Stats- Vocab Cram

Population Data

Data for all individuals from the target population

Sample Data

Data for some individuals from the target population

Variable

feature we record that may vary from individual to individual

Observational Unit

single individual entity in a study

Sampling Frame

list of units from which the sample is selected (survey, polls, etc.)

Selection Bias

distortion in data that occurs when participants chosen do not properly represent the intended population

Volunteer Sample

when self-selected participants take part in a study

Convenience Sample

when researchers select participants who are most easily accessible

Nonresponse Bias

when ppl refuse to participate resulting in the final sample being unrepresentative of the overall target population

Response Bias

tendency of participants to answer questions inaccurately

Probability Sampling

technique in which everyone in the population has a probability to be selected for the sample

Simple Random Sampling

Everyone in the population has an equal probability of being sampled

Stratified Random Sampling

Dividing population then taking a sample from each

Cluster Sampling

Population is divided into clusters —→ a random sample of clusters is selected

Systematic Sampling

method of selecting individuals from a population at fixed, periodic intervals—> has random starting point

Observational Study

observe subjects in natural environment without interfering or assigning treatments; not possible to conclude a cause

Randomized Experiment

participants randomly assigned treatment; can conclude cause

Case-Control

2 distinct random samples of individuals w/ differing by some feature; individuals w/ feature= cases; indiv. w/o= controls

Cross-Sectional

information collected by observing multiple subjects at a single specific point

Completely Randomized Experiment

Individuals are randomly assigned to different treatment groups

Matched Pairs Experiment

Imposed conditions are compared on pairs/sets of related individuals

Study Design

Completely randomized experiment; individuals are assigned to different treatment groups completely at random

A single-blind experiment

where participants do not know which treatment they have been assigned

A double-blind experiment

where participants nor researchers knows who had which treatment

Categorical

variable for which the raw data are group or category names

Nominal

variable does not necessarily have logical ordering

Ordinal

variable does have logical ordering

Quantitative

variable for which the raw data are numerical measurements/counts

Discrete

can take one of a countable list of distinct values (ex: 1, 2, 3, 4,)

Continuous

can take any of the infinite possible values in an interval (ex: 1-4)

Statistic

summary of sample data

Parameter

summary of population data

P-Hat

proportion of individuals in a certain category

Mean

represented by “x-bar”; the average of distribution

Median

the middle of distribution

“Mu”

population mean

Sigma

population standard deviation

Variance

average squared difference from the mean

Sigma (squared)

population variance

Capital Sigma

Summation Operator

Alpha

significance level/ probability of Type 1 error

Beta

probability of a Type II error

Type 1 Error

False positive; concluding that a relationship exists when it does not

Type 2 Error

False negative; concluding there is no relationship when there is

Rho

Population correlation

Pi

Population proportion

Epsilon

Error term

“estimate”

Confidence Interval

Test

Hypothesis

N

sample size

What is this equation used for?

standard deviation of sampling distribution; tells you how much sample mean (x-bar) is expected to vary from true population size (mu) for a given sample size (n)

What is this equation used for?

Expected Value E(X) or “mu”; for calculating theoretical average outcome if repeated many times

What is this equation used for?

Variance; measures how spread out the outcomes are around the expected value

Central Limit Theorem

If a large sample size is taken (greater than or equal to 30) the distribution of the sample means will look like a bell curve

What is this equation for?

Calculating Confidence intervals; to determine how confident the range of values contains the true population parameter

Critical Value

threshold of probability for result to be statistically sigificant