Documentos de Académico
Documentos de Profesional
Documentos de Cultura
0 Reference Booklet
End-to-End Software
for Statistical Analysis
Table of contents
Specification summaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
SPSS 16.0 is a modular, tightly integrated, full-featured Statistical software help at your fingertips
product line. Along with related products from SPSS Inc., One thing that sets SPSS apart from its competitors is its
it provides capabilities for all stages of the analytical ease of use. To ensure you get the most from the software,
process—planning, data collection, data access, data SPSS provides you with many types of easy-to-access online
and output management, data analysis, reporting, and help for selected items, including:
deployment. SPSS’ intuitive user interface makes n Case studies that show you how to use selected statistics
it easy to use, yet the software includes all of the data and interpret results through realistic, hands-on examples
management, statistics, and reporting methods you n Tutorials with examples to guide you through the software
need to perform even the toughest analysis. and have you up and running quickly
n Context-sensitive help with links to tutorials for step-by-
Use SPSS to answer all your difficult questions step instructions when you need them
Analysts and researchers have used SPSS software for n Statistics Coach™ to help you choose which statistical
more than 39 years to uncover key facts, patterns, and procedure or graph to use
trends in a variety of areas, including: n Chart Advisor to help you select the best chart for deeper
n Survey and market research and direct marketing insight into your data
n Academia n Statistical glossary for clarifying unfamiliar statistical terms
n Administrative research, human resources, and resource n “What’s This” help to explain and define terms
planning
n Medical, scientific, clinical, and social science research
n Planning and forecasting
n Quality improvement
n Reporting and ad hoc decision making
“Super… I’m very stimulated and pleased with the innovative thinking
that has going into this major upgrade to the SPSS base product.”
– King Douglas
Senior analyst
American Airlines
2
The analytical process
Get everything you need—from one source collecting data, re-enter them when doing analysis, and
Important steps in the analytical process include preparing enter them again when creating a report. If you collect
data for analysis and, once analysis is complete, putting data in Dimensions or SPSS Data Entry, however, variables
results in a format other people can use. If you use established in the data collection stage are carried through
software that has limited capabilities, you have to piece to SPSS. So you can concentrate on analysis instead of data
together products from a variety of vendors to get your preparation and management.
job done. SPSS can give you everything you need for the
analytical process—and all of our products work together. Build a complete system for the analytical process
SPSS Inc. has products for each stage of the analytical
Using SPSS products exclusively streamlines the entire process:
analytical process. Because SPSS products work together, n Save time and money by planning your analytical project
you don’t have to repeat work that might be required if n Collect/enter data—even on the Web—efficiently
you use products from a variety of vendors. For example, n Reach the analysis stage faster with easy data access
with a multi-vendor solution you would have to enter n Prepare for analysis quickly with efficient
variable labels (such as gender or income level) when data management
n Analyze data using a variety of statistics and
procedures for more accurate models
n Clearly report your results to the people who
can use them
1 Planning n Share results with others—on the Web or in
SamplePower ®,
presentations or publications
Deployment 7 SPSS Complex Samples ™,
and SPSS Conjoint ™
Content management The following pages of this booklet describe highlights
and Web delivery with
for each stage of the analytical process and present
2 Data collection
SPSS Predictive Enterprise Services™
specification summaries for all SPSS products that you
Reporting 6 Dimensions™ and
SPSS Data Entry ™
can incorporate into your system. This booklet also
SPSS Base, SPSS Tables, includes a product/application chart that guides you
and Dimensions
to the appropriate products to fit your requirements.
3
Everything you need to prepare data for analysis
Save time and money with pre-analysis planning Efficiently collect data from surveys and forms
By taking the time to plan your analytical project before Would you like to collect and analyze survey data using a
you begin analysis, you can save time and money in the system that works seamlessly with SPSS? Dimensions and
long run. If you don’t already have data for your analysis, SPSS Data Entry products provide you with options for
use SamplePower to plan your data collection. fielding surveys.
When working with sample survey data, use the complex Create interactive surveys to engage survey takers. Here, respondents
can drag virtual dollars to one or more areas in which they’d like to
samples plan procedure (in SPSS Complex Samples). allocate funds.
This procedure enables you to specify how to draw or
analyze stratified, clustered, or multistage complex With Dimensions and SPSS Data Entry products, you can
sample designs, with or without replacement. Methods create professional surveys for almost any data type. Using
for sampling with probability proportional to size are their intuitive survey design interfaces, you can create new
also available. questions from scratch, start with tested sample questions
from the included Question Library, reuse questions from
The complex samples plan procedure includes two existing surveys, or choose selections from response lists.
wizards for planning through the interface—the Sampling
Plan Wizard and the Analysis Preparation Wizard. Which Dimensions provides options that enable you to conduct
one you use depends on your data source. When you surveys online, over the phone, on paper, through laptops
have created plan files, you can save them and treat them or hand-held devices, or by using a combination of these
as templates. This enables you to save all the decisions methods. If you offer surveys online, Secure Sockets Layer
you made when creating the plan. And it saves time and (SSL) encryption keeps your forms and surveys secure.
improves accuracy for yourself and others who may want Additionally, you can create surveys that include graphics,
to plug your plans into the data to replicate results or pick images, video, and sound files—to create vibrant and
up where you left off. Once you’ve created a sample or engaging surveys that people will want to complete. You
specified standard errors, you can confidently analyze your can even deploy surveys in many different languages,
data and produce results knowing that the planning stage including those, such as Asian languages, that use double-
addressed this specific data type. byte characters. Results obtained in multiple languages
are stored in a single, central source and are immediately
available for analysis. For desktop data entry of surveys or
forms, use SPSS Data Entry.
4
Both Dimensions and SPSS Data Entry create SPSS data Open multiple datasets within a single SPSS session
(SAV) files, so your data are always ready for analysis— In SPSS 16.0, you can open multiple datasets in a single
there’s no need for additional data cleaning or preparation. SPSS session. Each dataset has its own Data Editor
window. Use the interface or syntax to select the active
Reach the analysis stage faster dataset for a particular procedure. Apply commands across
Before you begin analysis, you often have to bring data multiple open datasets, as well as copy and paste data
from many types of data sources into SPSS. And you may dictionary information between multiple files.
have to transform the data to prepare them for analysis.
SPSS includes a number of wizards, tools, techniques, Prepare continuous-level data for analysis
and features that help you prepare and manage data prior If you have continuous-level data, such as income or age,
to analysis. Highlights are described below. you can easily prepare them for analysis with the Visual
Binning procedure. This procedure provides a quick,
The Database Wizard enables you to access massive interactive method by which you can become familiar with
amounts of data from numerous database sources from your dataset. A data pass provides you with a histogram
the interface. You can access databases without writing that enables you to specify cutpoints in an intelligent
code or syntax. The Database Wizard guides you through manner. For example, break income into “bins” of 10,000
the data access process and generates code in the or age into bins of 10 years. Then save time by automatically
background. This Wizard has drivers for many ODBC- creating value labels from the specified cutpoints (for
compliant databases, including Oracle , Microsoft SQL
® ® example, “21-30”).
Server™, Microsoft Access™, IBM DB2® UDB, and Sybase®.
With the right drivers, you can connect to any ODBC- Categorize one or more continuous variables by distributing
compliant database. You can also import data from OLE the values of each variable into bins. Optimal Binning
DB data sources without having to go through ODBC. (in SPSS Data Preparation) enables you to determine
Additionally, SPSS gives you easy access to SAS , Stata ® ® cutpoints to help you reach the best possible outcome
Microsoft Excel®, and text data. for algorithms designed for nominal attributes. Because
you can use a guide variable to determine cutpoints, you
Easily set up data using an efficient data dictionary can maximize the relationship between guide variables
Easily set up data dictionary information (value labels, and binned variables.
variable labels, and variable types) to organize your data
and prepare them for analysis more quickly using the Easily eliminate duplicate records
Define Variable Properties tool. A data pass made first Are your datasets likely to have duplicate records that
enables SPSS to present a list of values and counts so could compromise results if not removed? With the
you can add the information in a more intelligent manner. Identify Duplicate Cases tool, SPSS makes it easy for you
to prevent this. Set parameters, view a list of duplicates,
Improve data preparation for more accurate results flag records, and remove them prior to analysis. You can
Use specialized techniques in the SPSS Data Preparation* sort duplicate cases, if desired, to more easily select
add-on module to get your data ready for analysis faster the record you want to keep for analysis. You have the
and reach more accurate conclusions. The Validate Data flexibility to define duplicates using various methods. For
procedure enables you to apply rules to perform data example, you can “un-duplicate” based on multiple variables.
checks based on each variable’s measure level (whether
categorical or continuous). By using this procedure, you Work with long variable names
can eliminate tedious manual data validation checks. With SPSS, you can more accurately describe your data
The Anomaly Detection procedure searches for unusual using variable names up to 64 bytes in length. This
cases based upon deviations from similar cases and gives enables you to work with data from databases or spread-
reasons for such deviations. The information that results sheets that allow longer variable names. In addition, you
from these procedures enables you to determine data can ensure that data containing long text strings (up to
validity and remove or correct suspicious cases at your 32,767 bytes) is not truncated or lost when working with
discretion prior to analysis. open-ended question responses, databases, or data from
other software that allows long data strings, and other
types of long text strings.
* SPSS Data Preparation was previously called SPSS Data Validation™.
5
Incorporate dates and times into your analysis If you need to access results from SPSS jobs in other
In SPSS, you can easily work with dates and times using jobs or applications, use the Output Management
the Date and Time Wizard. Make calculations with dates System (OMS). The OMS provides you with the ability
and times, create date/time variables from strings to automatically write selected categories of output to
containing date variables (such as “3/29/06”), and bring different output files in different formats, including
date/time data from a variety of sources into SPSS. You SPSS (SAV) data files, XML, HTML, and TXT.
can also parse individual date/time units, such as year,
from date/time variables to apply filters. For example, With the OMS, for example, you can save pivot table
parse start dates to examine employees who started with output to SPSS-format data files, and then use that output
your organization in 2005. as input for subsequent commands that SPSS can process
further. The OMS provides you with a myriad of
Combine records without painstaking programming possibilities for working with SPSS output. For example:
Do you work with data that require lengthy programming n Automatically generate selected output as HTML for
before you can analyze them? The Data Restructure Wizard viewing in a browser
enables you to restructure a data file that has multiple n Process SPSS output in XML format using standard XSLT
records per subject so that data for each subject are in a tools to extract information for use in other applications
single record. No need to set up vectors or loops. You can n Generate a table of contents from XML using a supplied
also do the reverse—take data from a single record and XSLT stylesheet and free XSLT tools from third parties
spread them across multiple cases for linear mixed models
(see page 7). For more information on how to use the OMS, refer to
Programming and Data Management for SPSS 16.0:
Transform your data for easier analysis A Guide for SPSS and SAS users. This book covers several
Do you have surveys that ask respondents to give positively other topics, and includes a chapter explaining how you
worded and negatively worded responses? “How easy is can perform many common data management tasks in
it to reach customer service?” “How difficult is it to return both SAS and SPSS. See the “Resources” section on
items?” Multiple-item indices need all questions to go www.spss.com/spss to learn more about this book.
in the same direction. SPSS’ data transformations make
working with combined data more reliable by allowing you Use the SPSS Programmability Extension™ to create
to “flip” responses—so that all of your data are in the same procedures and applications—and perform even the most
direction. This enables you to turn negatives into positives, complex jobs—within SPSS. This powerful feature enables
creating overall, multiple-item indices. your organization to extend SPSS with external programming
languages, such as Python, R, and the .NET version of
SPSS gives you a variety of other transformation Microsoft® Visual Basic®. The SPSS Programmability
techniques that enable you to compute new variables Extension is included with SPSS 16.0—making SPSS an
using arithmetic, cross-case, logical, missing-value even more powerful statistical solution. See page 21 for
random number, statistical, or string functions. You more information.
can also use such programming structures as repeat-end
repeat, loop-end loop, vectors, and much more. Unlock the value in text responses
Do you include open-ended questions in your surveys?
You can also create your own dictionary information for How often have you bypassed analysis of open-ended
variables with Custom Attributes. For example, create custom responses because it takes too much time and money
attributes describing transformations for a derived variable to quantify this kind of data? What valuable information
with information explaining how you transformed the variable— remains hidden in these responses? SPSS Text Analysis for
which helps you keep track of transformations. Surveys supports the use of linguistic technologies and
manual techniques so you can create categories or “code
Expand your capabilities for programming frames” and categorize or “code” open-ended survey
With SPSS, you have some powerful ways to go beneath responses more quickly and reliably than manual-only
the SPSS interface to create applications and programs— approaches. Then export results as categories or
producing a limitless set of possibilities for using data dichotomies for further analysis in SPSS or Excel.
and output.
6
A broad range of statistics for data analysis
SPSS 16.0 helps you better analyze data because it gives SPSS Complex Samples provides you with complex samples
you the tools you need to solve a variety of business and descriptives and complex samples tabulate to use with
research problems—not just the problem for which you data resulting from complex sampling methods. Use
initially purchased the software. It provides you with a complex samples descriptives to analyze measures of
wide range of statistics so you can get the most accurate continuous data, including ratios, and complex sample
response for specific data types. Add-on modules give tabulate to analyze measures of categorical types, including
you even more analytical power and flexibility—and they crosstabulations. These procedures enable you to achieve
easily plug into SPSS Base. You can add as much analytical more statistically valid inferences for populations measured
capability to your system as you need and work confidently, in your complex sample data. Because they incorporate
moving seamlessly from one product to the next. sample design into survey analysis, you can more confidently
state results.
SPSS’ statistics include a wide variety of procedures
for descriptive analysis, numerical prediction, group Predict numerical outcomes
identification, and forecasting. Statistical highlights are Sometimes what you want to predict is numerical, such as
described on the following pages. income, customer lifetime value, standardized test scores,
or patient length of stay. You might want answers to such
Understand your data questions as “How well can an outcome be predicted by
Before you run your data through models or create reports a set of predictors?” or “What are the best predictors of
and summaries, you want to understand your data. You an outcome?” SPSS has a variety of statistics, from linear
might ask such questions as “What’s the distribution?” regression to non-linear techniques, which enable you to
or “What does the average person score?” SPSS Base gives answer these questions and more.
you a variety of ways to summarize your data and accurately
describe variables of interest, including data displays, Access a wide range of powerful models
such as frequency distributions, boxplots, and histograms; Use generalized linear models (GENLIN) and generalized
measures of central tendency; and estimates of variability. estimating equations (GEE) (in SPSS Advanced Models)
to address a wide range of statistical modeling problems
More confidently reach sample survey results for numerical outcomes. GENLIN covers not only widely
When you work with complex samples, such as those that used statistical models, such as linear regression for
use stratified, clustered, or multistage sampling, you need normally distributed responses, but also many useful
specialized statistical techniques to account for the sample statistical models via its very general model formulation.
design and its associated standard errors. GEE extends generalized linear models to accommodate
correlated longitudinal data and clustered data.
6
7
designs, and random coefficient models. You gain more value customers.” With the complex samples general
accurate predictive models because the linear mixed linear model (CSGLM) (in SPSS Complex Samples), you
models procedure takes the hierarchical structure of your can build linear regressions, models for analysis of
data into account. variance (ANOVA), and models for analysis of covariance
(ANCOVA) to accurately analyze and predict numerical
You can also use the linear mixed models procedure if outcomes from your complex sample design.
you’re working with repeated measures data, including
situations in which there are different numbers of repeated Estimate nonlinear equations
measurements or different intervals for different cases, Do you work with models that have nonlinear relationships?
or both. For example, a healthcare organization testing For example, do you want to predict coupon redemption
patients recovering from illness might not have uniformly as a function of time and number of coupons distributed?
collected data. The organization would ideally test Estimate nonlinear equations using one of two SPSS
individuals at regularly scheduled intervals (for example, procedures: nonlinear regression (NLR) for unconstrained
every 90 days). But people often miss appointments or are problems or constrained nonlinear regression (CNLR)
tested earlier or later than the optimal time. More standard (in SPSS Regression Models) for constrained and
procedures discard data that do not meet test conditions. unconstrained problems. NLR enables you to estimate
The linear mixed models procedure, however, uses all of models that have arbitrary relationships between
your data—for a more accurate picture of patient recovery. independent and dependent variables using iterative
estimation algorithms. With CNLR, you can use linear and
Handle related multiple dependent variables nonlinear constraints on any combination of parameters.
Do you need a flexible procedure that provides
simultaneous analysis with related multiple dependent Fit structural equation models
variables? The general linear models (GLM) multivariate How do you test and confirm observed or latent (hidden)
procedure (in SPSS Advanced Models) provides flexible variables in your data? Using Amos structural equation
design and contrast options to estimate means and modeling (SEM) software enables you to create more
variances and to test and predict means. For example, realistic models than if you used standard multivariate
a financial services company might want to know what methods or regression alone. Build attitudinal and
predicts account balances for various account types. behavioral models in Amos that more realistically reflect
You can also mix and match categorical and continuous complex relationships because you can use observed
predictors to build models. You’re not limited to one type or latent numeric variables to predict any other numeric
of predictor variable. Instead, you have options that give variable. For example, confirm that motivation is more
you a wealth of model-building possibilities. important than intelligence in predicting job performance.
Compare measures from the same people over time Amos’ interactive, visual approach to SEM makes it easy
Do you need to measure the same people over time to, for you to learn and use. You can create path diagrams
for example, measure how perceived quality of healthcare of hypothesized models using drawing tools, rather than
changes? Using the GLM repeated measures procedure writing equations or typing commands. You can also
(in SPSS Advanced Models), you can take the same extend the capabilities of Amos by using popular Microsoft
measurement a fixed number of times on individual programming languages, including Visual Studio® and C#.
subjects or cases, and mix and match categorical and Once you complete a model, you can assess its fit with
continuous-level predictors, including interactions. a mouseclick. Amos can even suggest how the model may
be improved—for example, by adding an arrow to connect
Make more accurate numerical predictions when two variables. Graphs and statistics help you find an
using complex samples optimum trade-off between model simplicity and goodness
Do you use complex sampling methods to survey your of fit. And, when you’re done, you can print presentation-
customers or citizens? For example, in a customer quality output or copy and paste it into other applications.
satisfaction survey, you might survey 1,000 customers, The latest version of Amos enables you to add latent class
of which 500 people are categorized as “big-ticket analysis (mixture modeling) to your statistical options. For
customers” and 500 people are identified as “smaller- more details, please visit www.spss.com/amos.
8
or low-performance) to market a moderate-performance
car to the right people. SPSS procedures for identifying
groups fall within two types: segmentation/dimension
reduction and classification.
12
Better processing throughout
your entire organization
Sometimes you need more processing power, functionality, and data security than your desktop computer
provides—especially if you work with large datasets. For more efficient and secure handling of your large
SPSS Server’s two-tier client/server architecture, in which the SPSS desktop or client machine is connected
to the server machine, shifts data handling, preparation, and analytical tasks to the server. This prevents
open transportation of data across the network to the desktop. To further reduce network traffic, network
administrators can block client-side viewing of data that has been loaded on the server side. For total data
security over the network, administrators can create a connection that uses Secure Sockets Layer (SSL)
Several features unique to the server version of SPSS further increase its functionality. For true predictive
analysis, you can load one or more models created by SPSS or Clementine® on sample datasets in
SPSS Server to score new data as they are collected. Use the powerful predictor selection and Naïve
Bayes algorithms to identify relevant predictors in datasets that potentially have hundreds of
predictors. Conduct lengthy data preparation and analysis tasks, allowing for unsupervised,
unconnected overnight processing of very large datasets with the SPSS Batch Facility (SPSSB).
And, with the SPSS Adapter for SPSS Predictive Enterprise Services, you can store and manage
a variety of assets, including Python script files, while enjoying increased performance during
SPSS Server works with a variety of popular server platforms, including Microsoft Windows
Server ® 2003, Sun™ Solaris™, IBM® AIX®, HP-UX™, and Red Hat® Linux®.
13
Report your results to people who can use them
Summarize, display, and explore results
Once your analysis is complete, you usually need to
summarize results so your audience can understand
them. SPSS includes many graph types that enable
you to provide the right visual to describe your results.
Graph types include:
n Categorical charts
n Quality control charts
n Scatterplots
n Density charts
n Diagnostic and exploratory plots
n Multiple use charts
These charts were created using the Graphics Production Language (GPL).
SPSS Base includes a presentation graphics system that The background image shows a bubble chart that consists of multiple
enables you to easily create the chart you need to describe elements, lines, and points. Population is mapped to bubble size. The
foreground image shows a chart in which each histogram is sub-divided
your results and select from a variety of editing features into groups and stacked accordingly. The gridlines are drawn in 3-D.
for easy customization. The presentation graphics system
is also easy to use when working in a production setting.
You can create a chart or graph once and then use your
Create custom tables in no time
specifications to create hundreds more just like it.
If you need to present and display results in a tabular
format, use SPSS Tables. With it, you can summarize your
Use the Chart Builder, a chart creation interface, to more
data for different audiences. The table preview builder
easily create commonly used charts. If you’re an advanced
interface updates in real time, so you always know what
user, the Graphics Production Language (GPL) gives you
your tables will look like.
an even broader range of chart and option possibilities.
14
Shown here are examples of three common table types—two-dimensional crosstabulation (top), multiple response set
(middle), and shared response categories, or comperimeter table (bottom)—that you may want to use when analyzing and
describing your data in SPSS Tables.
You can also use SPSS Tables as an analytical tool to Information consumers easily find what they need—on
better understand your data when you include inferential their own. Eliminate the frustrating “back-and-forth” of
statistics. Choose from test independence (Chi-square), traditional static reports. You don’t have to create new
compare means (t test), or compare proportions (z test). reports whenever managers want to see the analysis in
These statistics enable you to compare means or a different way. Recipients can pivot and explore reports
proportions for demographic groups, customer segments, online to find answers—rather than coming to IT. And your
time periods, or other categorical variables to identify organization can keep critical information secure, because
trends, changes, or major differences in your data. you decide who receives reports and how much they see.
15
Seamlessly take a project from
planning to deployment
1. Planning/survey design
Develop surveys and forms in Dimensions or SPSS
Data Entry. Create a survey with Dimensions to capture 1 3
results from online, telephone, in-person, or paper
survey responses. Use SPSS Data Entry to create paper-
based surveys or forms and for desktop data entry. Both Online or
products have an intuitive interface to make survey and desktop
form creation quick and easy—greatly reducing your
data preparation time prior to analysis.
2 4
2. Data collection/data entry
Centrally collect data from any source. You may want to
collect data online (from the Internet or your intranet), by
phone, through text messaging, using portable devices Central survey
(for example, at a tradeshow), by scanning responses to repository Content
Repository
paper surveys, or through manual data entry. You can
use all of these data collection methods with Dimensions
or manually enter data with SPSS Data Entry, and ensure
�
�
WWW or intranet
Telephone
5
� Text messaging
that results are centrally collected and ready for analysis � Portable devices
in SPSS. � Paper scanning
� Data entry WWW or intranet
6
16
Solve your business and research problems
using the right products
Need help deciding which products from SPSS Inc. you can use for specific applications? This diagram, which is organized by application
and product (in the order used in the analytical process), will guide you to the appropriate product selection.
P ro d u c ts
tion
nt
s ysis
ion
sts
me
ies
ry
ed
fica
l
ple plex
Ent
t
Sur xt Ana
is
t Te
er
ess
loy
gor
oin
ing
del vanc
ds
rks l
lys
es
ons
Pow
re
si
n
ase
Net Neur
dep
ata
ata
xac
onj
ate
Sam om
Val Miss
egr
atio
Tre Clas
ren
abl
vey
Sco
Ana
s
d
nsi
e
ple
SD
SD
s
SB
SA
SC
SC
SC
SR
SE
ST
ST
ST
par
wo
del
line
art
os
S
e
ue
es
S
Sam
Dim
SPS
SPS
SPS
SPS
SPS
SPS
SPS
SPS
SPS
SPS
SPS
SPS
SPS
SPS
SPS
Sm
Am
Pre
Mo
Mo
for
On
Survey research, market research, and direct marketing
Customer satisfaction surveys ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Product attribute testing ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Pricing and promotion analysis ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Market segmentation studies ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Demographic studies
and opinion polling ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Database and
direct marketing ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Academia
Teaching ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Administration ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Administrative research, human resources, and resource planning
Program effectiveness ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Employee attitude and
satisfaction surveys ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Applicant selection and testing ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Compensation and
employment analysis ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Medical, scientific, clinical, and social science research
Treatment outcome analysis ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Behavioral and biomedical
research ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Environmental impact studies ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Criminal justice studies ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Outcomes management ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Project management ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Research and
development engineering ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Planning and forecasting
Sales and marketing
forecasting and budgeting ✓ ✓ ✓ ✓ ✓
Resource requirements
and forecasting ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
New product forecasting ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Econometrics ✓ ✓ ✓ ✓ ✓
Enrollment and admissions
forecasting ✓ ✓ ✓ ✓ ✓ ✓
Financial account balance ✓ ✓ ✓ ✓ ✓
Quality improvement
Productivity and service quality ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Manufacturing and
productivity analysis ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Predictive maintenance
planning ✓ ✓ ✓ ✓ ✓ ✓
Utilization studies ✓ ✓ ✓ ✓ ✓ ✓
Performance measurement ✓
Reporting and ad hoc decision making
Fraud, waste, and
abuse detection ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Risk and credit
management ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Reports to or from
governmental agencies ✓ ✓ ✓ ✓ ✓ ✓ ✓
Industry trend analysis ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Program evaluation
and funding decisions ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Institutional research ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓
Property and tax assessment ✓ ✓ ✓ ✓ ✓
17
Specification summaries – Work with cumulative distributions produce more readable graph output in
Discover which powerful statistics and and random number generators for a production setting.
procedures are included in each SPSS product. discrete distribution functions n Categorical charts
SPSS Base is listed first. The other products are – Use cumulative distributions for – Bar: simple, cluster, and stacked
listed in the order in which they are used in non-central distributions – Line: simple, multiple, and drop-line
the analytical process. Products listed with – Use density/probability functions for – Area: simple and stacked
an asterisk require the use of SPSS Base 16.0 continuous and discrete distributions – Pie: simple and exploding
in order to operate. – Work with non-central – High-low and high-low-close
density/probability functions – Boxplot: simple and clustered
SPSS Base 16.0 – Select from two-tail probabilities: – Error bar: simple and clustered
Chi-square and F n Scatterplots
Access, manage, prepare, and analyze data,
and report results. – Use auxiliary function: logarithm of the – Simple, grouped, scatterplot matrix,
Data access, data preparation, and data complete Gamma function and 3-D
and output management – Fit lines: linear, quadratic or cubic
n Database Wizard Descriptive statistics regression, and Lowess smoother;
n “GET SAS” command (to import SAS data) n Crosstabulations confidence interval control for total or
n Text Wizard n Frequencies subgroups; and display spikes to line
n Excel interface n Descriptives – Bin points by color or marker size to
n Export to Database Wizard n Explore prevent overlap
n Density charts
n Data editor
n Multiple datasets open within a single – Dot charts: stacked dots show
SPSS session distribution; symmetric, stacked,
Correlation
Custom Attributes and linear
n n Bivariate**
number generator functions graphics system. Included features make – Intel® or AMD x86 processor running at
it easier to create the graph you want and 1GHz or higher
18 **Multithreaded algorithm, resulting in improved performance and scalability on multiprocessor or multicore machines.
–
RAM: 512MB RAM; 1GB recommended depends on test) with varied sample sizes; cumulative link function for samples draw
–
450MB of available hard-disk space power only or power with varied effect sizes by complex sampling methods
–
CD-ROM drive and Alphas n Complex samples logistic regression
–
Super VGA (800x600) or a higher- n Create pivot tables (CSLOGISTIC) to perform binary logistic
resolution monitor n Print and export tables; optionally, regression analysis, as well as multiple
– For connecting with an SPSS Server, a export data to Excel or other spreadsheet logistic regression (MLR) analysis, for samples
network adapter running the TCP/IP programs drawn by complex sampling methods
network protocol n Save graphs to a number of formats, including
n Software WMF, EMF, and BMP; easily export to SPSS Conjoint*
– Internet Explorer 6.0 programs such as Word or PowerPoint Discover consumer preferences in order
n Use statistical tests for means, proportions, to help improve products and price
For SPSS 16.0 for Mac® correlations, ANOVA, regression, logistic them effectively.
n Operating system regression, survival analysis, and n Orthogonal main effects fractional designs
– Apple® Mac OS X 10.4 (Tiger™) or equivalence tests n Plancards to produce printed cards for
– PowerPC or Intel processor n Operating system: Windows 2000, 98, or 95 squares analysis of preference
– RAM: 512MB RAM; 1GB recommended n Hardware: Pentium-compatible processor
– 800 MB of available hard-disk space n Memory: 16MB RAM Dimensions
– CD-ROM drive n Minimum free drive space: 10MB Support the entire survey research process,
– Super VGA (800x600) or a higher- n SVGA monitor from questionnaire design to multimodal
resolution monitor data collection and analysis to Web-based
n Software SPSS Complex Samples* reporting, using an integrated suite of products
– Safari™ 1.3.1 or higher, Mozilla® Incorporate complex sample designs into data for survey research.
Firefox® 1.5 or higher, or Netscape® 7.2 analysis for more accurate analysis of this n Create surveys in any language, for
– Sampling Plan Wizard for use when phone, text messages, portable devices,
SPSS 16.0 for Linux® collecting data and paper) and store surveys in a
n Operating system† – Analysis Preparation Wizard for use when centralized location for easy access
– Any Linux OS that meets the following working with public-use datasets n Make the most of your existing technology
n Hardware sums, and ratios and compute their standardized programming language and
– Processor: Intel or AMD x86 processor associated standard errors, design effects, method of data management
running at 1 GHz or higher confidence intervals, and hypothesis tests Enjoy unlimited creativity in designing
– RAM: 512MB RAM; 1GB recommended for samples drawn by complex sample survey projects, increasing the quality of
– 450 MB of available hard-disk space methods results and the satisfaction of internal and
– CD-ROM drive n Complex samples tabulate (CSTABULATE) external clients
– Super VGA (800x600) or a higher- to display one-way frequency tables or n Deliver useful results to more people faster
*This software requires SPSS Base 16.0 in order to run. There are no other system requirements for this product.
**Multithreaded algorithm, resulting in improved performance and scalability on multiprocessor or multicore machines. 19
SPSS Data Entry – Create reports describing invalid data term inclusion, and a semantic network
Build paper surveys and forms and more – Save variables that record rule violations (available in the English version only)
accurately enter data on your desktop. and use them to help clean data and n Customize extraction dictionaries included
n Drag-and-drop form design filter out bad cases in the product to improve performance and
n Toolbox with many response options n Use the Anomaly Detection procedure results
n Ability to define data file and SPSS to identify unusual cases based upon n Export data as dichotomies or categories
dictionary as you build forms deviations from their peer group, and to SPSS (SAV) or Excel (XLS)
n Copy and paste variable properties receive reasons for such deviations n Different native lanuguage versions are
n Drag variables to automatically – Specify variables to be used by available for analyzing English, Dutch,
create questions the procedure French, German, or Spanish survey text
n Question library of sample questions and – Specify categorical, continuous, and
responses ID variables (to identify cases), and list System requirements
n Flexible formatting capabilities variables that should be excluded n Operating system: Microsoft Windows Vista
n Powerful data cleaning rules from analysis Business or Home Basic (32- and 64-bit) or
– Validation rules – Specify the methods of handling missing Windows XP Professional, Service Pack 2
– Checking rules values (32-bit)
– Skip-and-fill rules – Specify settings such as the percentage n Hardware: Pentium-class processor; 3.0
n Open multiple surveys at once of cases considered as anomalies and GHz recommended
n Up to 4,000 characters in open-ended minimum and maximum number of n Memory: 256MB RAM minimum; 512MB
System requirements – Write the model to a specified filename minimum; more recommended for larger
System requirements vary depending on the – Control the display of the output results datasets
product(s) you choose. Refer to www.spss.com n Use Optimal Binning to categorize one or n Monitor: 1024 x 768 (SVGA) resolution
or your sales representative for more details. more continuous variables by distributing n CD-ROM drive for installation
the values of each variable into bins. This n Web browser: Internet Explorer 6.0 or later
SPSS Missing Value Analysis* procedure is helpful for reducing the number for online help
Create higher-value data and build better of values in the given binning input variables,
models when you estimate missing data. which can greatly improve the performance SPSS Regression Models*
n Analyze patterns of missing data of algorithms. Make better predictions with powerful
n Impute missing values – Select from three types of binning for regression procedures.
n Expectation maximization (EM) algorithm preprocessing data prior to model building: n Multinomial logistic regression
n Regression algorithm Unsupervised binning via the equal n Binary logistic regression
n Univariate: compute count, mean, standard frequency algorithm, supervised binning n Unconstrained nonlinear regression (NLR)
deviation, and standard error of mean for all via the MDLP (Minimal Description Length n Constrained nonlinear regression (CNLR)
cases, excluding those containing Principle) algorithm, or hybrid MDLP binning n Two-stage least squares
missing values for all variables – Specify criteria such as how to define the n Probit analysis
n Listwise: compute mean, covariance matrix, minimum and maximum cut point for each
and correlation matrix for all quantitative binning input variable SPSS Advanced Models*
variables for cases, excluding missing values – Save the following new variables Analyze complex relationships using
n Pairwise: compute frequency, mean, containing binned values and syntax a high-end modeler’s toolkit.
variance, covariance matrix, and to an SPSS syntax file n Generalized linear models (GENLIN) and
correlation matrix – Control output results display with the generalized estimating equations (GEE)
PRINT subcommand – GENLIN represents a unifying framework
SPSS Data Preparation* that includes classical linear models with
Improve the data preparation process by SPSS Text Analysis for Surveys normally distributed dependent variable,
using specialized techniques designed for Reliably categorize your open-ended survey logistic, and probit models for binary
your data—and reach more accurate results. responses by using automatic linguistic data, and loglinear models for count
n Use the Validate Data procedure to validate processing technologies and manual techniques. data, as well as various other nonstandard
data in the working data file n Import data from ODBC-compliant databases, regression-type models
– Specify basic checks to apply to SPSS (SAV), Excel (XLS), and the – GEE extends the generalized linear model
variables and cases in your file Dimensions Data Model to correlated longitudinal data and
– Use standard rules to describe data, view n Extract key concepts—terms, types, and clustered data. More particularly,
single variable rules, and apply them to patterns—automatically using linguistic generalized estimating equations model
analysis variables resources correlations within subjects.
– Define cross-variable rule expressions in n Automate the creation of categories and – GENLIN and GEE provide a common
which respondents’ answers violate logic categorize responses using term derivation, framework for the following outcomes:
20 *This software requires SPSS Base 16.0 in order to run. There are no other system requirements for this product.
Discover unlimited
programming capabilities
Have you ever wanted to apply programming logic to an SPSS job? Program a job so that it responds in a certain
way to produce results? Or even create an application that calls upon the SPSS engine for analytical tasks but
does not show the SPSS user interface? The SPSS Programmability Extension (introduced on page 6) provides
you with the ability to do all of this and more through the use of general purpose programming languages.
The SPSS Programmability Extension enables your organization to extend the SPSS command syntax language
with external programming languages, such as Python*, R**, and the .NET version of Microsoft Visual Basic.
You can use scripts written in general purpose programming languages to conditionally execute syntax or make
decisions about which syntax is executed based on a particular condition, such as:
Additionally, you can implement custom algorithms or procedures—including those for statistical analyses not
included in SPSS—written in Python, R, or other programming languages. With this capability, you can also:
n Create a user interface (UI) in SPSS for any custom algorithm or procedure you implement
n Send results to an SPSS pivot table—essentially extending the analytical capabilities of SPSS
The SPSS Programmability Extension is included with SPSS Base 16.0. It provides the ability to integrate SPSS Base
with an SPSS Programmability Plug-In, which is necessary to take advantage of this advanced programmability
functionality. Use the freeware plug-ins that SPSS Inc. has already built for Python, R, and .NET, or follow the instructions
in the SPSS Programmability Extension Software Developer’s Kit (SDK) to build your own. You can download freeware
SPSS Developer Central is the online resource for end users and software developers interested in SPSS-related
programming and development. Here, SPSS Inc. regularly publishes freeware projects, example libraries and syntax
*SPSS Inc. is not the owner or licensor of the Python software. All Python users must agree to the terms of the Python license located on the
Python Web site (www.python.org). SPSS does not make any statement about the quality of the Python program. SPSS fully disclaims all liability
associated with your use of the Python program.
** SPSS Inc. is not the owner or licensor of R. All R users must agree to the terms of the license agreement located on the R project Web site. SPSS
does not make any statement about the quality of R. SPSS full disclaims all liability associated with your use of R. For more information on R,
visit www.r-project.org/.
21
SPSS Advanced Models* (continued) – Use automated specification for quick the hidden and output layers
n Continuous outcomes: Linear regression, specification and testing n RBF users can also specify the Gaussian
analysis of variance, analysis of n Analyze multiple models simultaneously radial basis function used in the hidden
covariance, repeated measures n Impute missing values or latent factor layer (either Normalized RBF or Ordinary
analysis and Gamma regression scores. Choose from three data imputation RBF)
n Ordinal data: Ordinal regression methods: Regression, stochastic regression, n The CRITERIA subcommand specifies the
n Count data: Loglinear models, logistic or Bayesian. computational and resource settings to be
regression, probit regression, Poisson n Obtain Bayesian estimates. Markov chain used by the chosen procedure
regression, and negative binomial Monte Carlo (MCMC), the underlying
regression estimation method, specifies the Bayesian SPSS Categories*
n Event/trial data: Logistic regression technique that Amos will use. (Developed in partnership with the Data
n Claim data: Inverse Gaussian regression n Estimate and impute numerical values for Theory Group at Leiden University)
n Combination of discrete and continuous ordered-categorical and censored data Understand groupings using perceptual maps
outcomes: Tweedie regression n Determine probable values for missing or and predict categorical outcomes.
n Linear mixed models partially missing data values in a latent n Multidimensional scaling of proximity data
– Fixed effect ANOVA, analysis of n Use latent class analysis (mixture modeling) n Correspondence analysis
analysis of variance (MANOVA), and – Perform market segmentation studies optimal scaling
multivariate analysis of covariance – Estimate the size of each cluster or n Multiple correspondence analysis
– Random or mixed ANOVA and ANCOVA – Perform mixture regression modeling more sets of variables via alternating
– Repeated measures ANOVA and MANOVA and mixture factor analysis least squares
n Variance component estimation (VARCOMP) – Estimate the probability of group n Preference scaling via multidimensional
– Specify each individual candidate model covariates or scale dependent variables graphs, and data tables
as a set of equality constraints on n The PARTITION subcommand specifies the n Easily build trees using the intuitive
from a large number of candidate models n The ARCHITECTURE subcommand is used to n Collapse and expand branches, and change
n Perform confirmatory factor analysis: specify the neural network architecture other cosmetic properties, such as fonts
variance components, errors in variables, n MLP users can specify whether to use the and colors
measurement models, and latent variables automatic architecture or, if automatic is not n View and print trees
n Analyze mean structures and multiple- chosen, the number of hidden layers, and n Specify the exact zoom percentage for
group datasets the activation function to use for all units in viewing tree models in the interface
22 *This software requires SPSS Base 16.0 in order to run. There are no other system requirements for this product.
n Automate tree building using the factors for periodic time series – Minimum free disk space: 5GB or more
production mode n Decompose a times series into its harmonic – A CD-ROM drive is required for installation
n Force one predictor into the model components, a set of regular periodic n Software:
n Specify prior probabilities, misclassification functions at different wavelengths – Database: Microsoft SQL Server 2000 or
costs, revenues, expenses, and scale scores or periods 2005; Oracle® 9i or 10g, IBM® DB2® 8.1,
n Perform analysis using one of four powerful 8.2, or 9
tree-growing algorithms SPSS Tables* – Application Server: JBoss® 4.0.3; BEA
– CHAID by Kass (1980) Easily, quickly, and clearly report results WebLogic® 9.1; IBM WebSphere® 6.1,
– Exhaustive CHAID by Biggs, de Ville, in a tabular format to the people who can Oracle Application Server 10gR3
and Suen (1991) use them. – If using with SPSS, version 14.0 or later is
– Classification & regression trees (C&RT) n Drag-and-drop table builder interface to required
by Breiman, Friedman, Olshen, and preview tables as you select variables
Stone (1984) and options SPSS Predictive Enterprise Manager
– QUEST by Loh and Shih (1997) n Inferential statistics that you can include (Administrative client)
n Generate risk and classification tables with tables to highlight opportunities or The minimum hardware and software
n Summarize node performance with problem areas requirements are:
evaluation graphs and tables to help – Chi-square tests n Operating system: Windows Vista,
identify the best segments – Column proportion tests Windows XP, Windows Server 2003,
n Partition data between training and test – Pairwise comparison of means Windows 2000 Server or Windows 2000
data to verify accuracy n A variety of summary statistics for even n Hardware:
n Display summary graphs or classification deeper insight into data – Processor: Pentium processor 1.8 GHz
rules for selected nodes using the node n Control contents to gain more control or faster
summary window over output – Memory: 512 MB RAM or more
n Export n Formatting controls for more – Minimum free drive space: 100 MB or
– Tree diagrams and charts. Export formats customized tables more
include Windows metafile (WMF), n Easy-to-understand syntax to create reports – A network adapter running TCP/IP protocol
Windows bitmap (BMP), encapsulated in a production setting – A CD-ROM drive is required for installation
PostScript® (EPS), JPEG, TIFF, PNG, and n Software:
Macintosh® PICT SPSS Predictive Enterprise Services – If using with SPSS, version 14.0 or later
– Summary tables as HTML, text, Microsoft Put valuable reports into the right hands, is required
Word/RTF, and Microsoft Excel files quickly and securely.
n Save information from the model as n Publish results instantly with Web-based SmartScore
variables in the working data file analytical reports Develop applications that generate scores
n Export decision rules: n Easily share information with customers, (predictions) for individual cases and deliver
– That define selected segments in SQL partners, and the public accurate profiles using this software
to score databases n Quickly distribute SPSS analysis and development kit (SDK).
– As SPSS syntax to score SPSS files other reports n Deploy scoring models enterprise wide
– As simple text (through syntax) n Enable applications to publish and for more efficient and effective
n Export trees as XML models for use with update content in the repository for decision support
SPSS Server and SmartScore to score new greater flexibility n Use Java-class files, which offer an
cases or data files n Keep critical information secure even if you application programming interface (API)
n Publish trees as images and tables as have large security domains n Score various models created in SPSS,
static or interactive tables to SmartViewer n Manage security with a centralized system AnswerTree®, and Clementine
Web Server n Score your own models created using
n Select interesting segments in the working SPSS Predictive Enterprise Repository (Server) an XML-based Predictive Model Markup
data file via tree nodes, and run more The minimum hardware and software require- Language (PMML)
analyses for additional insight ments are: n Develop applications using C++, C, Java™,