Medical Statistics Flashcards
What are the types of variables?
Categorical - binary - ordinal - nominal Numerical - continuous - discrete
what are binary variables?
can only have 1 of 2 variables
- e.g. yes or no
what are nominal variables?
no distinct hierarchy
- e.g. hair colour
what are ordinal variables?
distinct hierarchy/order
- e.g. social class
what are continuous variables?
can take any value in a given range
- e.g. height, age
what are discrete variables?
values must be whole numbers
- e.g. number of people
how would you display categorical data?
display on frequency
- bar charts
- pie charts
how would you display numerical data?
display using different graphs due to nature of data
- histograms
- box & whisker plots
- scatter plots
what is the sensitivity?
ability of a test to detect true positives
how would you calculate sensitivity?
TP/(TP+FN)
how would you calculate specificity?
TN/(TN+FP)
how would you calculate positive predictive value?
TP/(TP+FP)
how would you calculate negative predictive value?
TN/(FN+TN)
what is the specificity?
ability of a test to successfully exclude negatives
what is the positive predictive value?
probability that subjects with a positive screening truly have the disease
what is the negative predictive value?
probability that subjects with a negative screening test truly does not have the disease
how would you calculate the test accuracy?
(TP+TN)/(TP+TN+FP+FN)
what are the types of hypothesis?
- null hypothesis H0
- alternative hypothesis HA/H1
what is the null hypothesis?
there will be no significant difference
what is the alternative hypothesis?
there will be a significant difference
what is the P-value?
the probability of obtaining the observed results, assuming the null hypothesis is true
- the lower the p-value, the greater the statistical signficance
when would you use a scatter plot?
demonstrate correlation
when would you use a regression model?
to try to predict data using an equation
which axes do the independent and dependent variable go on?
x - independent
y - dependent
what is the simple linear regression model?
y = a + bx
what does the multiple linear regression take into account?
confounders
what is the multiple linear regression model?
y = (m1x1, m2x2, …) + c
when is logistic regression used?
when outcome variable is a binary variable
what is the logistic regression model?
logit(p) = a + bx
a - where the probability is
b - how fast it increases
