DATA TABLE and SUMMARY

Page Christy

Wall Test 2

10:39
09/22/15

WALL TESTING #2 4' WIDE x 8' PANEL

38.2

Distance
0.00
0.25
0.50
0.75
1.00
1.25
1.50

String 1
4010
2863
1783
1097
682
399
213

4000
2822
1867
1133
622
334
267

String 2
4075
3000
2159
1510
967
591
341

4047
3036
2183
1489
953
577
358

String 3
4095
2447
1442
934
601
368
182

was
3839 at 0.25"
normalized to 3000 to create a smooth
curve
4500
4000
3500

Column C

3000

Column E

2500

Column G
Column D

2000

Column F

1500

Column H

1000
500
0
0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40

1.60

3917
2634
1625
890
429
243
331

row 20

DATA TABLE and SUMMARY

Page Christy

Wall Test 2

10:39
09/22/15

0.800
0.700
0.600
0.500
Column H

0.400

Column I
Column J

0.300
0.200
0.100
0.000
160

Pressure
87
x
x
170.6
184.7
193.3
198.8
201.6
206.6
229.3
241.1
243.5
259
270.4
272.8
274.6
273.7
273.7
278.7
284.7
295.8
300.8
307

String 1
2229.3
x
x
2216.4
2216.5
2217
2217
2216.3
2216.9
2216.6
2217.1
2216.9
2216.7
2216.9
2216.9
2216.9
2216.9
2217
2217.1
2216.8
2217
2217.1
2217

String 2
4046
x
x
4053.6
4053.3
4053.1
4053.1
4053.4
4053.4
4053.7
4054.1
4056.4
4058.2
4053
4053.5
4052.8
4053.1
4052.9
4053.1
4053.3
4053.5
4053.1
4053.1

String 3
1000.9
x
x
1058.1
1057.7
1058
1058
1058.2
1058.3
1058.1
1058.1
1058.2
1058.3
1070.6
1071
1071.1
1070.5
1071
1071.1
1070.8
1070.7
1071.2
1071.1

180

200

220

240

260

time
10:39:25

15:35:04
15:35:15
15:35:27
15:35:43
15:35:58
15:36:14
15:36:27 567.112468
15:36:43
15:36:58
15:37:14
15:37:27
15:37:44
15:37:57
15:38:15
15:38:28
15:38:44
15:38:57
15:39:14
15:39:28
15:39:44

280

300

320

340

360

Pressure

String 1

String 2

182
199
210
217
220
226
254
269
272
291
305
308
310
309
309
315
323
336
342
350

0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328
0.328

0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023

String 3 time

0.757
0.757
0.757
0.757
0.757
0.757
0.757
0.757
0.757
0.757
0.749
0.749
0.749
0.749
0.749
0.749
0.749
0.749
0.749
0.749

DATA TABLE and SUMMARY

Page Christy

Wall Test 2

10:39
09/22/15

0.800
0.700
0.600
0.500
Column H

0.400

Column I
Column J

0.300
0.200
0.100
0.000
100

186.9
318.1
312.7
354.7
368.3
395.9
402.2
416.2
424
428.1
432.1
433.1
470.7
492.8
507.6
519.3
537.7
549
557.7
588.3
595.3
597.3
598.2
600.3
600.7
600.7
602
58.6

2218.3
2218.5
2218.8
2218.7
2218.6
2218.5
2218.5
2218.7
2218.6
2218.3
2218.3
2218.3
2218.3
2218.6
2218.4
2218.4
2218.5
2218.3
2218.5
2218.6
2218.5
2218.7
2218.5
2218.2
2218.6
2218.5
2218.4
2219.7

4054.7
4053.8
4053.4
4053.2
4052.4
4050.9
4050.9
4050.8
4050.5
4050.5
4050.4
4050.5
4050.5
4049.9
4049.5
4049.3
4049.6
4049.6
4049.3
4049
4048.9
4049.2
4048.8
4049.1
4049.2
4048.7
4049
4052.9

1070.7
1071.5
1071.2
1071.4
1071.2
1071.3
1071.4
1071.7
1071.8
1071.6
1071.8
1071.9
1072.1
1072.3
1084.2
1084.3
1084.2
1084.4
1084.1
1084.9
1085.1
1085.2
1084.7
1085.2
1085.5
1085.3
1084.9
1079.1

200

15:46:38
15:47:16
15:47:28
15:47:44
15:47:58
15:48:14
15:48:28
15:48:43
15:48:58
15:49:14
15:49:29
15:49:45
15:49:59
15:50:14
15:50:28
15:50:43
15:50:58
15:51:14
15:51:28
15:51:44
15:51:57
15:52:14
15:52:28
15:52:44
15:52:58
15:53:14
15:53:29
15:54:23

300

400

500

202
364
357
409
426
460
467
485
494
499
504
505
552
579
597
612
634
648
659
697
705
708
709
711
712
712
713

600

700

0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327

800

0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.024
0.023
0.023
0.024
0.024
0.024
0.024
0.024
0.024
0.024
0.024
0.024

0.749
0.748
0.749
0.749
0.749
0.749
0.749
0.748
0.748
0.748
0.748
0.748
0.748
0.748
0.741
0.741
0.741
0.741
0.741
0.740
0.740
0.740
0.740
0.740
0.740
0.740
0.740

DATA TABLE and SUMMARY

Page Christy

Wall Test 2

10:39
09/22/15

0.800
0.700
0.600
0.500
Column H

0.400

Column I
Column J

0.300
0.200
0.100
0.000

62.7
150.9
172.8
176.8
181.8
248.3
267.8
289.5
300.9
340.6
355.4
357.2
385.7
405.8
416.6
425
427.8
439.8
503.4
577.7
597.3
604.4
621
646.4
673.8
705
715
714.8
715.8
715.7
715.8
716.4
716.6
717.6
717
717.6
716.8
733.6
746.2
737.3
723.1
665.2
87

2219.3
2219.9
2220.1
2219.7
2219.8
2220.3
2219.7
2219.9
2219.9
2219.7
2220.1
2219.9
2219.9
2220.1
2219.9
2220.1
2220.1
2219.7
2220.1
2220
2220.2
2220
2219.9
2219.9
2219.7
2220
2218.8
2245.8
2245.9
2245.8
2246.3
2246.4
2246.3
2246.4
2246.6
2246.5
2246.8
2246.8
2247
2246.4
2246.7
2246.7
2229.3

4053
4053
4052.5
4053.2
4053.1
4053.1
4052.7
4053
4052.6
4052.8
4052.4
4052.4
4052.2
4051.8
4051.6
4051.9
4051.9
4051.7
4051
4050.9
4050.8
4050.6
4050.2
4049.9
4049.2
4049.1
4049
4049.4
4049.5
4049.7
4049.5
4049.7
4049.7
4049.8
4049.8
4049.8
4049.9
4049.8
4049.8
4049.8
4049.9
4049.8
4046

500

1079.3
1080.2
1080.6
1080.8
1081.1
1081.8
1082.3
1082.6
1082.6
1082.6
1082.7
1082.9
1082.8
1083.1
1083.2
1082.9
1083.1
1083
1083.9
1084.5
1084.3
1084.9
1084.4
1084.7
1085.1
1086.2
1085.8
1101.7
1101.7
1102
1101.7
1101.6
1101.9
1101.8
1101.9
1101.9
1102.1
1102.2
1102.1
1101.3
1100.1
1098.5
1000.9

15:59:20
15:59:38
15:59:54
16:00:09
16:00:24
16:00:38
16:00:54
16:01:09
16:01:24
16:01:38
16:01:54
16:02:09
16:02:25
16:02:38
16:02:54
16:03:09
16:03:24
16:03:38
16:03:54
16:04:09
16:04:24
16:04:38
16:04:53
16:05:09
16:05:24
16:05:38
16:05:53
16:06:06
16:06:09
16:06:14
16:06:19
16:06:24
16:06:28
16:06:32
16:06:38
16:06:44
16:06:49
16:06:53
16:06:59
16:07:03
16:07:08
16:07:14
16:07:19

1000

1500

109
352
412
423
437
619
673
733
764
873
914
919
997
1052
1082
1105
1112
1145
1320
1524
1578
1598
1643
1713
1788
1874
1902
1901
1904
1903
1904
1905
1906
1909
1907
1909
1906
1953
1987
1963
1924
1765

2000

2500

0.327
0.327
0.327
0.327
0.327
0.326
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.327
0.318
0.318
0.318
0.318
0.318
0.318
0.318
0.318
0.318
0.318
0.318
0.318
0.318
0.318
0.318

0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.024
0.024
0.024
0.024
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023
0.023

0.744
0.743
0.743
0.743
0.743
0.742
0.742
0.742
0.742
0.742
0.742
0.741
0.742
0.741
0.741
0.741
0.741
0.741
0.741
0.740
0.741
0.740
0.741
0.740
0.740
0.739
0.740
0.730
0.730
0.730
0.730
0.730
0.730
0.730
0.730
0.730
0.730
0.730
0.730
0.730
0.731
0.732

Page 14 - 5

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls
E

GETTING STARTED
This spreadsheet tool calculates by the "least squares" method.

Definition:
Y = mX + b describes a straight line on a graph

Click on Tools, Data Analysis, Regression

where:
Y
m
X
b

values plotted along the vertical Y axis

slope of the line
values plotted along the horizontal
X-axis
the Y-axis intercept
the point at which the line starts
for a given value of X and Y
along the slope m

row 20

Definition: Y = m1X1 + m2X2 +b

where:

the y variable depends upon values

of X1 and X2 for a single line
row 30

NOTE: Some references use the equation y = a + bx.

where a is the Y intercept and b is the slope.
This can get a bit confusing.
Note: Capitalized X and Y refer to actual plot values.
Lower case x and y are the distances from the actual
plot to the calculated values.

If the Data Analysis selection is not on your Tools

drop down menu, go to Add-Ins in the Tools
drop down menu and click on
Analysis ToolPak

Figure 14-1 The Analysis window.

row 40
NOTE: This template is printed at roughly 80 lines per
page because of the tables of calculations. Normally,
we print at 70 lines per page.

row 50

row 60

Figure 14-2 The Add-Ins window.

row 70

Page 14 - 6

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls

AN EXAMPLE
Analyze the following data used in calibrating a string gage.
The quickest way to get results is to use the Regression tool.
The solution is an attempt to fit a curve to the data.
Note that the resulting table does not change with adjustments to the
data.
row 80

to read the entire label(s),

increase the column width
row 90
In terms of y = mx + b
String 1
4010
2863
1783
1097
682
399
213

x^2
16080100
8194479
3177306
1202312
464988
159281
45369

Distance
0.00
0.25
0.50
0.75
1.00
1.25
1.50

compare
0.044
0.152
0.492
0.827
1.075
1.264
1.396

SUMMARY OUTPUT
Figure 14-3 The Regression analysis window.
Regression Statistics
Multiple R
0.9902058
R Square
0.98050753
Adjusted R
0.9707613
Standard Err 0.09234693
Observation
7

row 100

ANOVA
y data

df

x
x2
X Variable 1 X Variable 2
variables must be arrayed in this order -do not reverse

SS
MS
F
Significance F
2 1.71588818 0.857944 100.60373 0.00038
4 0.03411182 0.008528
6
1.75
Labels cut short by column width

Regression
Residual
Total

Coefficients Standard Err t Stat

Intercept
1.55639636 0.0816755 19.05585
X Variable 1 -0.0007731 0.00011106 -6.961648
X Variable 2 9.873E-008 2.614E-008 3.776497

m_1
m_2

P-value
4.47E-005
0.0022377
0.0194947

Lower 9 Upper 95 Lower 95 Upper 95.0%

1.32963 1.78316 1.32963 1.78316
-0.0011 -0.0005 -0.0011 -0.00046
3E-008 2E-007 3E-008 2E-007

1.60
1.40
1.20

The Distance curve is the raw data.

The compare curve is the distance
calculated using the values resulting
values from the regression analysis.

1.00
0.80
Distance
compare

0.60
0.40
0.20

Now that we have all of this data, what does it mean?

The following pages address the information presented
by the regression tool and the linest function.

0.00
0

500

1000

1500 2000

2500

3000 3500 4000 4500

xy scatter graph

Figure 14-4 Curve fitting with regression analysis.

row 130

Page 14 - 7

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls
E

STRAIGHT LINE REGRESSION ANALYSIS -- LEAST SQUARES

Straight Line Regression Analysis Y = mX + b
Where a least squares line is plotted using the values of x and y from the gage readings.
x
y
x2
y2
X - Xm
Y - Ym (X - Xm )2 (Y - Ym )2 straight line x
X
Y
X2
Y2
XY
y
53
45
2809
2025
2385
2.1
11.1
4.2
123.2
22
24.9
36
43
1296
1849
1548
19.1
13.1
362.9
171.6
89
88.1
88
89
7744
7921
7832
33.0
32.9
1085.7
1082.4
84
79
7056
6241
6636
29.0
22.9
838.1
524.4
86
84
7396
7056
7224
31.0
27.9
957.9
778.4
64
66
4096
4356
4224
9.0
9.9
80.1
98.0
row 140
45
49
2025
2401
2205
10.1
7.1
101.0
50.4
48
48
2304
2304
2304
7.1
8.1
49.7
65.6
39
43
1521
1849
1677
16.1
13.1
257.6
171.6
67
76
4489
5776
5092
12.0
19.9
142.8
396.0
54
59
2916
3481
3186
1.1
2.9
1.1
8.4
73
77
5329
5929
5621
18.0
20.9
322.2
436.8
65
56
4225
3136
3640
10.0
0.1
99.0
0.0
29
28
841
784
812
26.1
28.1
678.6
789.6
52
51
2704
2601
2652
3.1
5.1
9.3
26.0
22
27
484
729
594
33.1
29.1
1092.3
846.8
row 150
76
76
5776
5776
5776
21.0
19.9
438.9
396.0
32
34
1024
1156
1088
23.1
22.1
531.3
488.4
51
60
2601
3600
3060
4.1
3.9
16.4
15.2
37
32
1369
1024
1184
18.1
24.1
325.8
580.8
1101

1122

68005

20 each
Ym

Xm
55.05

56.1

Solve for the equations:

SY =
SXY

Sxy =

325.2

7394.95

Regression Statistics
Multiple R
0.965872
R Square
0.932909
Adjusted R 0.929182
Standard Err 5.126058
Observation
20

m SX + n b
b SX + m SX2

for b

Sxi Syi - Sxi Syi

2

322.2

7049.8

SUMMARY OUTPUT

3437

1374800
1360100

n * Sxi2 - (Sxi)2
b

68740
-61766
6974

XYm

n * (Sxi yi) - Sxi Syi =

n * Sxi2 - (Sxi)2
0.94306

for m

69994

1235322
1212201

76301610

75682740

1360100

1212201

ANOVA
Regression
Residual
Total

row 160
r2 is the correlation of estimated and actual y values

standard error for the y estimate

the F statistic to determine
degrees of
the degree of chance
freedom n - k
df
SS
MS
F Significance F
1 6576.8235 6576.82 250.29 5E-012
18 472.97646 26.2765
row 170
19
7049.8

4.184410

Note: move table labels

for easier reading

100

Standard Error
Lower 95%
Lower 95.0%
Upper95.0%
Coefficients
t Stat
P-value
Upper 95%
Intercept
4.18440963 3.475932 1.2038237 0.24425 -3.1183 11.4871 -3.11826 11.4871
X Variable 1 0.9430625 0.05961 15.820659 5E-012 0.81783 1.0683 0.81783 1.0683

90
80
70
60
50

linest

40

Given Y
slope m
Calculated Y standard error se

30

R Square, r2
chance relation F
SS Regression

20
10

m1
0.94306
0.05961
0.93291
250.293
6577

b
4.184
3.476
5.126
18
472.98

y intercept
standard error for the coefficient m
standard error for the y estimate
degrees of freedom n - k
SS Residual, SSE

0
10

20

30

40

50

60

70

80

90 100

Figure 14-5 The straight line regression analysis process..

row 190

Page 14 - 8

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls

STRAIGHT LINE REGRESSION ANALYSIS -- LEAST SQUARES

The Long Hand Approach
n
20 each
number of samples
k
2 each
number of constants in the
equation

S ( Y - Ym )2 - b1 / n ( nSXY - SX SY )
7049.8 -0.0471531
1374800
472.98

s
s2

SSE / (n - k)
26.276

n SXY - S X SY
(n S X2 - (SX )2 )

(n SY2 - (SY )2 )

1374800
1360100

1122
1399880

r
r
r
r2
SYX

139478
147899
0.966
0.933

Sy
n-1
2

s2Y
sY
sb
sb

18

371
19.3
SYX /

Sx2
0.0596

Continued
where:
Yc
Xm Ym
x y

Yc = mX + b
the computed/expected value of y
the mean value(s) of X and Y
deviations from their means x = X - Xm
and y = Y - Ym
the y-axis intercept
slope of the line (y2 - y1) / (x1 - x2)
sum of squares for error = S( Y - Ym )2
degrees of freedom n - k
row 200

1122.00
where:
n
k

number of samples
number of constants in the equation, hence:
y = mx + b
m and b = 2 constants
y = m1 + m2 + b
3 constants

Sx2

Sx2 represents standard

error of regression coefficient: the
dispersion of X values around their mean.

r2

compares estimated and actual y-values

value ranges from 0 to 1 with 1 being a
perfect correlation also known as R Square

Syx

Standard Error of Estimate of the dependent

variable Y regressed against the
independent variable(s) X

sey
se

standard error of the y estimate

the standard error values for the
coefficients m1, m2, etc.

sb, seb

Standard Error of the Regression Coefficient

which is a measure of the sampling error
in b.
describes the sampling distribution of a
deviation from the population mean divided
by the standard error
row 230

140996
R square

-1258884

7049.8 -0.9430625
6974
20
-2
5.1261
standard error of estimate

SYX

Sy2 - mSxy
n-k

SYX

2
Y

-1101.00

472.98 /

-1101
-1212201

b
m
SSE
df

SSE
SSE
SSE
2

7049.8
20
-1
estimated total variance
standard deviation
5.1261
7395
error of regression coefficient

ssresid

row 220

the sum of the squared differences for

estimated y y values and the actual y values
for each point sum (y estimated - y actual)2
S ( Y - Yc )2 called the
residual sum of the squares
the ( Y - Yc ) 's above the line roughly equal
the sum of the ( Y - Yc ) 's below the line
hence: S ( Y - Yc ) = 0

row 250

Page 14 - 9

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls

STRAIGHT LINE REGRESSION ANALYSIS -- LEAST SQUARES

S( X = Xm)2

Xm - s

19.7
55.05
X
35.3
35.3

Y
0
90

74.8
74.8

0
90

2s

2s

3s

3s
68.27%

Xm + s

count
%

row 260

95.45%
data points between the s cursors
13 points / n = 19 * 100

13
68.4 %

Continued

7394.95
19
standard deviation
from above

n-1
s
Xm

99.73%
Figure 14-6 Standard deviations.

Sf

SYX

1 + 1/n + x2 / Sx2

5.1261

X
Sf

55
5.253

0.0500

x2 /

standard deviation,
standard measure of dispersion

ssreg

the sum of the squared differences for the

actual values and the average y values for
each point
sum (y actual - y average)2
called the regression sum of squares

row 270

7394.95 F

sample value for X

standard error of forecast
Spurr and Bonini pg 472

Standard Error of an Individual Forecast

x2
X
x
15
-40.05
1604.00
35
-20.05
402.00
55
-0.05
0.00
75
19.95
398.00
95
39.95
1596.00

s, s

SYX

Sf

Sf 95%

Yc

Sf + 95%

5.1261
5.1261
5.1261
5.1261
5.1261

5.77
5.39
5.25
5.39
5.77

12.116
11.312
11.031
11.310
12.111

18.330
37.192
56.053
74.914
93.775

30.447
48.504
67.083
86.224
105.887

120

row 280
Sf - 95% where the value 2.10 * SYX and Sf comes
6.214 from the table VALUES of t for Probability P
25.879 see row 412
45.022
63.604
81.664
estimated y value
slope m

row 290

y-axis

100

b=
y intercept

80

Calculated
Data
+95%
-95%
s

60
40

deviation from
the line y
actual y value
X mean, Y mean
x-axis
Figure 14-7 Straight line regression analysis.
row 300
total sum of squares
regression sum of squares + residual sum
of the squares

20
0
10

20

30

40

50

60

70

80

90

100

Figure 14-8 Regression analysis with the least squares concept.

F statistic determines whether the observed

relationship between dependent and
independent variables occurs by chance.
Always positive.
row 310

Page 14 - 10

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls
E

LINEST VALUE
The linest function is the simpler, easier, and more dynamic way to do straight line, multiple straight line, and curvilinear regression analysis.
You can fit a straight or curved line and watch the linest function change to reflect your iterations.

String 1

Distance
X1

4010
2863
1783
1097
682
399
213

Y
0.00
0.25
0.50
0.75
1.00
1.25
1.50

11045.6

5.25

linest
m
se
r2
F
ssreg

To get help with the equation entry, click on the "=" sign in front of the
edit bar.

row 320

m1
b
-0.0003661 1.3276942
5.117E-005 0.1047367
0.91100747 0.17648622
51.184493
5
1.59426308 0.15573692

row 330

row 340

row 350

Figure 14-9 The Linest value process.

row 360

row 370

Page 14 - 11

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls
E

LINEST VALUE -- Continued

To utilize the linest tool, first select an area to edit or input the
formula and then park it with a space in front of the "=" sign.
Degrees of P = 0.1 P = 0.05
freedom tail 0.05
1
6.314 12.706
2
2.900
4.303
3
2.330
3.182
4
2.132
2.776
5
2.015
2.571
6
1.943
2.447
7
1.895
2.365
8
1.860
2.306
9
1.833
2.262
10
1.812
2.228

Describe an array with your cursor and press the [F2] key.

Remove the " ' " and space in front of the "=" sign.
Press the [Ctrl] [Shift] [Return] keys at the same time.

11
12
13
14
15
16
17
18
19
20

1.796
1.782
1.771
1.761
1.753
1.746
1.740
1.734
1.729
1.725

2.201
2.179
2.160
2.145
2.131
2.120
2.110
2.101
2.093
2.086

21
22
23
24
25
26
27
28
29
30

1.721
1.717
1.714
1.711
1.708
1.706
1.703
1.701
1.699
1.697

2.080
2.074
2.069
2.064
2.060
2.056
2.052
2.048
2.045
2.042

infinity

1.645

1.960

row 380

row 390

row 400

row 410

Figure 14-10 Values of t and probability P.

To get this result:

row 420
one tail

Figure 14-12 The one tail curve.

Figure 14-11 The Linest process.
row 430

Page 14 - 12

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

MULTI-STRAIGHT LINE REGRESSION ANALYSIS

We will use fabricated data for this example
Area
Height
Value
X1
X2
X12
Y
14.7
155
4.1
216.1
14.2
155
3.9
201.6
12.7
158
3.2
161.3
13.8
158
2.9
190.4
14.4
155
3.9
207.4
17.4
157
4.1
302.8
21.8
172
5.8
475.2
14.0
170
5.1
196.0
17.5
175
6.8
306.3
23.0
185
6.8
529.0
18.3
185
6.5
334.9
19.4
205
7.0
376.4
15.2
215
5.8
231.0
18.3
195
5.1
334.9
21.7
178
5.3
470.9
16.7
160
4.9
278.9
13.6
205
6.0
185.0
14.5
190
5.3
210.3
12.1
203
4.8
146.4
17.4
125
4.3
302.8
330.7
3501
101.6
5657.41
n
20
mean ________________________
X1m
X2m
Ym
X2m x1
16.535

175.05

5.08

5468.12
Sx12
189.29

287361355.xls
E

X22

Y2
16.81
15.21
10.24
8.41
15.21
16.81
33.64
26.01
46.24
46.24
42.25
49.00
33.64
26.01
28.09
24.01
36.00
28.09
23.04
18.49
543.44

24025
24025
24964
24964
24025
24649
29584
28900
30625
34225
34225
42025
46225
38025
31684
25600
42025
36100
41209
15625
622729

X2m x2

Ym y

612850
Sx22
9879

516.13
Sy2
27.312

Long Hand Calculations

Simultaneous Equations for y = m1x1 + m2x2 + b

X1 Y

X2 Y

X1 X2

60.27
55.38
40.64
40.02
56.16
71.34
126.44
71.4
119
156.4
118.95
135.8
88.16
93.33
115.01
81.83
81.6
76.85
58.08
74.82
1721.48

635.5
604.5
505.6
458.2
604.5
643.7
997.6
867
1190
1258
1202.5
1435
1247
994.5
943.4
784
1230
1007
974.4
537.5
18119.9

2278.5
2201
2006.6
2180.4
2232
2731.8
3749.6
2380
3062.5
4255
3385.5
3977
3268
3568.5
3862.6
2672
2788
2755
2456.3
2175
57985.3

Ym x1

X2m y

X2m x1

1679.96 17785.08
Sx1y
Sx2y
41.524
334.82

57889.0
Sx1x2
96.3

linest

m1

plot the mean of the two values

through the data scatter
X1
X2
Y
9.921 105.03
3.048
16.535 175.05
5.08
24.803 262.58
7.620

4.067383

row 440

row 450

row 460

m2

Sx1 y

= m1 x12

+ m2 x1 x2

slope m

41.524

= m1 189.29

+ m2 96.3

standard error se

0.0070014 0.05058 1.44269 standard error for m

0.7000701 0.69416 #N/A
19.839956
17 #N/A

standard error for the y est

degrees of freedom n - k

19.120315 8.19169 #N/A

SSE

Sx2 y

= m1 x1 x2

+ m2 x22

R Square, r2
chance relation F

334.82

= m1 96.3

+ m2 9879

SS Regression

multiply Sx2y by

1.966

to get

658.36

0.0319

0.2031 -3.8653 y intercept

row 470
189.29

19425

-19328.676

subtract from Sx1y

to get

-616.83

m1
m2

0.2031
0.0319
Y mean -m1X1 mean -m2X2 mean
5.08
3.359
5.586
-3.8653

row 480

To predict y with inputs for x1 and x2

x1
15
x2
y
y
n
k

180
3.047
4.926
20 each
3 each

5.744

-3.865
number of samples from above
number of constants in the regression equation
row 490

Page 14 - 13

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls
E

MULTI-STRAIGHT LINE REGRESSION ANALYSIS -- Continued

Sy2 - m1Sx1y - m2Sx2y
Syx
n-k
Syx

27.312
20
0.6942

Syx
Sf

SYX

-0.2031 41.52
-0.0319 334.82
-3
standard error of estimate

Sy2
n-1
1.437

s2Y

27.312
19
estimated total variance

R2
R2

1 - S2yx / s2Y
0.6648

0.8153

8
7

1 + 1/n + x2 / Sx2

s2Y

row 500

6
5

x1
regression

4
3

coefficient of multiple determination

also known as R Square

2
1

row 510

Beta Coefficients
b1
b1

0.2031

b1

b2

14

16

18

20

22

24

26

Figure 14-13 Regression Graph

189.29
27.312
b is a pure number representing the
net regression coefficient expressed in units of its
own standard deviation.

0.535

b2

12

Sx12
Sy2

b1

10

Sx22
Sy2

b2

0.0319

b2

0.607

row 520
Compare these values to see which values,
X1 or X2, have greater significance.

9879
27.312

Standard Error of Regression Coefficient

r212
(Sx1x2)2
(Sx21)(S2x2)
r212

0.0050

sm1

Syx
(1 - r212 )

Sx21
sm1

0.694
189.29

sm1

0.0506

sm2
Sx2
sm2
sm2

96.3 ^2
189.29

9879

7
6
x2
regression

row 530

4
3

0.9950
Standard Error of the Regression
Coefficient

Syx
(1 - r212 )
0.694
9879
0.0070

2
1
0
80 100 120 140 160 180 200 220 240 260 280

0.9950

row 540

Figure 14-14 Regression Graph No. 2

m1

4.016 standard errors from B1 = 0

0.2031 m1 / 0.0506 sm1
m2
4.558
For n - k degrees of freedom, the one tailed t value is 2.567 at a significance of 0.01.
Both m1 and m2 are greater than 2.567 at a significance of 0.01.
row 550

Page 14 - 14

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls
E

THE F COEFFICIENT
Are the methods different enough from each other to be significant?
Method
1
2
3
4
scores
65
75
59
94
87
69
78
89
73
83
67
80
79
81
62
88
81
72
83
69
79
76
90
T X
group mean
Tm (SX)2

454
75.67
34353

n
p

549
78.43
43057

23
4

CM
CM

(SX)2 / n
137602

Total SS
Total SS

S(X - Xm)2 - CM

SST
SST
SSE
SSE

425
70.83
30104

351
87.75
30800

1779

SX2

4225
7569
5329
6241
6561
4761

5625
4761
6889
6561
5184
6241
8100

3481
6084
4489
3844
6889
5776

8836
7921
6400
7744

34686

43361

30563

30901

row 560
139511

Xmm

overall mean denoted in text books as X

with double bars on top

groups, populations

CM

correction for the mean

1779
23
correction for the mean

Total SS
SST
SSE

SST + SSE
row 569
sum of squares for treatment
pooled sum of squares of the deviations,
sum of the squares for error

MST
MSE

mean squares for treatment

mean square for error

139511

137602

Sn(Tm - X)2 = ST2/n -CM

712.6

138314

137602

Total SS - SST
1196.6

1909.2

712.6

138314.4

1909.2

Mean squares for treatment and error

MST
237.5

a = 0.05
0.25
712.6

MSE
MSE

SSE / ( n - p )
63.0

1196.6

19

F
F

MST / MSE
3.77

237.53

62.98

a = 0.05
0.25

MST / MSE the chance relation between

data groups and within a group the F
statistic determines whether the observed
relationship between dependent and
independent variables occurs by chance.
Always positive.

n1

factor for the groups in the F statistic

n2

factor for the number of observations within

3
n-p

a = 0.05

-1.96
1.96
1.645
two tailed statistical test
single tailed statistical test
rejection region shaded
rejection region shaded
Figure 14-15 Two tailed and single tailed curves.

to find the F statistic for

n1
p-1

n2

19
for F.05

3.13

OK 3.13 < 3.77

a number of groups
r2

the coefficient of correlation between y and x,

A measure correlation varies from 0 to 1
with 1 being the best correlation

row 610

Page 14 - 15

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls

THE F COEFFICIENT -- Continued

Data Table from above
Method
score #
1
1
65
2
87
3
73
4
79
5
81
6
69
7
0

linest
2
75
69
83
81
72
79
90

3
59
78
67
62
83
76
0

6
75.7

7
78.4

6
70.8

4
94
89
80
88
0
0
0

m
b
Err:504
Err:504 Err:504 Err:504 Err:504
Err:504
Err:504 Err:504 Err:504 Err:504
Err:504
Err:504 Err:504 Err:504 Err:504
Err:504
Err:504 Err:504 Err:504 Err:504
Err:504
Err:504 Err:504 Err:504 Err:504
the chance relation F in this linest is not valid

slope m
standard error se
R Square, r2
chance relation F
SS Regression

linest must have 0's as place holders

linest
4 slope m
-0.056603 -0.124712 0.12109
87.8 standard error se
0.011377 0.0648682 0.06152
R Square, r2
0.971371 0.6330973 #N/A
chance relation F
16.96454
2 #N/A
SS Regression
27.19838 0.8016245 #N/A

0.1367
0.06516
#N/A
#N/A
#N/A

row 620

-3.9531
5.63671
#N/A
#N/A
#N/A
row 630

Sort each group in ascending order before using linest

score #
1
2
3
1
65
69
59
2
69
72
62
3
73
75
67
4
79
79
76
5
81
81
78
6
87
83
83
7
90
7
3.0

6
75.8

7
65.9

4
80
88
89
94

61
95

1
0.156
7.908

Group 1
slope m
standard error se
R Square, r2

6
71.3

4
88.8

chance relation F
SS Regression
Group 2
slope m
standard error se
R Square, r2

100
90
80
70

Column B

60

Column C
Series 3

50

Column E

40

slope 1

30

slope 2
slope 3

20

slope 4

chance relation F
SS Regression

2
66 0.2761054
90 7.4670743

3
55 0.44606
95 8.16128

linest
0.228 -13.752
0.0113578 0.86357
0.9901714 0.20736
402.97674
4
17.328
0.172
linest
0.2996237
0.0219888
0.973777
185.67265
27.265757

4
80 0.84615
95 4.04715

row 640

-19.499
1.73062
0.38321
5
0.73424
row 650

Group 3
slope m
standard error se
R Square, r2
chance relation F
SS Regression

linest
0.1928805 -10.162
0.0153079
1.09
0.9754242
0.33
158.76182
4.00
17.069924
0.43

10
0
0

Figure 14-16 A comparason of methods.

Group 4
slope m
standard error se
R Square, r2
chance relation F
SS Regression

linest
0.2133995 -16.226
0.0452131 3.97393
0.9176179
0.45
22.277108
2
4.5880893 0.41191

row 660

row 670

Page 14 - 16

REGRESSION ANALYSIS

Christy
10:39
09/22/15

14
ENGINEERING with the SPREADSHEET
Copyright 2006 American Society of Civil Engineers
A

287361355.xls
E

CURVILINEAR REGRESSION ANALYSIS

And now, for the real reason that I got into this essay.
unwind
Record the readings of a string gage at about 0.25 inch intervals. This string gage uses a linear
potentiometer but the readings from linear potentiometers and LVDT's are never quite linear. To
interpret readings from the gage between the calibration measurements, we create the equation
that can plot a straight line through the scatter of points and give fairly accurate results.

a spring causes the

potentiometer to rewind
Figure 14-17 A diagram of the string gage.

The creation of an equation to fit these points is fitted essentially by hand. We

use the linest function to help us fit this curve. Linest is dynamic -- as you
change your values, linest re-computes its value.

r2

the correlation of estimated and

actual y values

The r2 value is of critical value.

Note:

an r2 greater than 0.95 is usually acceptable

LVDT

linear voltage displacement transducer an

instrument which uses a power supply to
create and AC voltage. The measurement is
made by the position of a plunger within a
coil whose voltage changes as the plunger
moves. Measurements must be
"normalized" to be truly accurate.

String 1
x1

x2 = x11.01

4010
2863
1783
1097
682
399
213

4357
3100
1921
1176
728
424
225

power of
n

m
se
r2
F
ssreg

m
se
r2
F
ssreg

Distance
y
0.00
0.25
0.50
0.75
1.00
1.25
1.50

curve
4011
2861
1782
1098
683
400
212

x1 mean
3627
1578
-471

y mean
0
0.75
1.5

1.60

1.01
7

1.40
1.20

linest x1
-0.0003661 1.3276942
5.117E-005 0.1047367
0.91100747 0.17648622
51.184493
5
1.59426308 0.15573692
linest x1 + x1power of
0.03267229 -0.0359284
0.00459171 0.00499789
0.99348401 0.05339237
304.937274
4
1.73859702 0.01140298

Colum
nC
straigh
t line
curve

1.00
0.80
0.60

row 700

0.40
1.75986483
0.06850498
#N/A
#N/A
#N/A

0.20
0.00
-1000

1000

2000

3000

4000

5000

row 710

Figure 14-18 The straight line approximation of a curve.

2
This plot can be also be fitted with Excel's add trendline
feature in the graph pop up menu.
Click on the curve of actual values and select add trendline.
You have the option of displaying the formula and the value for r 2
on the graph.

f(x) = -0.5077143644 ln(x) + 4.2762264009

R = 0.9941917142

1
1
1

Colum
nC

1
1

row 720

0
0
0
-1000

1000

2000

3000

4000

5000

Figure 14-19 Using Excel's add trendline function to fit a curve.

row 730