Class 11: research 2 Flashcards
what is a type 1 error
fasle positive - you are preg (man)
When you incorrectly reject a
true null hypothesis.
1 > P
what is a type II error
false negative - you are not preg (women)
Failing to reject a false null
hypothesis.
II > N
what are Likelihood Ratios used for
To determine how much a test result changes the probability of having a disease.
Combine both sensitivity and specificity into a single measure, making them more comprehensive.
Likelihood Ratios help clinicians make better diagnostic decisions.
Positive Likelihood Ratio (LR+) - what does a higher LR+ indicate
positive test result is more likely to be
seen in someone with the disease.
Positive Likelihood Ratio (LR+) - is the test more likely to rule in or rule out the condition
ruling in
Positive Likelihood Ratio (LR+) - what is a large shift in probability
> 10 = large shift in probability
Negative Likelihood Ratio (LR-) - what does a low LR- indicate
better the test is at ruling out the condition.
Negative Likelihood Ratio (LR-) - what is a large shift in prob
<0.1 = large shift in probability away from a diagnosis.
Values between 0.1 and 10 are not a large shift and not definitive
in ruling in or out.
what does a p-value of 0.8 indicate
indicates 8% probability that the difference
between the groups is due to chance
what is Alpha value:
Predefined threshold that determines if your result is statistically significant
Level of statistical significance, usually set before the study
what is a alpha value normally set to
0.05 à allowing 5% chance of error
in term of p and alpha values what must be true for a study to be statistically significant
p value must be lower or
equal to alpha level
Parametric - distribution
equal bell curve
Parametric - sample
randomized
Parametric - quantitative or qualitative
quantitative
Non-Parametric - distribution
non bell shaped -skewed curve
unequal distribution
Non-Parametric - quantitative or qualitative
qualitative
love
what is more powerful Non-Parametric or Parametric
Parametric
what are the different type of Parametric tests
t-test
Analysis of Variance (ANOVA):
ANCOVA:
what are the different types of Non-Parametric Tests
Chi square test
Mann Whitney U test
Kruskal Wallis test
what is Correlation
Describes the strength and direction of a relationship between two variables.
what is Positive Correlation
Both variables show same direction of
change.
Ex. Number of hours studying and NPTE scores
what is Negative Correlation
One variable increases while the other
decreases.
Ex. Duration of exercise and body weight
Correlation - high
> .76 to 1.00
Correlation - mod
.51 to .75
Correlation - fair
.26 to .50
Correlation - low
0.00 to .25
what is Regression
Regression examines how changes in the independent variables (predictors) are associated with changes in the dependent
variable (outcome)
what is a strong regression
R square > 0.5 indicates
strong association between variables
what is the difference in correlation and regression
- Correlation > measurement of strength/direction of relationship
- Regression > prediction
Compares mean between
2 independent groups: P and NP
P: Independent T-test
NP: Mann-Whitney U-test
Compares mean between
2 dependent groups: P and NP
P: Paired T-test
NP: Wilcoxon Signed Rank test
Compares mean between
3 or more independent
groups: P and NP
P: ANOVA
NP: Kruskal Wallis test
Quantifies association
between 2 variables: P and NP
P: Pearson product correlation
NP: Spearman rho correlation
what is the sensitivity formula
TP/TP+FN
what is the specificity formula
TN/TN+FP
what is the formula for LR+
sn/1-sp
what is the formula for LR-
1-sn/SP
when is 1 tailed vs 2 tailed used
1 tailed - you know the direction of the hypothesis
- you know this medicine works really well
2 tailed - do not know the direction of the hypothesis
what is the differnce between one way anova and two way anova
one way: only looking at one factor
- age
two way: looking at two factors
- age and activty
what is repeated measure anova
looking at the same group of indivduals over a period of time
what makes ancova different
the present of a co-varient
what is the difference between paired and independent t-tests
paired - same group of indivduals at the different time points
independent - looking at the difference between two groups of indviduals
chi squared test is used for what kind of data
nominal - unequal, mutally exclusive groups
manny white U -test
T independent test
2 independent groups
unequal ordinal data
kruskal wallis tes
anova
3 more groups
unequal ordinal
over what number indicates a strong correlation
> .75
what is the correlation and regression coefficient
R = regression
r = correlation
R square > 0.5 indicates what
strong association between variables