SPSS for Instruction

SPSS Books and Manuals

Learn how to more productively and efficiently work with SPSS software. A number of SPSS books and manualsfor beginners and expert users alikecan show you how.
Below are descriptions of books and manuals of interest to SPSS users. Use these books to teach yourself SPSS or supplement techniques youve learned in the classroom.
SPSS 13.0 Guide to Data Analysis by Marija Norusis SPSS 13.0 Guide to Data Analysis is a friendly introduction to both data analysis and SPSS software. Easy-to-understand explanations and in-depth content make this guide both an excellent supplement to other statistics texts and an ideal primary text for any introductory data analysis course. With this book, you will learn how to describe data, test simple hypotheses, and examine relationships using real datasets. Exercises at the end of each chapter let you test your skills. SPSS 13.0 Guide to Data Analysis is designed for use with SPSS 13.0 for Windows, including the SPSS Student Version. A data CD is included with this book. To view a description of this book, table of contents, and preface, or to purchase a copy, visit www.prenhall.com. ISBN: 0-13-186535-8
SPSS 13.0 Statistical Procedures Companion by Marija Norusis SPSS 13.0 Statistical Procedures Companion is an introduction to SPSS and its most frequently used statistical procedures. The book takes you from creating, cleaning, and describing data files to analyzing the data using sophisticated statistical procedures such as logistic regression; factor, cluster, and discriminant analysis; log linear models; and general linear models. Youll find discussion of the statistical backgrounds for each procedure and a review of the basics of hypothesis testing. The book offers practical suggestions, emphasizing topics that arise when analyzing real data for presentations, reports, and dissertations. A data CD is included with this book. To view a description of this book, table of contents, and preface, or to purchase a copy, visit www.prenhall.com. ISBN: 0-13-186539-0
SPSS 13.0 Advanced Statistical Procedures Companion by Marija Norusis SPSS 13.0 Advanced Statistical Procedures Companion covers the procedures in two popular SPSS add-on modules, SPSS Advanced Models and SPSS Regression Models. The book provides introductions to and examples for procedures such as loglinear and logit analysis for categorical data; ordinal, multinomial, two-stage, and weighted least squares regression; Kaplan-Meier and Cox regression models for survival analysis; and variance components analysis. A data CD is included with this book. To view a description of this book, table of contents, and preface, or to purchase a copy, visit www.prenhall.com. ISBN: 0-13-186540-4
SPSS 13.0 Brief Guide SPSS 13.0 Brief Guide uses a convenient tutorial system to acquaint beginning users with the components of the SPSS system. Learn how to use the Data Editor, import data into SPSS, work with statistics and output, create and edit charts, modify data values, manage syntax and data files, calculate new data values, and sort and select data. To view a description of this book and table of contents, or to purchase this book, visit www.spss.com/estore/softwaremenu/book.cfm. ISBN: 0-13-154242-7
SPSS Base 13.0 User's Guide SPSS Base 13.0 User's Guide provides a thorough explanation of SPSS features. Topics covered include the Text Wizard,
SPSS Programming and Data Management: A Guide for SPSS and SAS Users, Second Edition by Raynald Levesque SPSS Programming and Data Management: A Guide for SPSS and SAS Users, Second Edition documents the wealth of functionality beneath the SPSS user interface. It includes detailed examples of command syntax, the macro facility, scripting, and the output management system. The accompanying CD-ROM includes the command and data files used in the book. The book also contains a chapter for SAS users, showing equivalent SPSS code for many common data management tasks. With knowledge gained from this book, you will be able to use the many tools available within SPSS to import data from almost any source, clean it, transform it, merge it with other data, and get it into the condition required to produce reliable models and informative results. To view the complete table of contents or to purchase this book, visit www.spss.com/estore/softwaremenu/book.cfm. ISBN: 0-56827-355-X

Database Wizard, Data Editor, scripting, data definition and modification, file and output management (including the SPSS Viewer and report cubes), statistical and graphical procedures (including output examples), production mode operation, and utilities for getting information (including help) and controlling the environment. To view a description of this book and table of contents, or to purchase this book, visit www.spss.com/estore/softwaremenu/book.cfm. ISBN: 0-13-185723-1
Manuals for add-on modules and stand-alone software User manuals are available for all SPSS add-on modules and many stand-alone products, such as Amos. Visit www.spss.com/estore/softwaremenu/book.cfm for a full list of available manuals, or to make a purchase.
To learn more, please visit www.spss.com. For SPSS office locations and telephone numbers, go to www.spss.com/worldwide.
SPSS is a registered trademark and the other SPSS products named are trademarks of SPSS Inc. All other names are trademarks of their respective owners. 2005 SPSS Inc. SB13INS-0405

Customer Service

If you have any questions concerning your shipment or account, contact your local office, listed on the SPSS Web site at http://www.spss.com/worldwide. Please have your serial number ready for identification.

Training Seminars

SPSS Inc. provides both public and onsite training seminars. All seminars feature hands-on workshops. Seminars will be offered in major cities on a regular basis. For more information on these seminars, contact your local office, listed on the SPSS Web site at http://www.spss.com/worldwide.

Technical Support

The services of SPSS Technical Support are available to registered customers. Customers may contact Technical Support for assistance in using SPSS or for installation help for one of the supported hardware environments. To reach Technical Support, see the SPSS Web site at http://www.spss.com, or contact your local office, listed on the SPSS Web site at http://www.spss.com/worldwide. Be prepared to identify yourself, your organization, and the serial number of your system.

Additional Publications

Additional copies of SPSS product manuals may be purchased directly from SPSS Inc. Visit the SPSS Web Store at http://www.spss.com/estore, or contact your local SPSS office, listed on the SPSS Web site at http://www.spss.com/worldwide. For telephone orders in the United States and Canada, call SPSS Inc. at 800-543-2185. For telephone orders outside of North America, contact your local office, listed on the SPSS Web site. The SPSS Statistical Procedures Companion, by Marija Noruis, has been published by Prentice Hall. A new version of this book, updated for SPSS 13.0, is planned. The SPSS Advanced Statistical Procedures Companion, also based on SPSS 13.0, is forthcoming. The SPSS Guide to Data Analysis for SPSS 13.0 is also in development. Announcements of publications available exclusively through Prentice Hall will be available on the SPSS Web site at http://www.spss.com/estore (select your home country, and then click Books).

Tell Us Your Thoughts

Your comments are important. Please let us know about your experiences with SPSS products. We especially like to hear about new and interesting applications using the SPSS system. Please send e-mail to suggest@spss.com or write to SPSS Inc.,
Attn.: Director of Product Planning, 233 South Wacker Drive, 11th Floor, Chicago, IL 60606-6412.

About This Manual

This manual documents the graphical user interface for the procedures included in the Regression Models add-on module. Illustrations of dialog boxes are taken from SPSS for Windows. Dialog boxes in other operating systems are similar. Detailed information about the command syntax for features in this module is provided in the SPSS Command Syntax Reference, available from the Help menu.

Contacting SPSS

If you would like to be on our mailing list, contact one of our offices, listed on our Web site at http://www.spss.com/worldwide.

Contents

1 Choosing a Procedure for Binary Logistic Regression Models Logistic Regression
Logistic Regression Set Rule. 6 Logistic Regression Variable Selection Methods. 6 Logistic Regression Define Categorical Variables. 7 Logistic Regression Save New Variables. 9 Logistic Regression Options. 10 LOGISTIC REGRESSION Command Additional Features. 11
Multinomial Logistic Regression
Multinomial Logistic Regression Models. 15 Multinomial Logistic Regression Reference Category. 17 Multinomial Logistic Regression Statistics. 18 Multinomial Logistic Regression Criteria. 20 Multinomial Logistic Regression Options. 21 Multinomial Logistic Regression Save. 23 NOMREG Command Additional Features. 24

Probit Analysis

Probit Analysis Define Range. 27 Probit Analysis Options. 28 PROBIT Command Additional Features. 29

Nonlinear Regression

Conditional Logic (Nonlinear Regression). 33 Nonlinear Regression Parameters. 34 Nonlinear Regression Common Models. 35 Nonlinear Regression Loss Function. 36 Nonlinear Regression Parameter Constraints. 37 Nonlinear Regression Save New Variables. 38 Nonlinear Regression Options. 38 Interpreting Nonlinear Regression Results. 39 NLR Command Additional Features. 40

Logistic Regression

Logistic regression is useful for situations in which you want to be able to predict the presence or absence of a characteristic or outcome based on values of a set of predictor variables. It is similar to a linear regression model but is suited to models where the dependent variable is dichotomous. Logistic regression coefficients can be used to estimate odds ratios for each of the independent variables in the model. Logistic regression is applicable to a broader range of research situations than discriminant analysis.
Example. What lifestyle characteristics are risk factors for coronary heart disease
(CHD)? Given a sample of patients measured on smoking status, diet, exercise, alcohol use, and CHD status, you could build a model using the four lifestyle variables to predict the presence or absence of CHD in a sample of patients. The model can then be used to derive estimates of the odds ratios for each factor to tell you, for example, how much more likely smokers are to develop CHD than nonsmokers.
Statistics. For each analysis: total cases, selected cases, valid cases. For each
categorical variable: parameter coding. For each step: variable(s) entered or removed, iteration history, 2 log-likelihood, goodness of fit, Hosmer-Lemeshow goodness-of-fit statistic, model chi-square, improvement chi-square, classification table, correlations between variables, observed groups and predicted probabilities chart, residual chi-square. For each variable in the equation: coefficient (B), standard error of B, Wald statistic, estimated odds ratio (exp(B)), confidence interval for exp(B), log-likelihood if term removed from model. For each variable not in the equation: score statistic. For each case: observed group, predicted probability, predicted group, residual, standardized residual.
Methods. You can estimate models using block entry of variables or any of the
following stepwise methods: forward conditional, forward LR, forward Wald, backward conditional, backward LR, or backward Wald.

4 Chapter 2

Data. The dependent variable should be dichotomous. Independent variables can be interval level or categorical; if categorical, they should be dummy or indicator coded (there is an option in the procedure to recode categorical variables automatically). Assumptions. Logistic regression does not rely on distributional assumptions in the same sense that discriminant analysis does. However, your solution may be more stable if your predictors have a multivariate normal distribution. Additionally, as with other forms of regression, multicollinearity among the predictors can lead to biased estimates and inflated standard errors. The procedure is most effective when group membership is a truly categorical variable; if group membership is based on values of a continuous variable (for example, high IQ versus low IQ), you should consider using linear regression to take advantage of the richer information offered by the continuous variable itself. Related procedures. Use the Scatterplot procedure to screen your data for multicollinearity. If assumptions of multivariate normality and equal variance-covariance matrices are met, you may be able to get a quicker solution using the Discriminant Analysis procedure. If all of your predictor variables are categorical, you can also use the Loglinear procedure. If your dependent variable is continuous, use the Linear Regression procedure. You can use the ROC Curve procedure to plot probabilities saved with the Logistic Regression procedure. Obtaining a Logistic Regression Analysis

9 Logistic Regression

Logistic Regression Save New Variables
Figure 2-4 Logistic Regression Save New Variables dialog box
You can save results of the logistic regression as new variables in the working data file:
Predicted Values. Saves values predicted by the model. Available options are
Probabilities and Group membership.
Probabilities. For each case, saves the predicted probability of occurrence of the
event. A table in the output displays name and contents of any new variables.
Predicted Group Membership. The group with the largest posterior probability,
based on discriminant scores. The group the model predicts the case belongs to.
Influence. Saves values from statistics that measure the influence of cases on predicted
values. Available options are Cooks, Leverage values, and DfBeta(s).
Cook's. The logistic regression analog of Cook's influence statistic. A measure
of how much the residuals of all cases would change if a particular case were excluded from the calculation of the regression coefficients.
Leverage Value. The relative influence of each observation on the model's fit. DfBeta(s). The difference in beta value is the change in the regression coefficient
that results from the exclusion of a particular case. A value is computed for each term in the model, including the constant.
Residuals. Saves residuals. Available options are Unstandardized, Logit, Studentized, Standardized, and Deviance.

10 Chapter 2

Unstandardized Residuals. The difference between an observed value and the
value predicted by the model.
Logit Residual. The residual for the case if it is predicted in the logit scale. The
logit residual is the residual divided by the predicted probability times 1 minus the predicted probability.
Studentized Residual. The change in the model deviance if a case is excluded. Standardized Residuals. The residual divided by an estimate of its standard
deviation. Standardized residuals which are also known as Pearson residuals, have a mean of 0 and a standard deviation of 1.

14 Chapter 3

Obtaining a Multinomial Logistic Regression
E From the menus choose: Analyze Regression Multinomial Logistic. Figure 3-1 Multinomial Logistic Regression dialog box
E Select one dependent variable. E Factors are optional and can be either numeric or categorical. E Covariates are optional but must be numeric if specified.
15 Multinomial Logistic Regression
Multinomial Logistic Regression Models
Figure 3-2 Multinomial Logistic Regression Model dialog box
By default, the Multinomial Logistic Regression procedure produces a model with the factor and covariate main effects, but you can specify a custom model or request stepwise model selection with this dialog box.
Specify Model. A main-effects model contains the covariate and factor main effects but no interaction effects. A full factorial model contains all main effects and all factor-by-factor interactions. It does not contain covariate interactions. You can create a custom model to specify subsets of factor interactions or covariate interactions, or request stepwise selection of model terms. Factors and Covariates. The factors and covariates are listed with (F) for factor and

(C) for covariate.

Forced Entry Terms. Terms added to the forced entry list are always included in the

model.

16 Chapter 3
Stepwise Terms. Terms added to the stepwise list are included in the model according to one of the following user-selected methods: Forward entry. This method begins with no stepwise terms in the model. At each
step, the most significant term is added to the model until none of the stepwise terms left out of the model would have a statistically significant contribution if added to the model.
Backward elimination. This method begins by entering all terms specified on the
stepwise list into the model. At each step, the least significant stepwise term is removed from the model until all of the remaining stepwise terms have a statistically significant contribution to the model.
Forward stepwise. This method begins with the model that would be selected by

slope.

Fiducial Confidence Intervals. Confidence intervals for the dosage of agent
required to produce a certain probability of response. Fiducial confidence intervals and Relative median potency are unavailable if you have selected more than one covariate. Relative median potency and Parallelism test are available only if you have selected a factor variable.
Natural Response Rate. Allows you to indicate a natural response rate even in the
absence of the stimulus. Available alternatives are None, Calculate from data, or Value.

29 Probit Analysis

Calculate from Data. Estimate the natural response rate from the sample data. Your
data should contain a case representing the control level, for which the value of the covariate(s) is 0. Probit estimates the natural response rate using the proportion of responses for the control level as an initial value.
Value. Sets the natural response rate in the model (select this item when you know
the natural response rate in advance). Enter the natural response proportion (the proportion must be less than 1). For example, if the response occurs 10% of the time when the stimulus is 0, enter 0.10.
Criteria. Allows you to control parameters of the iterative parameter-estimation
algorithm. You can override the defaults for Maximum iterations, Step limit, and Optimality tolerance.
PROBIT Command Additional Features
The SPSS command language also allows you to: Request an analysis on both the probit and logit models. Control the treatment of missing values. Transform the covariates by bases other than base 10 or natural log. See the SPSS Command Syntax Reference for complete syntax information.
Nonlinear regression is a method of finding a nonlinear model of the relationship between the dependent variable and a set of independent variables. Unlike traditional linear regression, which is restricted to estimating linear models, nonlinear regression can estimate models with arbitrary relationships between independent and dependent variables. This is accomplished using iterative estimation algorithms. Note that this procedure is not necessary for simple polynomial models of the form Y = A + BX**2. By defining W = X**2, we get a simple linear model, Y = A + BW, which can be estimated using traditional methods such as the Linear Regression procedure.
Example. Can population be predicted based on time? A scatterplot shows that there seems to be a strong relationship between population and time, but the relationship is nonlinear, so it requires the special estimation methods of the Nonlinear Regression procedure. By setting up an appropriate equation, such as a logistic population growth model, we can get a good estimate of the model, allowing us to make predictions about population for times that were not actually measured. Statistics. For each iteration: parameter estimates and residual sum of squares. For

each model: sum of squares for regression, residual, uncorrected total and corrected total, parameter estimates, asymptotic standard errors, and asymptotic correlation matrix of parameter estimates. Note: Constrained nonlinear regression uses the algorithms proposed and implemented in NPSOL by Gill, Murray, Saunders, and Wright to estimate the model parameters.
Data. The dependent and independent variables should be quantitative. Categorical variables, such as religion, major, or region of residence, need to be recoded to binary (dummy) variables or other types of contrast variables.

32 Chapter 5

Assumptions. Results are valid only if you have specified a function that accurately describes the relationship between dependent and independent variables. Additionally, the choice of good starting values is very important. Even if youve specified the correct functional form of the model, if you use poor starting values, your model may fail to converge or you may get a locally optimal solution rather than one that is globally optimal. Related procedures. Many models that appear nonlinear at first can be transformed to
a linear model, which can be analyzed using the Linear Regression procedure. If you are uncertain what the proper model should be, the Curve Estimation procedure can help to identify useful functional relations in your data.
Obtaining a Nonlinear Regression Analysis
E From the menus choose: Analyze Regression Nonlinear. Figure 5-1 Nonlinear Regression dialog box
E Select one numeric dependent variable from the list of variables in your working

data file.

33 Nonlinear Regression E To build a model expression, enter the expression in the Model field or paste
components (variables, parameters, functions) into the field.
E Identify parameters in your model by clicking Parameters.
A segmented model (one that takes different forms in different parts of its domain) must be specified by using conditional logic within the single model statement.
Conditional Logic (Nonlinear Regression)
You can specify a segmented model using conditional logic. To use conditional logic within a model expression or a loss function, you form the sum of a series of terms, one for each condition. Each term consists of a logical expression (in parentheses) multiplied by the expression that should result when that logical expression is true. For example, consider a segmented model that equals 0 for X<=0, X for 0<X<1, and 1 for X>=1. The expression for this is: (X<=0)*0 + (X>0 & X < 1)*X + (X>=1)*1. The logical expressions in parentheses all evaluate to 1 (true) or 0 (false). Therefore: If X<=0, the above reduces to 1*0 + 0*X + 0*1=0. If 0<X<1, it reduces to 0*0 + 1*X +0*1 = X. If X>=1, it reduces to 0*0 + 0*X + 1*1 = 1. More complicated examples can be easily built by substituting different logical expressions and outcome expressions. Remember that double inequalities, such as 0<X<1, must be written as compound expressions, such as (X>0 & X < 1). String variables can be used within logical expressions: (city=New York)*costliv + (city=Des Moines)*0.59*costliv This yields one expression (the value of the variable costliv) for New Yorkers and another (59% of that value) for Des Moines residents. String constants must be enclosed in quotation marks or apostrophes, as shown here.

38 Chapter 5

Nonlinear Regression Save New Variables
Figure 5-5 Nonlinear Regression Save New Variables dialog box
You can save a number of new variables to your active data file. Available options are Predicted values, Residuals, Derivatives, and Loss function values. These variables can be used in subsequent analyses to test the fit of the model or to identify problem cases.
Predicted Values. Saves predicted values with the variable name pred_. Residuals. Saves residuals with the variable name resid. Derivatives. One derivative is saved for each model parameter. Derivative names
are created by prefixing 'd.' to the first six characters of parameter names.
Loss Function Values. This option is available if you specify your own loss
function. The variable name loss_ is assigned to the values of the loss function.
Nonlinear Regression Options
Figure 5-6 Nonlinear Regression Options dialog box

39 Nonlinear Regression

Options allow you to control various aspects of your nonlinear regression analysis:
Bootstrap Estimates. A method of estimating the standard error of a statistic using repeated samples from the original data set. This is done by sampling (with replacement) to get many samples of the same size as the original data set. The nonlinear equation is estimated for each of these samples. The standard error of each parameter estimate is then calculated as the standard deviation of the bootstrapped estimates. Parameter values from the original data are used as starting values for each bootstrap sample. This requires the sequential quadratic programming algorithm. Estimation Method. Allows you to select an estimation method, if possible. (Certain choices in this or other dialog boxes require the sequential quadratic programming algorithm.) Available alternatives include Sequential quadratic programming and Levenberg-Marquardt. Sequential Quadratic Programming. This method is available for constrained and

unconstrained models. Sequential quadratic programming is used automatically if you specify a constrained model, a user-defined loss function, or bootstrapping. You can enter new values for Maximum iterations and Step limit, and you can change the selection in the drop-down lists for Optimality tolerance, Function precision, and Infinite step size.
Levenberg-Marquardt. This is the default algorithm for unconstrained models.
The Levenberg-Marquardt method is not available if you specify a constrained model, a user-defined loss function, or bootstrapping. You can enter new values for Maximum iterations, and you can change the selection in the drop-down lists for Sum-of-squares convergence and Parameter convergence.
Interpreting Nonlinear Regression Results
Nonlinear regression problems often present computational difficulties: The choice of initial values for the parameters influences convergence. Try to choose initial values that are reasonable and, if possible, close to the expected final solution. Sometimes one algorithm performs better than the other on a particular problem. In the Options dialog box, select the other algorithm if it is available. (If you specify a loss function or certain types of constraints, you cannot use the Levenberg-Marquardt algorithm.)

40 Chapter 5

When iteration stops only because the maximum number of iterations has occurred, the final model is probably not a good solution. Select Use starting values from previous analysis in the Parameters dialog box to continue the iteration or, better yet, choose different initial values. Models that require exponentiation of or by large data values can cause overflows or underflows (numbers too large or too small for the computer to represent). Sometimes you can avoid these by suitable choice of initial values or by imposing constraints on the parameters.
NLR Command Additional Features
The SPSS command language also allows you to: Name a file from which to read initial values for parameter estimates. Specify more than one model statement and loss function. This makes it easier to specify a segmented model. Supply your own derivatives rather than use those calculated by the program. Specify the number of bootstrap samples to generate. Specify additional iteration criteria, including setting a critical value for derivative checking and defining a convergence criterion for the correlation between the residuals and the derivatives. Additional criteria for the CNLR (constrained nonlinear regression) command allow you to: Specify the maximum number of minor iterations allowed within each major iteration. Set a critical value for derivative checking. Set a step limit. Specify a crash tolerance to determine if initial values are within their specified bounds. See the SPSS Command Syntax Reference for complete syntax information.

Standard linear regression models assume that variance is constant within the population under study. When this is not the casefor example, when cases that are high on some attribute show more variability than cases that are low on that attributelinear regression using ordinary least squares (OLS) no longer provides optimal model estimates. If the differences in variability can be predicted from another variable, the Weight Estimation procedure can compute the coefficients of a linear regression model using weighted least squares (WLS), such that the more precise observations (that is, those with less variability) are given greater weight in determining the regression coefficients. The Weight Estimation procedure tests a range of weight transformations and indicates which will give the best fit to the data.
Example. What are the effects of inflation and unemployment on changes in stock
prices? Because stocks with higher share values often show more variability than those with low share values, ordinary least squares will not produce optimal estimates. Weight estimation allows you to account for the effect of share price on the variability of price changes in calculating the linear model.
Statistics. Log-likelihood values for each power of the weight source variable tested, multiple R, R-squared, adjusted R-squared, ANOVA table for WLS model, unstandardized and standardized parameter estimates, and log-likelihood for the WLS model. Data. The dependent and independent variables should be quantitative. Categorical
variables, such as religion, major, or region of residence, need to be recoded to binary (dummy) variables or other types of contrast variables. The weight variable should be quantitative and should be related to the variability in the dependent variable.
Assumptions. For each value of the independent variable, the distribution of the
dependent variable must be normal. The relationship between the dependent variable and each independent variable should be linear, and all observations should be independent. The variance of the dependent variable can vary across levels of the

42 Chapter 6

independent variable(s), but the differences must be predictable based on the weight variable.
Related procedures. The Explore procedure can be used to screen your data. Explore
provides tests for normality and homogeneity of variance, as well as graphical displays. If your dependent variable seems to have equal variance across levels of independent variables, you can use the Linear Regression procedure. If your data appear to violate an assumption (such as normality), try transforming them. If your data are not related linearly and a transformation does not help, use an alternate model in the Curve Estimation procedure. If your dependent variable is dichotomousfor example, whether a particular sale is completed or whether an item is defectiveuse the Logistic Regression procedure. If your dependent variable is censoredfor example, survival time after surgeryuse Life Tables, Kaplan-Meier, or Cox Regression, available in the SPSS Advanced Models option. If your data are not independentfor example, if you observe the same person under several conditionsuse the Repeated Measures procedure, available in the SPSS Advanced Models option.

Obtaining a Weight Estimation Analysis
E From the menus choose: Analyze Regression Weight Estimation.
43 Weight Estimation Figure 6-1 Weight Estimation dialog box
E Select one dependent variable. E Select one or more independent variables. E Select the variable that is the source of heteroscedasticity as the weight variable.
Weight Variable. The data are weighted by the reciprocal of this variable raised to
a power. The regression equation is calculated for each of a specified range of power values and indicates the power that maximizes the log-likelihood function.
Power Range. This is used in conjunction with the weight variable to compute
weights. Several regression equations will be fit, one for each value in the power range. The values entered in the Power range test box and the through text box must be between -6.5 and 7.5, inclusive. The power values range from the low to high value, in increments determined by the value specified. The total number of values in the power range is limited to 150.

44 Chapter 6

Weight Estimation Options
Figure 6-2 Weight Estimation Options dialog box
You can specify options for your weight estimation analysis:
Save best weight as new variable. Adds the weight variable to the active file. This variable is called WGT_n, where n is a number chosen to give the variable a unique name. Display ANOVA and Estimates. Allows you to control how statistics are displayed in the output. Available alternatives are For best power and For each power value.
WLS Command Additional Features
The SPSS command language also allows you to: Provide a single value for the power. Specify a list of power values, or mix a range of values with a list of values for the power. See the SPSS Command Syntax Reference for complete syntax information.
Standard linear regression models assume that errors in the dependent variable are uncorrelated with the independent variable(s). When this is not the case (for example, when relationships between variables are bidirectional), linear regression using ordinary least squares (OLS) no longer provides optimal model estimates. Two-stage least-squares regression uses instrumental variables that are uncorrelated with the error terms to compute estimated values of the problematic predictor(s) (the first stage), and then uses those computed values to estimate a linear regression model of the dependent variable (the second stage). Since the computed values are based on variables that are uncorrelated with the errors, the results of the two-stage model are optimal.

asymptotic regression in Nonlinear Regression, 35
Cox and Snell R-square in Multinomial Logistic Regression, 18 custom models in Multinomial Logistic Regression, 15
backward elimination in Logistic Regression, 6 binary logistic regression, 1
categorical covariates, 7 cell probabilities tables in Multinomial Logistic Regression, 18 cells with zero observations in Multinomial Logistic Regression, 20 classification in Multinomial Logistic Regression, 13 classification tables in Multinomial Logistic Regression, 18 confidence intervals in Multinomial Logistic Regression, 18 constant term in Linear Regression, 10 constrained regression in Nonlinear Regression, 37 contrasts in Logistic Regression, 7 convergence criterion in Multinomial Logistic Regression, 20 Cooks D in Logistic Regression, 9 correlation matrix in Multinomial Logistic Regression, 18 covariance matrix in Multinomial Logistic Regression, 18 covariates in Logistic Regression, 7
delta as correction for cells with zero observations, 20 density model in Nonlinear Regression, 35 deviance function for estimating dispersion scaling value, 21 DfBeta in Logistic Regression, 9 dispersion scaling value in Multinomial Logistic Regression, 21
fiducial confidence intervals in Probit Analysis, 28 forward selection in Logistic Regression, 6 full factorial models in Multinomial Logistic Regression, 15
Gauss model in Nonlinear Regression, 35 Gompertz model in Nonlinear Regression, 35 goodness of fit in Multinomial Logistic Regression, 18

58 Index

Hosmer-Lemeshow goodness-of-fit statistic in Logistic Regression, 10
intercept include or exclude, 15 iteration history in Multinomial Logistic Regression, 20 iterations in Logistic Regression, 10 in Multinomial Logistic Regression, 20 in Probit Analysis, 28
residuals, 9 saving new variables, 9 set rule, 6 statistics, 3 statistics and plots, 10 string covariates, 7 variable selection methods, 6 log-likelihood in Multinomial Logistic Regression, 18 in Weight Estimation, 41 log-modified model in Nonlinear Regression, 35

Johnson-Schumacher model in Nonlinear Regression, 35
leverage values in Logistic Regression, 9 likelihood ratio for estimating dispersion scaling value, 21 goodness of fit, 18 Linear Regression Two-Stage Least-Squares Regression, 45 weight estimation, 41 Logistic Regression, 3 binary, 1 categorical covariates, 7 classification cutoff, 10 coefficients, 3 command additional features, 11 constant term, 10 contrasts, 7 define selection rule, 6 display options, 10 example, 3 Hosmer-Lemeshow goodness-of-fit statistic, 10 influence measures, 9 iterations, 10 predicted values, 9 probability for stepwise, 10
main-effects models in Multinomial Logistic Regression, 15 McFadden R-square in Multinomial Logistic Regression, 18 Metcherlich law of diminishing returns in Nonlinear Regression, 35 Michaelis Menten model in Nonlinear Regression, 35 Morgan-Mercer-Florin model in Nonlinear Regression, 35 Multinomial Logistic Regression, 13, 18 command additional features, 24 criteria, 20 exporting model information, 23 models, 15 reference category, 17 save, 23 statistics, 18
Nagelkerke R-square in Multinomial Logistic Regression, 18 nonlinear models in Nonlinear Regression, 35 Nonlinear Regression, 31 bootstrap estimates, 38 command additional features, 40 common nonlinear models, 35 conditional logic, 33

59 Index

derivatives, 38 estimation methods, 38 example, 31 interpretation of results, 39 Levenberg-Marquardt algorithm, 38 loss function, 36 parameter constraints, 37 parameters, 34 predicted values, 38 residuals, 38 save new variables, 38 segmented model, 33 sequential quadratic programming, 38 starting values, 34 statistics, 31
ratio of quadratics model in Nonlinear Regression, 35 reference category in Multinomial Logistic Regression, 17 relative median potency in Probit Analysis, 28 Richards model in Nonlinear Regression, 35
parallelism test in Probit Analysis, 28 parameter constraints in Nonlinear Regression, 37 parameter estimates in Multinomial Logistic Regression, 18 Peal-Reed model in Nonlinear Regression, 35 Pearson chi-square for estimating dispersion scaling value, 21 goodness of fit, 18 Probit Analysis command additional features, 29 criteria, 28 define range, 27 example, 25 fiducial confidence intervals, 28 iterations, 28 natural response rate, 28 parallelism test, 28 relative median potency, 28 statistics, 25, 28
separation in Multinomial Logistic Regression, 20 singularity in Multinomial Logistic Regression, 20 step-halving in Multinomial Logistic Regression, 20 stepwise selection in Logistic Regression, 6 in Multinomial Logistic Regression, 15 string covariates in Logistic Regression, 7
Two-Stage Least-Squares Regression, 45 command additional features, 47 covariance of parameters, 47 example, 45 instrumental variables, 45 saving new variables, 47 statistics, 45