The Model

• The first order linear model

Simple Linear Regression
y  b0  b1x  e
Example
y = dependent variable b0 and b1 are unknown population
x = independent variable y parameters, therefore are estimated
from the data.
b0 = y-intercept
b1 = slope of the line
Rise
e = error variable b1 = Rise/Run
b0 Run
x
1 2

The Estimated Coefficients The Simple Linear Regression Line

Alternate formula for the slope b1
To calculate the estimates of the slope and
intercept of the least squares line , use the sy • Example:
formulas: b1  r
sx – A car dealer wants to find
the relationship between
SS xy Car Odometer Price
b1  The regression equation that estimates the odometer reading and 1 37388 14636
SS xx
the equation of the first order linear model the selling price of used cars. 2 44758 14122
b0  y  b1 x is: 3 45833 14016
– A random sample of 100
  x   y 
4 30862 15590
SS xy   xi yi 
i i cars is selected, and the data 5 31705 15568
n recorded. 6 34010 14718

 x 
2 ŷ  b0  b1 x – Find the regression line.
.
.
.
Independent
.
.
Dependent
.
SS xx   x variable x variable y
i
2
i   (n  1) s 2
x . . .
n
3 4
 The Simple Linear Regression Line Assessing the Model
• Solution • The least squares method will produces a
– Solving by hand: Calculate a number of statistics regression line whether or not there is a linear
 x 
2

x  36,009 .45; SS xx   xi 2 
i
 43,528, 690 relationship between x and y.
n
• Consequently, it is important to assess how well
y  14 ,822 .823 ; SS xy   ( xi yi )  
xi  yi
 2, 712,511
n the linear model fits the data.
where n = 100.
SS xy 2, 712,511
• Several methods are used to assess the model.
b1 
(n  1) s 2
x

43,528, 690
 .06232 All are based on the sum of squares for errors,
b0  y  b1 x  14,822.82  (.06232)(36, 009.45)  17, 067 SSE.
ŷ  b0  b1x  17,067  .0623 x 5 6

Sum of Squares for Errors Standard Error of Estimate

– This is the sum of differences between the points – The mean error is equal to zero.
and the regression line. – If se is small the errors tend to be close to zero
– It can serve as a measure of how well the line fits the (close to the mean error). Then, the model fits the
data well.
data. SSE is defined by – Therefore, we can, use se as a measure of the
n
suitability of using a linear model.
SSE  (y  ŷ ) .
i 1
i i
2
– An estimator of se is given by se
– A shortcut formula
S tan dard Error of Estimate
SSE   yi2 b0  yi  b1  xi yi se 
SSE
7 n2 8
 Standard Error of Estimate,
Example Testing the Slope
• Example: • We can draw inference about b1 from b1 by testing
– Calculate the standard error of estimate for the previous
example and describe what it tells you about the model fit.
H0: b1 = 0
H1: b1 = 0 (or < 0,or > 0)
• Solution
– The test statistic is

SSE  9, 005, 450 b1  b1 se

t where sb1 
SSE 9, 005, 450 s b1 SS xx
se    303.13
n2 98
It is hard to assess the model based The standard error of b1.
on se even when compared with the – If the error variable is normally distributed, the statistic
mean value of y.
9
is Student t distribution with d.f. = n-2. 10
s e  303.1 y  14,823

Testing the Slope, Testing the Slope,

Example Example
• Solving by hand
• Example – To compute “t” we need the values of b1 and sb1.
– Test to determine whether there is enough evidence
b1  .0623
to infer that there is a linear relationship between the
se 303.1
car auction price and the odometer reading for all sb1  
(99)( 43,528,690)
 .00462
(n  1) s x2
three-year-old Tauruses in the previous example .
b1  b1  .0623  0
Use a = 5%. t
sb1

.00462
 13.49

– The rejection region is t > t.025 or t < -t.025 with n = n-2 = 98.
Approximately, t.025 = 1.984
11 12