Technical Support
Frequently Asked Questions about SPSS
Q. I have some SPSS for the Macintosh 6.1 portable files
that I would like to transfer to the VAX. The procedure I used to use in
SPSS for the Macintosh version 4 doesn't seem to be working. How can I do
this with version 6.1 of SPSS for the Macintosh?
A.The format
of the SPSS portable files has changed since version 4 to for allow 8-bit
characters (extended ASCII) and also to change the Macintosh file type
from TEXT to SPOR (SPSS PORtable file) so that Macintosh text editors and
word processors do not mistake portable files for regular TEXT files.
Because a portable file is no longer TEXT, Fetch transfers it differently
and alters the line endings.
There are two solutions:
1) Change the type of the portable files to TEXT, and then use Fetch like you could in version 4 of SPSS for the Macintosh. We have attached an AppleScript file to the end of this message that will do this for you: Just drop one or more portable files onto the script icon, and it will change them into text files (you'll see the icon change). This script requires the scriptable Finder in Macintosh System 7.5. or
2) If you have NCSA Telnet, you can start Telnet and use your Macintosh as an FTP server. Then *from* the VAX, ftp to your Macintosh and get the file (or mget the files) you need.
We have successfully used both methods to transfer portable files from a Macintosh machine to our VAX and read them with version 4 of SPSS for VMS.
Download the AppleScript file: SPSS Portable to Text
Q. I have a file in SPSS on my IBM PC which I
would like to transfer to SPSS for the Macintosh (or vice versa). How do I
do this?
A. The transfer procedure differs depending upon
which platforms you are using and which versions of SPSS you have.
cd /pub/spss/general/utilSet ftp to BINARY mode by issuing the command "bin" and copy PORTFIX1.EXE with the command "get PORTFIX1.EXE".
Q. Is SPSS affected by the Pentium II processor
bug?
A. We do not expect the floating point unit bug in
Intel's Pentium II processor to affect calculations in SPSS. We have
reviewed the characteristics of the Pentium II problem and the techniques
used in the underlying code in SPSS and are confident that the software
carefully checks for conditions related to the floating point unit.
Q. SPSS software and the year
2000
A. SPSS software is year 2000 enabled (with the
exception of the X11ARIMA procedure in the SPSS Trends module). SPSS
correctly handles dates beyond the year 1999. If a user enters a date with
only two digits for the year, by default SPSS assumes that the year is
19xx and stores that information. If the user wants to enter a date before
1900 or after 1999, then all four digits for the year must be entered.
Beginning with SPSS 8.0, the user can specify how two digits years will be
interpreted. For example, the user can specify that values greater than or
equal to 05 should be interpreted as 19xx and values less than or equal to
04 should be interpreted as 20xx. All computations and functions involving
dates work correctly under all date formats.
For more information, visit the Y2K section of this web site.
Q. Restructuring the case base of an SPSS
file
I currently have many records for each unique id in my
data set. I would like to have a single case per id without losing
information. That is, I would like to restructure my data such that a
variable that currently has multiple values across records but within
cases, is restructured into many variables with multiple values within a
single case and on a single record.
I currently have many records for each unique id in my
data set. I would like to have a single case per id without losing
information. That is, I would like to restructure my data such that a
variable that currently has multiple values across records but within
cases, is restructured into many variables with multiple values within a
single case and on a single record.
A. The following is a worked
example restructuring one variable. Included in this syntax is output from
LIST commands. This output shows you how the sample file is being
restructured at each step. If you run the sample syntax, don't include the
output portions that are included here.
* To make one from many *.
DATA LIST FREE / ID Y.
BEGIN DATA
1 3
1 4
1 5
1 6
1 3
2 5
2 5
2 7
2 1
END DATA .
COMPUTE REC=1.
IF ID EQ LAG(ID) REC=LAG(REC)+1 .
LIST.
ID Y REC
1.00 3.00 1.00
1.00 4.00 2.00
1.00 5.00 3.00
1.00 6.00 4.00
1.00 3.00 5.00
2.00 5.00 1.00
2.00 5.00 2.00
2.00 7.00 3.00
2.00 1.00 4.00
VECTOR Y1_(5).
COMPUTE Y1_(REC)=Y.
LIST.
ID Y REC Y1_1 Y1_2 Y1_3 Y1_4 Y1_5
1.00 3.00 1.00 3.00 . . . .
1.00 4.00 2.00 . 4.00 . . .
1.00 5.00 3.00 . . 5.00 . .
1.00 6.00 4.00 . . . 6.00 .
1.00 3.00 5.00 . . . . 3.00
2.00 5.00 1.00 5.00 . . . .
2.00 5.00 2.00 . 5.00 . . .
2.00 7.00 3.00 . . 7.00 . .
2.00 1.00 4.00 . . . 1.00 .
AGGREGATE OUTFILE *
/ BREAK ID
/ Y1_1 TO Y1_5 = MAX( Y1_1 TO Y1_5).
LIST.
ID Y1_1 Y1_2 Y1_3 Y1_4 Y1_5
1.00 3.00 4.00 5.00 6.00 3.00
2.00 5.00 5.00 7.00 1.00 .
Q. How can I troubleshoot or diagnose an ODBC problem?
If the query with the first data source and ODBC driver works in Excel, you will need to contact SPSS Technical Support. In order for Technical Support to help you efficiently, you will need to supply the following information.
For items 4, 5, 6, SPSS Technical Support can e-mail or fax instructions for obtaining the information.
Q. How can I find the version number of my ODBC driver in Windows95?
If you were using Windows 3.1 or Windows NT, you would find similar options available, though the ODBC Administrator control panel might appear different.
Q. How do I generate an ODBC or SQL trace log?
Choose the first if you just want to generate a trace log for one query or one SPSS session. Choose the second if you want to generate a trace log for several SPSS sessions. If you want to compare the trace logs for two queries, one which is successful and one which is not, for example, you might find it useful to choose "One-time only," run SPSS and execute the query, exit SPSS, rename the SQL or ODBC trace log, return to the "32bit ODBC" control panel, select the "Tracing" tab and choose "One-time only," click on Apply and OK, restart SPSS and execute the second query.
The "Log file path" text entry box will tell you the current path and filename of the trace log; you can use this text entry box to change the path and filename. Click on Apply and then OK for your selections to take effect.
If you were using Windows 3.1 or Windows NT, you would find similar options available, though the ODBC Administrator control panel might appear different.
Q. How do I get a copy of the GET CAPTURE ODBC command syntax?
SPSS 7.5 for Windows lets you copy SPSS text, charts and pivot tables
into other applications. However, if you used SPSS 7.0 for Windows, you'll
notice a change in these capabilities. This document describes the
changes, what's available in SPSS 7.5 for Windows and how you can use
these features.
Please note: For the purposes of this document, "drag and drop" and "copy and paste" are interchangeable. That is, whether you use the menu to select copy and paste, or use drag and drop to move output, the end results are the same and the same rules apply.
What's new with moving SPSS output? Why?
With SPSS 7.5 for
Windows, SPSS implemented a new technology called Active-X. Using this
advanced technology, SPSS 7.5 now allows you to drag and drop pivot
tables, drag and drop multiple output objects at the same time, and
interactively edit them in your destination application. These features
require that the destination application is also Active-X enabled.
Microsoft Office '97 products will support the Active-X technology, but previous versions of Microsoft products, as well as some other word processing and presentation programs do not currently support Active-X technology.
Drag and drop without Active-X applications
When you use
SPSS 7.5 for Windows with non-Active-X applications, you can still drag
and drop SPSS text, pivot tables and charts into other documents. However,
you can not edit the SPSS output within the destination application.
To copy SPSS tables into other documents as text, use copy and paste or drag and drop. To keep the table formatting intact, choose Edit, Paste Special and select Picture.
When you install SPSS 7.5 for Windows, it assumes that your other applications do not support Active-X technology. When you begin to use new products that support Active-X objects, you must turn-on the Active-X support in SPSS.
Drag and drop with new Active-X applications (ex: Microsoft Office
'97)
To turn on the Active-X support in SPSS 7.5 for Windows, run
the file OBJS-ON.BAT in your SPSS7.5 directory. You can then edit SPSS
pivot tables within your destination application and drag and drop
multiple SPSS output objects at once.
Please note: If you turn on Active-X support and your applications do not support Active-X objects, you will not be able to drag and drop. To turn off Active-X support, run the file OBJS-OFF.BAT.
SPSS is dedicated to using leading technology to provide you with products that meet your needs. We apologize for any inconvenience these drag and drop changes may have caused, and thank you for your continued support of SPSS products and services.
Q. Moving from SPSS LISREL 7 to Amos 3.6 for
Windows
A. Transition Guidelines
Setting up analyses and running models differ between SPSS LISREL 7 and Amos 3.6 for Windows. This document provides information to facilitate the transition between the two products. It gives step by step instructions on how to get SPSS LISREL 7 data into Amos for Windows and includes sample SPSS syntax.
Instructions in this document assume that SPSS Base for Windows, release 6.1, 7.0, or 7.5, and Amos 3.6 for Windows are already installed.
Special considerations for various releases of SPSS for Windows
Amos can be invoked either standalone or from the SPSS Statistics menu . Here are some special considerations for invoking Amos from the SPSS menu.
Patches for SPSS Base 7.0 for Windows
If you are running SPSS Base 7.0 for Windows, it may be necessary to install the following patches to your SPSS Base system. To decide if you need the patches, install Amos. Open a file in the SPSS Data Editor. If Amos does not appear as a menu item on the SPSS statistics menu, then you must install the patches. The patches can be downloaded from the Internet at:
You must apply the second patch after the first patch.
Starting Amos from the SPSS menu
To launch Amos from SPSS for Windows a file must be active in the Data Editor. Amos appears as a menu choice under SPSS' Statistics menu in the following windows:
| Release of SPSS | Statistics menu on |
| SPSS 7.0 for Windows | Data window |
| SPSS 6.1.4 for Windows | Data window
Output window Syntax window |
| SPSS 7.5 for Windows | Data Editor window
Output Navigator window Syntax window |
Creating a path diagram
Usually prior to specifying and estimating a model in SPSS LISREL 7, you create a path diagram to aid in specifying the parameters in the model. You must reproduce this path diagram in Amos for Windows using the drawing tools. Amos' graphical user interface is highly intuitive and removes two steps from the modeling process:
Getting data from SPSS LISREL 7 to Amos 3.6 for Windows
There are several ways to bring data from SPSS LISREL to Amos. This section gives step by step instructions for three different scenarios. Each example assumes that you have an active file in the SPSS Data Editor, and Amos is accessed from the SPSS Statistics menu.
Syntax to create the sample data file which is used by the syntax file template
INPUT PROGRAM.
LOOP I=1 TO 1000.
COMPUTE SEED=NORMAL(1).
COMPUTE KSI1=.8*Seed + Normal(1).
COMPUTE KSI2=.8*Seed + Normal(1).
COMPUTE Eta1=.6 * Ksi1 + .8 * Ksi2 + NORMAL(.5).
Compute X1=.7 * Ksi1 + Normal(1).
Compute X2=.7 * Ksi1 + Normal(1).
Compute X3=.7 * Ksi1 + Normal(1).
Compute X4=.7 * Ksi2 + Normal(1).
Compute X5=.7 * Ksi2 + Normal(1).
Compute X6=.7 * Ksi2 + Normal(1).
Compute Y1=.7 * Eta1+ Normal(1).
Compute Y2=.7 * Eta1+ Normal(1).
Compute Y3=.7 * Eta1+ Normal(1).
END CASE.
END LOOP.
END FILE.
END INPUT PROGRAM.
CORR VAR=X1 TO Y3
/MATRIX=OUT(*).
SAVE OUTFILE 'c:\SPSSDATA.SAV'.
* SPSS syntax file template
* The example which follows uses 9 variables:
* X1 X2 X3 X4 X5 X6 Y1 Y2 Y3.
* Convert the correlation matrix to a covariance matrix *.
mconvert.
* Create a counter within each group within each type
* of matrix content (MEAN SD COV N).
split file by rowtype_.
compute @=1.
create @=CSUM(@).
* Destroy upper triangle of Covariance matrix *.
* Use your variable names in place of X1 X2-- Y3
* and change 9 to the number of variables in your file.
DO REPEAT V=x1 x2 x3 x4 x5 x6 y1 y2 y3 / X=1 to 9.
IF (ROWTYPE_='COV' and X > @) V=$SYSMIS .
END REPEAT.
* Replace names and number of variables as before
(9F7.3 changes to ##F7.3)
* in the remaining jobstream* .
DO IF ROWTYPE_='MEAN'.
WRITE OUTFILE 'AMOSMAT0.DAT'
/ '$INPUT VARIABLES '
/ ' x1' / ' x2' / ' x3' / ' x4' / ' x5'/' x6'/' y1'/' y2'/' y3'
/ '$MEANS '
/'(9F7.3)'
/x1 x2 x3 x4 x5 x6 y1 y2 y3 (9F7.3) .
END IF.
DO IF ROWTYPE_='N' AND @=1 .
WRITE OUTFILE 'AMOSMAT0.DAT' / '$SAMPLE SIZE =' , x1 (F4.0) .
END IF .
DO IF ROWTYPE_ = 'COV' AND @=1.
WRITE OUTFILE 'AMOSMAT0.DAT'
/'$COVARIANCES' / '(9F7.3)'.
END IF.
DO IF ROWTYPE_='COV'.
WRITE OUTFILE 'AMOSMAT0.DAT' /x1 x2 x3 x4 x5 x6 y1 y2 y3 (9F7.3) .
END IF.
EXECUTE.
* Results of the SPSS job *.
$INPUT VARIABLES
x1
x2
x3
x4
x5
x6
y1
y2
y3
$MEANS
(9F7.3)
-.018 .113 .028 .001 .066 .012 .066 .068 .103
$SAMPLE SIZE = 526
$COVARIANCES
(9F7.3)
1.954
.853 1.783
.877 .840 1.813
.274 .355 .247 1.839
.300 .365 .344 .781 1.686
.209 .328 .257 .755 .730 1.720
.667 .777 .834 .836 .744 .836 2.158
.657 .875 .761 .789 .713 .727 1.139 2.055
.765 .808 .821 .741 .836 .717 1.120 1.186 2.208
* Here is a MACRO to do the same thing *.
DEFINE MATAMOS
(VARS !ENCLOSE('(',')') /NVARS !TOKENS(1) /GPVAR !TOKENS(1) ).
!LET !LIST=' /' !DO !I !IN (!VARS)
!LET !LIST=!CONCAT(!LIST,'" ',!I,'"/') !DOEND
!LET !FMT=!CONCAT(!NVARS,'F7.3')
mconvert.
split file by rowtype_ .
compute @=1.
create @=CSUM(@).
DO REPEAT V=!VARS / @VARNUM@ =1 to !NVARS.
IF (ROWTYPE_='COV' and @VARNUM@ > @) V=$SYSMIS .
END REPEAT.
DO IF ROWTYPE_='MEAN'.
WRITE OUTFILE 'MATOUT.DAT'
/ '$INPUT VARIABLES '
!LIST '$MEANS '
/ !QUOTE(!CONCAT('(',!FMT,')'))/ !VARS (!FMT).
END IF.
DO IF ROWTYPE_='N' AND @=1 .
WRITE OUTFILE 'MATOUT.DAT'
/ '$SAMPLE SIZE =' , !HEAD(!VARS) (F4.0) .
END IF .
DO IF ROWTYPE_ = 'COV' AND @=1.
WRITE OUTFILE 'MATOUT.DAT'
/'$COVARIANCES'
/ !QUOTE(!CONCAT('(',!FMT,')')).
END IF.
DO IF ROWTYPE_='COV'.
WRITE OUTFILE 'MATOUT.DAT' /!VARS (!FMT) .
END IF.
EXECUTE.
!ENDDEFINE .
SET MPRINT ON.
MATAMOS VARS (X1 X2 X3 X4 X5 X6 Y1 Y2 Y3)
NVARS 9 .
Q. Insufficient Memory for
Installation
A.The installation procedure reports that you
have insufficient memory to install SPSS 7.0 for Windows although you have
at least 8MB of memory, the minimum amount of physical memory (RAM)
required to run SPSS.
There is a switch you may add to the RUN line
that will allow the installation procedure to continue. In the RUN dialog,
enter the line:
a:setup -NOENVFAIL
Make sure that you specify the correct drive letter (A or B), that there is a space after 'setup', and that the actual switch, "-NOENVFAIL, is in all capital letters.
Q. Installation Does Not Finish
A.
If the installation of SPSS 7.0 is halted or does not reach 100% before
displaying the Finish Dialog Box, it is often attributable to a
conflicting TSR or other application running in the background.
Instead of uninstalling SPSS, just re-run Setup to replace the renamed files. However, before attempting to re-install SPSS, try simplifying your Windows environment to prevent TSRs, device drivers, and other applications from interfering with the proper decompressing and copying of SPSS files.
After you have created a simplified environment, re-run Setup from your installation diskette and choosing the same options you had chosen previously. SPSS will determine which files it needs to re-install and will instruct you for the appropriate disks to complete its installation. You do NOT have to re-install SPSS completely.
This is a common question for new users of SPSS 7.0 for Windows. One of the major innovations of this release is that output from the base procedures and the new GLM procedure are displayed in what are known as Pivot Tables. Quite technically a Pivot Table such as a long frequency table is an OLE 2 object. This allows SPSS the freedom to display and manipulate the item. When a Pivot Table is displayed in an OLE container such as the SPSS Output Navigator, it is displayed as a Windows Metafile. However Windows Metafiles are limited in size. In order to view the entire table you need to activate the pivot table. In general a large table is easier to view, pivot, and edit if activated into its own window. To activate a table into its own window, right-click on the table in the output navigator and Open from the SPSS Pivot Table Object sub-menu. To close the pivot table after editing, Exit & Return to the output navigator from the Pivot Table Editor menu. You may also choose to edit the table in place by choosing Edit from the SPSS Pivot Table Object sub-menu or by simply double clicking on the table within the Output Navigator.
Q. What is an easy way to print one table per page?
A. From the EDIT menu, chose SELECT... ALL. PIVOT TABLES.
Then, go to the INSERT menu and pick PAGE BREAK.
Q. Displaying Both Variable Names and
Labels
A. When I go to OPTIONS and the OUTPUT LABELS
tabbed dialog box to request that my table labels display both the
variable name and label, it does not display that way when I generate a
table after making that specification.
There is a bug in SPSS 7.0 in which hitting the OK button after making your tables labels request does not produce the desired results. The work-around is to first select Names for variables and Values for variable values and hit apply. Then select Both for both and hit OK.
Q. Changes to Batch Jobs
A.
Double-clicking on my .INC files (for running batch jobs) no longer works.
In SPSS 7.0, syntax files must have the extension .sps as that is the
extension registered in the Windows 95 registration database. Changing the
.inc to .sps will allow you to open them in SPSS by double-clicking them,
but this will not automatically run them. Select all and then press the
run button on the toolbar. You can also put an SPSS shortcut into your
Windows 95 SendTo directory to send .inc files into SPSS without having to
rename them.
SPSS 6.1.3 for Windows: Installation Questions
Q. Overview
A. For best results when
installing SPSS for Windows, please specify a new or completely empty
directory when installing the product. If you presently have SPSS 6.1 for
Windows installed, you install over the existing copy. In addition, make
sure you have no other applications running when you begin to install SPSS
6.1.3 for Windows.
For improved performance of SPSS 6.1.3 for Windows, the following are recommended:
Q. Win32s Problems
A. The most
common errors that we have seen have been related to memory and are listed
below:
To set this parameter, edit the SYSTEM.INI file in the Windows directory using a text editor such as NOTEPAD or the System Configuration editor, SYSEDIT. Find the [386enh] section in this file and add the line:
PageOverCommit=8
Resave the file and restart Windows. Next go to the Virtual Memory Dialog box within the 386 Enhanced section of the Windows Control Panel. Change the swap file to temporary and use the maximum recommend size. Restart Windows. SPSS 6.1.3 for Windows should load and execute successfully under this configuration. Since paging is slower than the use of real memory, customers who will be using SPSS 6.1.3 for Windows extensively or for whom performance is highly valued, should seriously consider increasing memory to at least 8MB for optimal use..
If for some reason you need to delete an older version of Win32s, the following steps are required:
To install Win32s support without reinstalling SPSS 6.1.3 for Windows, place Base Disk 6 in your 3.5" drive, and select FILE --> RUN from the Program Manager. Enter the Win32s setup file, win32su.exe with the appropriate drive letter. This will execute the Microsoft Installation program. Note that the version of Win32s distributed with SPSS 6.1.3 for Windows is 1.25.