Application of biostatistics definitions Flashcards
descriptive statistics
analyses that help describe, show, summarize our data
inferential statistics
analyses that allow us make conclusions about relationship btwn indpndnt and dpndnt variables or generalize our samples findings to the population
Types of Data
discrete –> nominal and ordinal
continuous –> interval and ratio
mean
average
median
middle value in range
mode
most frequent value observed
perfect distribution - mean median mode
all identical
symmetrical distribution
mean mode median in middle
positive skew
maxima tends left
L–>R
mode median mean
negative skew
absolute maxima tends right
L —> R
mean, median, mode
null hypothesis
typically that independent variable has no effect on dependent variable
in study wanting to see if two options are equivalent
- no there is a difference bteween option A and B
alternative hypothesis
independent variable has significant effect on dependent variable
type one error
null is true and you reject
type II error
null is false and you accept
statistical power
likelihood that a study will detect an effect when there is an effect there to be detected
power = 1 - beta ; beta = p(TypeII)
primary outcome
outcome that will answer the primary (most important) question
ex. whether a treatment is better at preventing disease- related death than the standard therapy
sample size of trial is calculated based on change expect see in this outcome
secondary outcome
other relevant outcomes about the same study but less importance
ex. wheter also a reduction in disease measures other than death, or whether new treatmnet reduces overall cost
hypothesis generating only
composite outcome
other relevant outcomes are combined to form a single primary outcome
chi square test
when categorical independent and independent variable
t-test and ANOVA
when categorical independent and continuous dependent
regression test
continuous independent and dependent variable
paired t-test
both measures come from the same subject
-ex comparing means obtained in a cross over study
unpaired t-test
measures come from diff subjects
ANOVA
used when more than 2 independent variables
-like series of t-tests
efficacy
measure of capacity of a treatment to produce the desired effect in a controlled environment, such as in randomized controlled trial
effectiveness
actual effect of treatment in practice
accuracy
how close measurement is to true/accepted value
precise
how close measurements of same item are to eachother
validity
does the variable reported actually measure an outcome of interest
sensitivity
ability of a test to correctly ID those patients with the disease
- sensitive test has low rate of false negatives
specificity
ability of test to correctly ID those patients without the disease
-low rate of false positive
statistical significance
difference btwn two interventions resulting in p<0.05
clinical significance
difference bwtn two interventions that is meaningful to the patient and their outcomes