no change in infant mortality rates
infants born <500g
- account for 20% of total IMR
- birthrate of infants <500g increased 50% (0.12-0.18%)
What two measures of central tendency are EQUAL in NORMAL distribution (aka symmetric)
mean and median
*mode is NOT used to categorize the symmetry of data
Difference between:
- Risk Ratio/Relative Risk (RR)
- Relative Risk Reduction (RRR)
RR: ratio of incidence in exposed vs non exposed
RRR: ARR / risk in the control/non-exposed group
*both have the same denominator