Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
33 Cards in this Set
- Front
- Back
population
|
group of interest
|
|
variable
|
measured piece of information
|
|
distribution
|
list of possible values of the variable (different categories, %, #...)
|
|
center
|
median, mean
|
|
spread
|
how far apart are the values; variation
|
|
shape
|
values evenly spread; skewness; outliers
|
|
quantitative vs categorial graphs
|
quantitative -histogram
categorical - bar graph |
|
Q1
|
lowest 1/4 of data
|
|
Q3
|
upper 1/4 of data
|
|
IQR
|
Q3-Q1
|
|
Lower Limit
|
Q1-1.5xIQR
|
|
Upper Limit
|
Q3+1.5xIQR
|
|
Empirical Rule
|
68% within 1 SD
95% within 2 SD 99.7% within 3 SD |
|
r
|
correlation coefficient; measures the strength and direction of a linear relationship between two random variables 0<x<1
|
|
r^2
|
coefficient of determination; describes % variation in "y" accounted for by LSRL
|
|
residuals
|
(y-y^) tell how far observed values fall from the regression line (observed minus predicted)
|
|
slope
|
for every increase in "x" there is a b increase in "y"
|
|
experiment
|
deliberately imposing one treatment on one group and another treatment on another group in order to observe the effects of different treatments (response) of the population
|
|
observational study
|
individuals are observed and variables are recorded WITHOUT trying to affect the outcome
|
|
simulation
|
using a mathematical model to represent a real-world situation
|
|
stratified random sampling
|
population divided into non-overlapping strata (grade)
|
|
selection bias
|
excludes some part of the population
|
|
measurement bias/response bias
|
method of measurement gives results systematically different from the true value; designed to elicit a particular response
|
|
nonresponse bias
|
responses not obtained from all individuals of sample; nonresponders not represented
|
|
blocking design
|
experimental units grouped when they are known in some way to be similar; we might expect might have a different response to the treatment
|
|
matched pairs design
|
comparing two treatments; subjects are matched in pairs as closely related as possible; receive each treatment at a different time and responses compared
|
|
probability
|
the long-term frequence of an event
|
|
Law of large numbers
|
more trials, closer observed values get to the true population
|
|
mutually exclusive, disjoint
|
not independent
|
|
Central Limit Theorem
|
If we draw an SRS from any distribution and the sample size is large, the sample mean is CLOSE to normally distributed
|
|
X^2 goodness of fit
|
compares variable with 2 or more categories to a fixed distribution
|
|
X^2 test for homogeneity
|
tests whether 2 or more populations are distributed differently from one another (multiples samples/treatments)
|
|
X^2 test for association/independence
|
tests whether two or more variables measured on teh same individuals are
(Ho): independent/not associated (Ha): dependent/associated (one sample) |