NCSS Statistical Software NCSS.

com

Chapter 306

Multiple Regression
with Serial Correlation
Introduction
The regular Multiple Regression routine assumes that the random-error components are independent from one
observation to the next. However, this assumption is often not appropriate for business and economic data.
Instead, it is more appropriate to assume that the error terms are positively correlated over time. These are called
autocorrelated or serially correlated data.
Consequences of the error terms being serially correlated include inefficient estimation of the regression
coefficients, under estimation of the error variance (MSE), under estimation of the variance of the regression
coefficients, and inaccurate confidence intervals.
The presence of serial correlation can be detected by the Durbin-Watson test and by plotting the residuals against
their lags.

Autoregressive Error Model

When serial correlation is detected, there are several remedies. Since autocorrelation is often caused by leaving
important independent variables out of the regression model, an obvious remedy is to add other, appropriate
independent variables to the model. When this is not possible, another remedy is to use an autoregressive model.
The usual multiple regression model
Yt = β 0 + β 1 X 1t + β 2 X 2 t +  + β p X pt + ε t
is modified by adding the equation

ε t = ρε t −1 + ut
where
ρ < 1 is the serial correlation

ut ~ N (0, σ 2 )
The subscript t represents the time period. In econometric work, these u’s are often called the disturbances. They
are the ultimate error terms. Further details on this model can be found in chapter 12 of Neter, Kutner,
Nachtsheim, and Wasserman (1996).

306-1
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Cochrane-Orcutt Procedure
Several methods have been suggested to estimate the autoregressive error model. We have adopted the Cochrane-
Orcutt procedure as given in Neter, Kutner, Nachtsheim, and Wasserman (1996). This is an iterative procedure
that involves several steps.
1. Ordinary least squares. The regression coefficients are estimated using ordinary least squares. The array
of residuals is calculated.

( )
2. Estimation of ρ . The serial correlation is estimated from the current residuals et = Yt − Yˆt using the
formula
n

∑e e t t −1
ρ̂ = t =2
n

∑e
t =2
2
t −1

3. Obtain transformed data. A new set of data is created using the formulas.
Yt′ = Yt − ρˆYt −1
X 1′t = X 1t − ρˆX 1,t −1

X ′pt = X pt − ρˆX p ,t −1
4. Fit model to transformed data. Ordinary least squares is used to fit the following multiple regression to
the transformed data.
Yt′ = b0′ + b1′ X 1t + b2′ X 2 t +  + b′p X pt
5. Create the regression model for the untransformed data. The regression equation of the untransformed
data is created using the following equations.
b0′
b0 =
1 − ρˆ
b1 = b1′
b2 = b2′

b p = b′p
The estimated standard errors of the regression coefficients are given by
s (b0′ )
s (b0 ) =
1 − ρˆ
s (b1 ) = s (b1′ )
s (b2 ) = s (b2′ )

s (b p ) = s (b′p )
6. Iterate until convergence is reached. Steps 2 – 4 are then repeated until the value of P stabilizes. Usually,
only four or five iterations are necessary.

306-2
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

7. Calculate Durbin-Watson test on transformed residuals. As a final diagnostic check, the Durbin-Watson
( )
test may be run on the residuals et′ = Yt′ − Yˆt′ from the transformed regression model.

Durbin-Watson Test
The Durbin-Watson test is often used to test for positive or negative, first-order, serial correlation. It is calculated
as follows

∑ (e )
N
2
j − e j −1
j=2
DW = N

∑e
j =1
2
j

The distribution of this test is difficult because it involves the X values. Originally, Durbin-Watson (1950, 1951) gave
a pair of bounds to be used. However, there is a large range of ‘inclusion’ found when using these bounds. Instead of
using these bounds, we calculate the exact probability using the beta distribution approximation suggested by Durbin-
Watson (1951). This approximation has been shown to be accurate to three decimal places in most cases which is all
that are needed for practical work.

Forecasts
The predicted value for a specific set of independent variable values is given by

Yˆt = bˆ0 + bˆ1 X 1t + bˆ2 X 2 t +  + bˆ p X pt

For forecasts j periods into the future after the end of the series (period n is the final period on which we have
data), the formula is

Fn + j = bˆ0 + bˆ1 X 1,n + j + bˆ2 X 2,n + j +  + bˆ p X p ,n + j + ρ̂ j en

where en is the residual from the final observation. That is,

en = Yn − Yˆn
The approximate 1 − α prediction interval for this forecast is
Fn + j ± t1−α / 2,n −3 s F

where s F is the standard error of the prediction interval based on the transformed data.

Data Structure
The data are entered in two or more variables. An example of data appropriate for this procedure is shown below.
These data give the annual values for several economic statistics. Later in this chapter, these data will be used in
an example in which Housing is forecast from Mort5Yr and DispInc. These data are stored in a dataset called
Housing.NCSS. Note that only two decimal places are displayed here, while on the database, more decimal places
are stored.

306-3
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Housing dataset (subset)

Year Housing Mort5Yr DispInc TBill Unemp_rt
1981 403.34 18.25 27006.90 17.72 7.57
1982 407.92 17.93 26896.58 13.66 10.97
1983 446.87 13.17 26582.63 9.31 11.94
1984 457.22 13.54 27662.41 11.06 11.30
1985 485.25 12.08 28710.13 9.43 10.65
1986 475.87 11.17 29057.02 8.97 9.64
1987 491.30 11.12 29626.58 8.15 8.82
1988 493.23 11.61 31070.52 9.48 7.75
1989 487.14 12.01 32417.38 12.05 7.55
1990 491.00 13.31 32683.10 12.81 8.12
1991 512.39 11.07 31980.30 8.73 10.32
1992 523.07 9.50 32224.67 6.59 11.16
1993 533.20 8.76 32412.84 4.84 11.36
1994 497.75 9.53 32789.41 5.54 10.36
1995 502.59 9.14 33242.99 6.89 9.45
1996 522.73 7.91 33256.65 4.21 9.64
1997 538.72 7.05 33839.28 3.26 9.10
1998 533.61 6.92 34915.04 4.73 8.29
1999 531.89 7.54 35971.46 4.72 7.57
2000 528.09 8.32 37566.34 5.49 6.81
2001 544.91 7.38 38228.92 3.77 7.20
2002 547.70 6.99 38806.22 2.59 7.66
2003 561.19 6.36 38896.05 2.87 7.63
2004 581.54 5.38 39870.12 2.30 7.45
2005 6.00 41000.00
2006 6.25 42000.00

Procedure Options
This section describes the options available in this procedure.

Variables Tab
This panel specifies the variables used in the analysis.

Dependent Variable
Y: Dependent Variable(s)
This option specifies one or more dependent (Y) variables. If more than one variable is specified, a separate
analysis is run for each.

Numeric Independent Variables

X’s: Numeric Independent Variable(s)
Specify any numeric independent variables in this box. Numeric variables are those whose values are numeric and
are at least ordinal. Nominal variables, even when coded with numbers, should be specified as Categorical
Independent Variables. Although you may specify binary (0-1) variables here, they are more appropriately
analyzed when you specify them as Categorical Independent Variables.

306-4
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

If you want to create powers and cross-products of these variables, specify an appropriate model in the ‘Custom
Model’ field under the Model tab.
If you want to create predicted values of Y for values of X not in your database, add the X values to the bottom of
the database. These rows will not be used during estimation phase, but predicted values will be generated for them
on the reports.

Categorical Independent Variables

X’s: Categorical Independent Variable(s)
Specify categorical (nominal) independent variables in this box. By categorical we mean that the variable has only
a few unique, numeric or text, values like 1, 2, 3 or Yes, No, Maybe. The values are used to identify categories.
For further details about categorical variables, see the discussion on this topic in the Multiple Regression chapter.
Default Reference Value
This option specifies the default reference value to be used when automatically generating indicator variables
during the processing of selected categorical independent variables. The reference value is often the baseline, and
the other values are compared to it. For further details about this value, see the discussion on this topic in the
Multiple Regression chapter.
Default Contrast Type
Select the default type of numeric variable that will be generated when processing categorical independent
variables. The values in a categorical variable are not used directly in regression analysis. Instead, a set of
numeric variables is automatically created and substituted for them. This option allows you to specify what type
of numeric variable will be created. For further details about this option, see the discussion on this topic in the
Multiple Regression chapter.

Estimation Options
Maximum Cochrane-Orcutt Iterations
This is the maximum number of iterations that the procedure will cycle through. Some authors recommend only
one iteration. Others recommend stopping once the Durbin-Watson test is not significant. This option lets you
stop after a specific number of iterations. Usually, four or five iterations should be plenty.
Minimum Rho Change
If the change is rho (serial correlation) from one iteration to the next is less than this amount, the algorithm will
stop iterating. We suggest you use a small amount such as 0.00001.

Alpha Levels
Alpha of C.I.’s and Tests
The value of alpha for the statistical tests and confidence intervals is specified here. Usually, this number will
range from 0.1 to 0.001. A common choice for alpha is 0.05, but this value is a legacy from the age before
computers when only printed tables were available. You should determine a value appropriate for your particular
study.
Alpha of Assumptions
This value specifies the significance level that must be achieved to reject a preliminary test of an assumption. In
regular hypothesis tests, common values of alpha are 0.05 and 0.01. However, most statisticians recommend that
preliminary tests use a larger alpha such as 0.10, 0.15, or 0.20.
We recommend 0.20.

306-5
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Model Tab
These options control the regression model.

Model Specification
Which Model Terms
This option specifies which terms (terms, powers, cross-products, and interactions) are included in the regression
model. For a time-series regression model, select Up to 1-Way.
The other options on this tab are covered in detail in the Multiple Regression chapter. We refer you to that chapter
for further details.

Reports Tab
The following options control which reports and plots are displayed.

Select Reports
Run Summary ... Residuals
Each of these options specifies whether the indicated report is calculated and displayed. Note that since some of
these reports provide results for each row, they may be too long for normal use when requested on large
databases.

Report Options
Show All Rows
This option makes it possible to display predicted values for only a few designated rows.
When checked predicted values, residuals, and other row-by-row statistics, will be displayed for all rows used in
the analysis.
When not checked, predicted values and other row-by-row statistics will be displayed for only those rows in
which the dependent variable’s value is missing.
Precision
Specifies the precision of numbers in the report. Single precision will display seven-place accuracy, while the
double precision will display thirteen-place accuracy.
Variable Names
This option lets you select whether to display variable names, variable labels, or both.
Skip Line After
The names of the indicator variables can be too long to fit in the space provided. If the name contains more
characters than the number specified here, only the name is shown on the first line of the report and the rest of the
output is placed on the next line.
Enter 1 when you want the each variable’s results printed on two lines.
Enter 100 when you want each variable’s results printed on a single line.

306-6
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Report Options – Decimal Places

Probability ... Mean Square Decimals
Specify the number of digits after the decimal point to display on the output of values of this type. Note that this
option in no way influences the accuracy with which the calculations are done.
Enter 'All' to display all digits available. The number of digits displayed by this option is controlled by whether
the PRECISION option is SINGLE or DOUBLE.

Plots Tab
These options control the titles and style files used on each of the plots.

Select Plots
Histogram ... Residuals vs X Plot
Indicate whether to display these plots. Click the plot format button to change the plot settings.

Storage Tab
These options let you specify if, and where on the dataset, various statistics are stored.
Warning: Any data already in these columns are replaced by the new data. Be careful not to specify columns that
contain important data.

Data Storage Options

Storage Option
This option controls whether the values indicated below are stored on the database when the procedure is run.

• Do not store data

No data are stored even if they are checked.

• Store in empty columns only

The values are stored in empty columns only. Columns containing data are not used for data storage, so no
data can be lost.

• Store in designated columns

Beginning at the Store First Item In column, the values are stored in this column and those to the right. If a
column contains data, the data are replaced by the storage values. Care must be used with this option because
it cannot be undone.
Store First Item In
The first item is stored in this column. Each additional item that is checked is stored in the columns immediately
to the right of this column.
Leave this value blank if you want the data storage to begin in the first blank column on the right-hand side of the
data.
Warning: any existing data in these columns is automatically replaced, so be careful.

306-7
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Data Storage Options – Select Items to Store

Predicted Y ... Upper C.L. Individual
Indicate whether to store these row-by-row values, beginning at the column indicated by the Store First Item In
option.

Example 1 – Generating Forecasts (All Reports)

This section presents an example of how to generate forecasts for housing data that was presented earlier in this
chapter. This data is stored in the Housing dataset. We suggest that you open it now.
This example will run an adjusted multiple regression of Housing on Mort5Yr and DispInc. The adjustment will
use the Cochrane-Orcutt procedure. The data for housing ends in 2004. Forecasts will be generated for the years
2005 and 2006.
You may follow along here by making the appropriate entries or load the completed template Example 1 by
clicking on Open Example Template from the File menu of the Multiple Regression with Serial Correlation
window.

1 Open the Housing dataset.

• From the File menu of the NCSS Data window, select Open Example Data.
• Click on the file Housing.NCSS.
• Click Open.

2 Open the Multiple Regression with Serial Correlation window.

• Using the Analysis menu or the Procedure Navigator, find and select the Multiple Regression with
Serial Correlation procedure.
• On the menus, select File, then New Template. This will fill the procedure with the default template.

3 Specify the variables.

• On the Multiple Regression with Serial Correlation window, select the Variables tab.
• Set the Y: Dependent Variables box to Housing.
• Set the X’s: Numeric Independent Variables box to DispInc, Mort5Yr.
• Set the Maximum Cochrane-Orcutt Iterations to 1.

4 Specify the reports.

• Select the Reports tab.
• Make sure all reports and plots are checked.

5 Run the procedure.

• From the Run menu, select Run Procedure. Alternatively, just click the green Run button.

306-8
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Run Summary Section

Parameter Value Parameter Value
Dependent Variable Housing Rows Processed 32
Number Ind. Variables 2 Rows Filtered Out 0
Weight Variable None Rows with X's Missing 0
R2 0.8941 Rows with Weight Missing 0
Adj R2 0.8860 Rows with Y Missing 2
Coefficient of Variation 0.0361 Rows Used in Estimation 30
Mean Square Error 77.15598 Sum of Weights 29.000
Square Root of MSE 8.783848 Completion Status Normal Completion
Ave Abs Pct Error 1.801 Autocorrelation (Rho) 0.5121

This report summarizes the multiple regression results. It presents the variables used, the number of rows used,
and the basic results. The estimated value of the autocorrelation (rho) has been added to this report. Otherwise, it
is identical to the corresponding report in the regular Multiple Regression report.
Note that values such as R2, Mean Square Error, etc., are calculated on the transformed data.

Descriptive Statistics Section

Standard
Variable Count Mean Deviation Minimum Maximum
DispInc 29 31000.94 5157.438 21780.1 39870.13
Mort5Yr 29 10.53919 3.183494 5.380194 18.25095
Housing 29 491.0214 48.53722 403.3378 581.5398

For each variable, the count, arithmetic mean, standard deviation, minimum, and maximum are computed. This
report is particularly useful for checking that the correct variables were selected.

Correlation Matrix Section

DispInc Mort5Yr Housing
DispInc 1.0000 -0.5962 0.7913
Mort5Yr -0.5962 1.0000 -0.8874
Housing 0.7913 -0.8874 1.0000

Pearson correlations are given for all variables.

Regression Equation Section

Regression Standard T-Value Reject
Independent Coefficient Error to test Prob H0 at
Variable b(i) Sb(i) H0:B(i)=0 Level 5%?
Intercept 445.1365 35.4690 12.550 0.0000 Yes
DispInc 0.0044 0.0009 5.118 0.0000 Yes
Mort5Yr -8.5371 1.0522 -8.113 0.0000 Yes

Estimated Model
445.136489079996+ 4.4434007069797E-03*DispInc-8.53714263704248*Mort5Yr

This section reports the values and significance tests of the regression coefficients. Note that the intercept has
been corrected by dividing by 1-rho. Other than this, the report has the same definitions as in regular Multiple
Regression.

306-9
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Regression Coefficient Section

Independent Regression Standard Lower Upper Standardized
Variable Coefficient Error 95% C.L. 95% C.L. Coefficient
Intercept 445.1365 35.4690 372.2289 518.0441 0.0000
DispInc 0.0044 0.0009 0.0027 0.0062 0.4068
Mort5Yr -8.5371 1.0522 -10.7000 -6.3743 -0.6449
Note: The T-Value used to calculate these confidence limits was 2.056.

The report has the same definitions as in regular Multiple Regression.

Analysis of Variance Section

Sum of Mean Prob
Source DF R2 Squares Square F-Ratio Level
Intercept 1 1720724 1720724
Model 2 0.8941 16943.61 8471.806 109.801 0.0000
Error 26 0.1059 2006.055 77.15598
Total(Adjusted) 28 1.0000 18949.67 676.7738

This section reports the analysis of variance table. Note it was calculated from the transformed data on the last
iteration. Other than this, the report has the same definitions as in regular Multiple Regression.

Serial-Correlation and Durbin-Watson Test

Serial Correlation of Residuals from Uncorrected Model
Serial Serial Serial
Lag Correlation Lag Correlation Lag Correlation
1 0.5090 9 -0.4075 17 -0.1140
2 0.1980 10 -0.5085 18 -0.0147
3 0.0802 11 -0.3018 19 0.1512
4 0.0505 12 -0.1962 20 0.1290
5 0.2072 13 -0.1042 21 0.0519
6 0.2165 14 -0.1067 22 0.0275
7 -0.0649 15 -0.3178 23 0.0457
8 -0.0979 16 -0.2177 24 0.0875
Above serial correlations significant if their absolute values are greater than 0.365148

Serial Correlation of Residuals from Corrected Model

Serial Serial Serial
Lag Correlation Lag Correlation Lag Correlation
1 0.0261 9 -0.2371 17 0.0817
2 -0.0349 10 -0.3626 18 0.0420
3 0.0972 11 -0.0584 19 0.0314
4 -0.1182 12 0.0042 20 0.0473
5 0.1002 13 -0.0671 21 0.0248
6 0.2071 14 -0.0042 22 0.0761
7 -0.3095 15 -0.2443 23 0.0038
8 0.1301 16 0.0617 24 0.0388
Above serial correlations significant if their absolute values are greater than 0.371391

Durbin-Watson Test For Serial Correlation of Uncorrected Model

Did the Test Reject
Parameter Value H0: Rho(1) = 0?
Durbin-Watson Value 0.9234
Prob. Level: Positive Serial Correlation 0.0002 Yes
Prob. Level: Negative Serial Correlation 0.9974 No

306-10
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Durbin-Watson Test For Serial Correlation of Corrected Model

Did the Test Reject
Parameter Value H0: Rho(1) = 0?
Durbin-Watson Value 1.9221
Prob. Level: Positive Serial Correlation 0.3273 No
Prob. Level: Negative Serial Correlation 0.4923 No

This section reports the autocorrelation structure of the residuals both before and after the model is corrected for
serial correlation. It has the same definitions as in the regular Multiple Regression report.

Predicted Values with Confidence Limits of Means

Standard 95% Lower 95% Upper
Actual Predicted Error Of Conf. Limit Conf. Limit
Row Housing Housing Predicted Of Mean Of Mean
1 420.722 445.738
2 431.522 447.504 3.273 440.776 454.232
3 448.085 462.874 4.140 454.364 471.384
. . . . . .
. . . . . .
. . . . . .
26 528.086 541.005 3.211 534.406 547.605
27 544.913 552.012 2.883 546.086 557.938
28 547.703 557.854 3.010 551.667 564.041
29 561.186 563.635 2.950 557.571 569.698
30 581.540 576.364 3.536 569.095 583.633
31 578.744 3.901 570.725 586.762
32 579.760 4.192 571.143 588.377

Confidence intervals for the mean response of Y given specific levels for the IV’s are provided here.

Predicted Values with Prediction Limits of Individuals

Standard 95% Lower 95% Upper
Actual Predicted Error Of Pred. Limit Pred. Limit
Row Housing Housing Predicted Of Individual Of Individual
1 420.722 445.738
2 431.522 447.504 9.374 428.235 466.772
3 448.085 462.874 9.711 442.914 482.835
. . . . . .
. . . . . .
. . . . . .
26 528.086 541.005 9.352 521.782 560.229
27 544.913 552.012 9.245 533.009 571.015
28 547.703 557.854 9.285 538.768 576.940
29 561.186 563.635 9.266 544.588 582.681
30 581.540 576.364 9.469 556.900 595.828
31 578.744 9.611 558.988 598.500
32 579.760 9.733 559.753 599.766

A prediction interval for the individual response of Y given specific values of the IV’s is provided here for each
row. Note that the forecasts start where the actual housing values are blank.

306-11
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Residual Report
Absolute
Actual Predicted Percent
Row Housing Housing Residual Error
1 420.722 445.738
2 431.522 447.504 -15.982 3.704
3 448.085 462.874 -14.789 3.300
4 447.923 464.696 -16.773 3.745
5 451.401 454.340 -2.939 0.651
6 432.474 438.269 -5.795 1.340
7 403.338 409.328 -5.990 1.485
8 407.922 411.549 -3.627 0.889
. . . . .
. . . . .
. . . . .

This section reports on the sample residuals, or ei’s.

Histogram
The purpose of the histogram and density trace of the residuals is to evaluate whether they are normally
distributed. A dot plot is also given that highlights the distribution of points in each bin of the histogram. Unless
you have a large sample size, it is best not to rely on the histogram for visually evaluating normality of the
residuals. The better choice would be the normal probability plot.

306-12
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Probability Plot of Residuals

If the residuals are normally distributed, the data points of the normal probability plot will fall along a straight line
through the origin with a slope of 1.0. Major deviations from this ideal picture reflect departures from normality.
Stragglers at either end of the normal probability plot indicate outliers, curvature at both ends of the plot indicates
long or short distributional tails, convex or concave curvature indicates a lack of symmetry, and gaps or plateaus
or segmentation in the normal probability plot may require a closer examination of the data or model. Of course,
use of this graphic tool with very small sample sizes is not recommended.
If the residuals are not normally distributed, then the t-tests on regression coefficients, the F-tests, and any interval
estimates are not valid. This is a critical assumption to check.

Y vs X’s Plots
Actually, a regression analysis should always begin with a plot of Y versus each IV. These plots often show
outliers, curvilinear relationships, and other anomalies.

306-13
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Serial Correlation Plot: Residuals vs Lagged Residuals

This is a scatter plot of the jth residual versus the jth-1 residual. The purpose of this plot is to check for first-order
autocorrelation. Positive autocorrelation or serial correlation means that the residual in time period j tends to have
the same sign as the residual in time period (j-1). On the other hand, a strong negative autocorrelation means that
the residual in time period j tends to have the opposite sign as the residual in time period (j-1).

Sequence Plot: Residuals vs Row

Sequence plots may be useful in finding variables that are not accounted for by the regression equation. They are
especially useful if the data were taken over time.

306-14
© NCSS, LLC. All Rights Reserved.
NCSS Statistical Software NCSS.com
Multiple Regression with Serial Correlation

Residuals vs Yhat Plot

This plot should always be examined. The preferred pattern to look for is a point cloud or a horizontal band. A
wedge or bowtie pattern is an indicator of nonconstant variance, a violation of a critical regression assumption.
The sloping or curved band signifies inadequate specification of the model. The sloping band with increasing or
decreasing variability suggests nonconstant variance and inadequate specification of the model.

Residuals vs X’s Plots

This is a scatter plot of the residuals versus each independent variable. Again, the preferred pattern is a
rectangular shape or point cloud. Any other nonrandom pattern may require a redefining of the regression model.

306-15
© NCSS, LLC. All Rights Reserved.