n

Sample Mean Sum of all x values divided Sample mean is also the
∑ xi by the sample size population estimate
x́= ^μ = i=1
n
n
Sample Variance Sum of squares of how far Distance x values lie from the
∑ ( x i−x́)2 each x value is from the mean
s x2 =σ^ x2 = i=1 mean divided by the
n−1
sample size less one
Sample Standard s x =σ^ x =√ s x 2 Square root of the variance Expresses variability in
Deviation population (distance values
dispersed from estimated
mean)
z-score x−μ Subtract mean from x and Transforms data from a
z=
σ divide by the standard normal distribution into a
deviation standard normal distribution
Standard Error σ Population standard Describes spread of sampling
se( x́ )=
√n deviation divided by square distribution; estimate of
root of the sample size population standard deviation
Expected Value E ( x́ )=μ Average of sample should Expected mean of all x values
be the population average; should equal population
sample mean is an mean; same value as the mode
unbiased estimate of in a standard distribution
population mean
Test Statistic ^ ¿
θ−θ Sample estimate of θ If larger than critical value,
t= minus the hypothesized can reject the null hypothesis
^
se ( θ)
value of θ divided by the
standard error
Critical Value α α is the significance level The cutoff value for
t n−1 ,
2 we’re willing to accept determining if a test is
significant
Confidence CI 100 (1−α ) % ( μ )= μ^ ± t se ( μ^ ) The confidence interval at Interval of numbers within
Interval
[ n−1 ,
α
2 ] α level of a population which we believe the
mean is estimate of the parameter will fall; if
mean plus or minus the hypothesized mean not in
 margin of error interval, reject the null
Confidence CI 95% =[ x́−1.96 se ( x́ ) , x́+ 1.96 se (x́) ] With repeated sampling, 95%
Interval (95%) of these samples will contain
the true parameter μ
Confidence CI 99% =[ x́−2.59 se ( x́ ) , x́+ 2.59 se ( x́) ] With repeated sampling, 99%
Interval (99%) of these samples will contain
the true parameter μ
Stochastic Model y=f ( x )+ ε y is the function of x plus Relationship holds true
any uncertainty that generally, but there are some
account for small variations and uncertainties
differences
Simple Linear y i=β 0 + β 1 x i +ε i y is the function of β0 plus One unit of y changes as a
Regression Model x times β1 plus error; β0 is function of x by β1 units; null
the y intercept; β1 is the hypothesis says β1 is 0, so y
slope changes by β0
n
Estimated β1 The sum of x and y values Tells us the value of β1 to use
∑ ( x i−x́ ) ( y i− ý ) minus their means, divided in the simple linear
^β 1= i=1 by the sum of squares of x regression; one unit change in
n

∑ (x i−x́)2 values minus their means x predicts a β1 change in y

Estimated β0 ^β 0= ý −β1 x́ The mean value of y minus Tells us the value of β0 to use
the mean value of x times in the simple linear regression
β1
Residual ε^ = y− β^ 0 − ^β1 x= y− ^y Estimated error is the The deviation of the original y
difference between left over once relationship
observed y and estimated with x is taken into account
(fitted) y
n n
Sum of Squares Sum of residuals squared is The difference between our
SS Residual=∑ ε^ i2 =∑ ( y i ¿− ^y i )2 ¿ the sum of the difference estimated y line and our actual
Residual i=1 i=1
between y values and y values; want to minimize
estimated y the squares between them
n
Sum of Squares Total variation is sum of All variation in a given simple
SSTotal =∑ ( y i− ý )2 the difference between y linear regression model
Total i=1
 values and the mean of y
squared
n
Sum of Squares Explained variation is the Variation of the estimated y
SS Explained=∑ ( ^y i− ý )2 sum of difference between values around their mean
Explained i=1
estimated y and the mean
of y squared
Sum of Squares SSTotal =SS Residual+ SS Explained Total variation is the sum
Total of residual variation and
explained variation
Coefficient of 2 SS Explained SS Residual Goodness of fit is the ratio A measure of how well the
R= =1− of the explained variance regression line fits the data
Determination SSTotal SSTotal
to the total variance (least squares)
Standard Error of se Regression= √ σ^ ε2= σ^ ε Square root of the residual Estimate of the magnitude of
Regression variance the typical deviation from the
regression line
n
Residual Sum of residuals squared Estimate of the variance of
Variance SS Residual i=1 ∑ ε^ i2 divided by the sample size the population errors
σ^ ε2= = less two
n−2 n−2
n
Sample The sum of x and y values Measure of how two variables
Covariance
∑ ( x i−x́ ) ( y i− ý ) minus their means divided move or vary together
s xy = i=1 by sample size less one
n−1
Sample s xy Sample covariance divided Measure of the strength of
r xy = by the standard deviation association between two
Correlation sx s y
of x times the standard variables
deviation of y
Multiple Linear y=β 0 + β 1 x 1 + β 2 x 2 +…+ β k x k + ε The more x values you use, Regressing multiple x
Regression Model the less error there will be variables on y
F-Statistic SS Model (Sum of squares of the Tells the overall significance
MS Model k model divided by # of of the regression model; if
F= = regressors) over (sum of significant, at least one x
MS Residuals SS Residuals
squares of the residuals variable is related to y
n−k −1
divided by # of
 observations - # of
regressors -1)
OLS Formula for σx y Covariance of the variables Population parameter in a
β 1= 1

divided by the variance multiple regression;