Linear Regression in Excel

SIMPLE LINEAR REGRESSION EXAMPLE

Butlers Trucking Company is an independent trucking Company in southern California. A major portion
of Butlers business involves deliveries throughout its local area. To develop better work schedules, the
managers want to estimate the total daily travel time for their drivers.
nitially the managers believed that the total daily travel time would be closely related to the number of
miles traveled in making the daily deliveries. A simple random sample of !" driving assignments is
provided in Table !. #se $%cel to make a scatter diagram of these deliveries &to verify that a linear
relationship does e%ist' and develop a regression e(uation e%pressing this relationship.
Table !
)riving Assignment *
!
+,iles Traveled -+Travel Time &hrs.'
! !"" ../
0 1" 2.3
/ !"" 3..
2 !"" 4.1
1 1" 2.0
4 3" 4.0
5 51 5.2
3 41 4."
. ." 5.4
!" ." 4.!
Excel Instructions for Drawing a Scatter Plot
!. $nter the above information in the $%cel spreadsheet as shown in 6igure ! below.
0. Click on Insert on the toolbar and then click on the Chart tab. The Chart 7i8ard will appear. n step !
on select the *- &scatter' chart type &6igure 0', then click ne%t.
/. -our numerical data is contained in cells A0 through B!!. 9o in step two enter your data range as
shown in 6igure /, and click ne%t.
2. n steps / you can give your chart a title and label your a%es. n step 2 specify where you want the chart
to be placed. The finished chart is shown in 6igure 2.
1. After verifying that a linear trend does e%ist, determine the least s(uared regression e(uation.
Figure 1
Figure
Figure !
Figure "
Butler Trucking Example
0
2
4
6
8
10
0 50 100 150
Miles Traveled
T
r
a
v
e
l

T
i
m
e
Excel Instructions for Regression Anal#sis
!. The :egression ,acro &which is part of the Analysis Tool;ak' is standard with $%cel, however, it is not
always active and available for use. 9elect the Tools menu, if Analysis ToolPak is active then you should see a
Data Anaylsis item at the bottom of the menu. f this item is present skip to step /.
0. f this item is not there then you need to do one easy step. 9elect the Add Ins option under the Tools menu,
which brings up the following window.
Figure $
Click the Analysis ToolPak checkbo%, then OK. Analysis Toolpak should now be present under Tools in the
future.
0. 9elect the Data Analysis option under the Tools menu and select the Regression option &as shown below'.
Figure %
/. -our dependent variable &y' data is in cells B! through B!! &including the variable name or label', and your
independent variable data &%' is in cells A! through A!!. Click the labels bo% to indicate that the first row
contains the variable names, and then click ok. 9ee 6igure 5.
Figure &
2. A new worksheet will appear revealing the results of your regression analysis. The results from this analysis
are shown below.
SUMMARY OUTPUT
Regression Statistics
Multiple R 0.8149057
R Square 0.6640713
Adu!ted R
Square 0.6220802
Sta"dard
#rr$r 1.0017919
O%!er&ati$"
! 10
A'O(A
df SS MS F
Significance
F
Re)re!!i$" 1 15.87130435
15.87130
4 15.814578 0.004080177
Re!idual 8 8.028695652 1.003587
T$tal 9 23.9

Coefficient
s Standard Error t Stat P-value Lower 95% Upper 95%
Lower
95!%
Upper
95!%
*"ter+ept 1.273913 1.400744525
0.909454
2 0.3896874 ,1.95621171 4.5040378 ,1.9562117 4.5040378
-1.Mile!
Tra&eled 0.0678261 0.017055637
3.976754
7 0.0040802 0.028495691 0.1071565
0.0284956
9 0.1071565
Inter'reting Results
!. n your second model summary table, you will find the Coefficient of )etermination, :
0
, and the
Correlation Coefficient, :.
0. The A<=>A table gives the 6 statistic for testing the claim that there is no significant relationship
between your independent and dependent variables. The sig. value is your p value. Thus you should
reject the claim that there is no significant relationship between your independent and dependent
variables if p?.
/. The Columns below the Coefficients bo% gives the b
"
and b
!
values for the regression e(uation. The
intercept value is always b
"
. The b
!
value is ne%t to your independent variable, %.
2. n the last ;@value column of the coefficient output data, the p values for individual t tests for our
independent variable is given &in the same row as your independent variable'. :ecall that this t test tests
the claim that there is no relationship between the independent variable and your dependent variable.
Thus you should reject the claim that there is no significant relationship between your independent
variable and dependent variable if p?.
Coefficient of )etermination
Correlation
Coefficient
; value for
Anova Test
b
"
b
!
; value for
t test for *
!
II( M)L*IPLE REGRESSION EXAMPLE
n attempting to identify another independent variable, the managers felt that the number of deliveries
could also contribute to the total travel time. Table 0 includes the number of deliveries for each of the random
driving assignments provided in Table !.
Table 0
A B C
! *
!
+,iles
Traveled
*
0
+<umber of
)eliveries
-+Travel Time
&hrs.'
0 !"" 2 ../
/ 1" / 2.3
2 !"" 2 3..
1 !"" 0 4.1
4 1" 0 2.0
5 3" 0 4.0
3 51 / 5.2
. 41 2 4."
!" ." / 5.4
!! ." 0 4.!
To determine the regression e(uation for this scenario follow the same 9;99 steps provided for 9imple Ainear
:egression with the following modificationsB
$nter your multiple regression data in $%cel as shown above.
n 9tep /, specify your dependent variable &y' data is in cells C! through C!! &including the variable
name or label', and your independent variable data &%
!
and %
0
' is in cells A! through B!!. Click the
labels bo% to indicate that the first row contains the variable names, and then click ok. 9ee 6igure 3.
-our output for this multiple regression problem should be similar to the results shown below.
9#,,A:- =#T;#T
Regression Statistics
,ultiple :
"..1"453!4
4
: 9(uare
".."/533.5
1
Adjusted :
9(uare
".354/""!!
!
9tandard $rror
".15/!20!1
0
=bservations !"
A<=>A
df SS MS F
Significance
F
:egression 0
0!.4""1141
! !".3""03 /0.353/452/ "."""05402
:esidual 5
0.0..22/23
4 "./032.0
Total . 0/..
Coefficients
Standard
Error t Stat P-value Lower 95% !!er 95%
Lower
95"#%
!!er
95"#%
ntercept
@
".3435"!24
5
"..1!12550
1 @"..!0.2 "./.!4/2/"2
@
/.!!351043
/ !./3!/2.51
@
/.!!351043
/ !./3!/2.51
*!+,iles
Traveled
"."4!!/21.
.
"."".3332.
1 4.!30/.5 "."""210.4!
"."/5510"2
!
"."321!5!1
4
"."/5510"2
!
"."321!5!1
4
*0+<umber of
)eliveries
"..0/201/4
5
".00!!!/24
! 2.!5401! ".""2!14400
".2""15123
.
!.22405102
2
".2""15123
.
!.22405102
2
Inter'reting Results
!. n your second model summary table, you will find the Adjusted Coefficient of )etermination, Adjusted
:
0
, and the Correlation Coefficient, :.
0. The A<=>A table gives the 6 statistic for testing the claim that there is no significant relationship
between your all of your independent and dependent variables. The sig. value is your p value. Thus you
should reject the claim that there is no significant relationship between your independent and dependent
variables if p?.
/. The Coefficients bo% gives the b
"
and b
!,
and b
0
values for the regression e(uation. The constant value is
always b
"
. The b
!
value is ne%t to your %
!
value, and b
0
is ne%t to your %
0
value.
2. n the last column of the coefficient bo%, the p values for individual t tests for our independent variables
is given. :ecall that this t test tests the claim that there is no relationship between the independent
variable &in the corresponding row' and your dependent variable. Thus you should reject the claim that
there is no significant relationship between your independent variable &in the corresponding row' and
dependent variable if p?.