BIOS 600 Exam 1

View()

allows you to view dataframes in a spreadsheet format

glimpse()

shows you an overview of the dataframe - the columns and first few rows

names()

returns a list of variable/column names

getwd()

returns the path for your working directory

YAML

sets the document details -> name, format, date

install.packages()

installs a package to the first time onto your computer

library()

loads a package in your workspace

ggplot()

initializes data visualization object with ggplot2

tidyverse

collection of R packages that have common grammar and data structure

geom_point()

makes a scatterplot

geom_boxplot()

makes a boxplot

geom_bar()

makes a bar chart

geom_line()

makes a line graph

labs()

sets labels

size

sets the variable to determine size of points

color

sets the variable to determine color

alpha

sets the transparency

fill

sets the color

#| warning: false

suppresses warnings that result from code

#| message: false

suppresses messages that result from code

filter()

filters to rows matching certain criteria

slice()

pick rows using an index/range of indeces

distinct()

filters for unique rows

select()

chooses variables by name

arrange()

orders rows by a certain value, ascending is default--use desc() for descending

summarize()

creates a new variable and displays summary statistics

group_by()

groups observations by a variable

mutate()

creates a new variable

sample_n()

takes a random sample with n observations

sample_frac()

takes a random sample of a fraction of observations

case_when()

creates a set of conditions/formulas for defining new variables

|>

pipes the previous row into the next- used output of one row as input for next

na.rm = T

removes NA values

dibinom(______)

k, size=n, prob=p

dipois(_____)

x=k, lambda=λ

Representative Sample

characteristics of the sample are similar to that of the population

Generalizability

extent that results form a sample can be used to draw conclusions about a population

Probability Sample

all units have a known probability of being selected

Simple Random Sample

every individual has an exact, equal, and independent chance of being chosen

Stratified Sample

a sample drawn in such a way that known subgroups within a population are represented in proportion to their numbers in the general population

Cluster Sample

obtained by selecting all individuals within a randomly selected collection or group of individuals

Non-Probability Sample

some units can't be selected or there is no way of knowing the probability of selecting any individual

Quota Sample

selecting individuals until a specific number is met

Convenience Sample

only members of the population who are easily accessible are selected

Snowball Sample

a sample in which respondents are asked to identify additional members of a population

Experimental Studies

researchers control exposure/treatment

Observational Studies

exposures/treatments are not assigned by the researchers

Selection Bias

the sample is not representative of the population being studied

Reporting Bias

the tendency to under-report available info

Non-Response Bias

bias introduced to a sample when a large fraction of those sampled fails to respond

Attrition Bias

occurs when participants drop out of a long-term experiment or study

Falsified Data

Data that are fabricated, or made up, by researchers intentionally trying to pass off research results that are inaccurate. This is a serious ethical breach and can even be a criminal offense.

Why is transparency important in research?

to uphold integrity, promote trust, and prevent harm

Reproducibility

being able to use the original data and methods/code to get the same results

Replicability

being able to independently repeat a study with the same methods with new data

Exploratory Data Analysis

summarizes a dataset main characteristics to help you get familiar with the data and identify and data quality issues

Nominal Categorical Data

named categories with no numeric meaning

Ordinal Categorical Data

ordered categories with relative comparisons

Count or Rank Numerical Data

e.g. number of drinks consumed

Continuous Data

measurable quantities that can take on any value

Parameters

describe a population

Statistics

describe a sample

Frequentist

probability is defined by the frequency of an event

Bayesian

probability as a measure of certainty about an event, updates as information becomes available

Probability Space

models random experiments or real world random events

Sample Space

set of all possible outcomes

Event

any possible number of outcomes (subset of sample space)

Disjoint/Mutually Exclusive

P(A or B) = P(A) + P(B)

Complement Rule

P(Ac) = 1 - P(A)

Komolgorov Axioms

1. Probability of any event occurring in the sample space is between 0 and 1

2. Probability of the whole sample space is 1

3. If A and B are disjoint, P(A or B) = P(A) + P(B)

Independence

P(A and B) = P(A) x P(B)

Prevalence

% of people with a condition

Sensitivity

true positive rate

Specificity

true negative rate (1 - false positive rate)

Positive Predictive Value

The probability that subjects with a positive test truly have the disease.

Negative Predictive Value

The probability that a person with a negative test result is truly disease free

Discrete Probability Distributions

1. outcomes are disjoint

2. probability of each outcome is between 0 and 1

3. sum of probabilities of all outcomes is 1

Bernoulli Distribution

binary outcomes with probability of success = p

Binomial Distribution

1. fixed number of Bernoulli trials

2. outcomes are independent

3. probability of success p is the same for each trial

Poisson Distribution

1. within any interval, k can take on any value 0,1,2...infinity

2. each event occurs independently

3. the average rate that events occur in an interval is constant

4. two events can't occur simultaneously

Expected Value

weighted avg of possible outcomes

Bernoulli: p

Binomial: np

Poisson: lambda

Variance

expected spread of values around the expected value

Bernoulli: p(1-p)

Binomial: np(1-p)

Poisson: lambda

Continuous Probability Distributions

given by a probability density function with are under the curve = 1 -- described for ranged or values

z-score

z=(x-mu)/sigma --> tells us how many population standard deviations an observation is away from the population mean