Module 1.2 Goodness of Fit and Hypothesis Tests Flashcards
Total sum of squares (SST)
(actual Y values - mean Y value)^2
Regression sum of squares (RSS)
(predicted Y values - mean Y value)^2
Sum of squared errors (SSE)
(actual Y values - predicted Y values)^2
Total variation =
explained variation + unexplained variation
SST =
RSS + SSE
RSS
explained variation
SSE
unexplained variation
degrees of freedom of explained variation
1
degrees of freedom unexplained variation
n - 2
total degrees of freedom
n - 1
mean square regression (MSR)
= RSS
Mean squared error (MSE)
= SSE/(n - 2)
k =
number of slope parameters (degrees of freedom explained variable)
standard error of estimate (SEE)st
square root of MSE
percentage of total variation of the dependent variable explained by the independent variable
R^2
R^2 =
regression sum of squares (RSS)/total sum of squares (SST)
for simple linear regression, what is an alternative way to determine R^2?
r^2 (squaring correlation coefficient)
assesses how well set of independent variables explains the variation in the dependent variable
F test
used to test whether AT LEAST one independent variable in a set of independent variables explains a significant portion of the variation of the dependent variable
F statistic
F statistic =
MSR/MSE
how many tails is an F test?
one tailed
what is the null hypothesis for an F test
b1 = X
what is the alternate hypothesis for an F test?
b1 does not = X
degrees of freedom for numerator (top part of fraction) for F test?
1
degrees of freedom for denominator of F test?
n - k - 1 = n - 2
decision rule for F test?
reject null if F statistic > critical value
t statistic (n - 2 degrees of freedom) =
(point estimate - hypothesized value)/standard error of point estimate
rejection of null in a t test means what?
slope coefficient is DIFFERENT from the hypothesized value
t statistic for simple linear regression =
R * square root of (n - 2)/square root of (1 - r^2)
smallest level of significance for which the null hypothesis can be rejected
p value
tells us how the dependent variable moves in relation to independent variable
slope coefficient