Biostatistics Exam 2 Vocab

Define probability

The likelihood that an outcome will occur (relative frequency)

Define sample space

All possible outcomes

Define equiprobable

Each event is equally likely to happen

What approach do you use if events are equiprobable?

Classical approach

How do you measure probability if not equiprobable?

Empirical approach

Define permutation

The number of ways an object can be arranged

Set theory: what is a set? elements? sample space?

Set: collection of objects/interests

Elements: objects in the set

Sample space: all possible objects

What does this mean? P(A∩B)

Intersection, “and”, multiplication

What does this mean? P(A U B)

Union, “or”, adding

Define independent events

Outcome of one event does not affect the outcome of the other events

Define mutually exclusive events

Events/outcomes that can’t happen at the same time; Venn diagram shows no intersection (Ex. male or female)

What is the summation rule (product rule) and when do you use it?

Don’t count the intersection twice (use when events are not mutually exclusive- intersection is there)

How do you calculate union of independent events?

Union= summation - intersection

What is conditional probability?

Probability of a second event after another event occurred first

What is prevalence?

Existing rate of occurrence in population (ex. how common a disease is)

What is sensitivity?

Positive test, actually have disease (test accuracy); T+/E+

What is specificity?

Negative test, don’t have disease; T-/E-

What is a false positive? What is it the complement of?

Positive test, don’t have disease (T+/E-); Specificity complement (1 - specificity)

What is a false negative?

Negative test, have the disease (T-/E+); Sensitivity complement (1 - sensitivity)

What is sensitivity concerned with (sensitivity vs specificity)?

Minimizing false negatives, ensures individuals with condition are not missed by the test (goes untreated)

What is specificity concerned with (sensitivity vs specificity)?

Minimizing false positives, ensures that individuals without condition are not incorrectly identified as positive (don’t get treated if they don’t need it)

What are some things to think about in the specificity vs sensitivity debate?

Think about consequences of false positive vs false negative (false positive can retest with little consequences, )

Worries: disease manifestation, administering treatment that is not needed, side effects, interactions with other medications

Answer will change based on situation (no right answer)

What does a linear line mean in an ROC curve? What about a line with a very big curve?

Linear = guess, big curve = near perfect test

Define variable and random variable

Variable- any item being measured

Random variable- any value whose value is controlled by an element of chance

Define discrete and continuous

Discrete: yes/no, 2 outcomes

Continuous: many outcomes

What do binomial distributions assume?

1. fixed number of trials

2. independent events

3. probability is constant

What is the expected value?

Most likely outcome, mean

How do n and p determine the shape of binomial distributions?

As n approaches infinity, binomial distributions approximate a normal distribution. When P is 0.5 distribution is normal, when P is less than 0.5 it is right skewed, when P is above 0.5 it is left skewed

What is a poisson distribution? What are the three main characteristics?

Distribution of rare events, approximation of a binomial distribution; p is small, n is large, binomial probability (constant p, fixed # trials)

What happens to a curve when mean changes?

Location changes, shape does not (moves along the x axis)

What happens to a curve when st dev changes?

Shape changes, location does not (gets taller or shorter)

What is the empirical rule? (Give percentages)

Describes how much of the population each inflection point (1 stdev) represents

%: -1 to 1 (68.3%), -2 to 2 (95.5%), -3 to 3 (99.7%)

When is standard normal distribution a normal distribution?

When mean= 0 and stdev= 1

What is a Z value?

How many standard deviations away from the mean (inflection points)

What is the first step to tell if you have a normal distribution in research?

Make a graph (histogram)