Skip to Content
Skip to Table of Contents

← Previous Article Next Article →

ATPM 9.05
May 2003




How To



Download ATPM 9.05

Choose a format:

Review: SPSS 11.0

by David A. Zatz,


Developer: SPSS, Inc.

Price: $1,145 (base); maintenance (recommended) starts at $287; upgrade from SPSS 10 for $299; upgrade from SPSS 6 for $499. Academic pricing starts at $599 (base). Many users will need extra modules. Training manuals (recommended) are around $100 each.

Requirements: G3 with 192 MB RAM, 256 MB of hard disk space, Mac OS X 10.1.2. Not compatible with Mac OS X 10.4.

Recommended: 500 MHz G4, 256 MB RAM.

Trial: None

The statistical analysis application SPSS entered the Mac market with a lean, efficient, syntax-driven program. Version 4 would let you use menus, but those menus would simply build syntax on the screen, helping users to use the powerful program as efficiently as possible. It had many deficiencies—among other things, it produced its results in a mainframe-style monospaced font—but it ran rather quickly on a Mac Plus, and helped beginners get used to the SPSS syntax language.


Unfortunately, after a disastrously bad port of version 6—which switched to a more modern menu-driven system, with syntax “also” available—SPSS abandoned its Mac customers.

Then, about two years ago, SPSS contracted with MacKiev to port version 10 over to the Mac. The result was a program with the capabilities of the Windows version, including a wide variety of plug-in modules and good speed. Now, version 11 has been ported over—but only to OS X. If you use OS 9, you must stay with version 10. But those who recently purchased new computers, only to find they cannot boot into OS 9—and that SPSS 10 won’t work in the Classic environment—will be very happy about the new release.

As with prior versions, as you slice, dice, and weight data, and pop out scatterplots using the built-in menus, SPSS automatically records your syntax into a “journal,” which you can later modify as a kind of macro to quickly run lots of analyses very quickly. I recommend buying the optional syntax guide and typing your commands as often as you can, instead of using the menus. Using SPSS becomes much, much faster that way.


Most of SPSS’ capabilities remain from version 10, and I will refer you to that article rather than going back through them.

As with version 10, the major drawback to the program is a long boot time. Version 10 booted in 30 seconds on my beige G3, despite an upgraded hard drive and processor. Version 11 takes 45 seconds on my blue and white G3, with an upgraded hard drive and 400 MHz G4 processor. No doubt it’s considerably faster on more modern models. While screen redraws are very sluggish on older machines, calculations are tolerably quick. I suspect the part of the program that actually does the work uses only highly optimized (and probably very old in many cases) code, while the user interface was developed with less concern for processor time. SPSS’ Windows origins are otherwise hidden. It looks and acts like a real Mac program, using standard OS routines.

The data, output, and syntax views all have different sets of menus, as they do in Windows: for example, you have to be in the data window before you can use a menu to split the file. This can be annoying, but it does keep the menus to a manageable size.

SPSS has done a wonderful job of maintaining the syntax through all Mac and Windows versions, so that files written for older versions work without a hitch. The data format is also compatible without conversion across versions and platforms, making it easy to share data and scripts with Windows people or users of older Macs. Data is stored in a highly compressed format, in a break from Microsoft’s norms of oversized files, but matching SPSS’ mainframe heritage. There are only two major issues with compatibility: SPSS 11 for Windows output files cannot be shared with SPSS 10 (or earlier versions) for the Mac, and SPSS’ scripting language (not to be confused with syntax!) was not ported over, since SPSS 11 can now use AppleScript.

SPSS 11, like prior versions, also reads tab-delimited and fixed-format ASCII text and Excel files. Output files can be saved as HTML for posting on the Web. SPSS also sells programs that allow other people to view the results and delve deeper into the data.

SPSS also deserves credit for porting over just about all the modules available under Windows, so Mac users aren’t faced with overpriced crippleware. Indeed, the prices match the Windows versions, despite porting costs. The main deficiencies of the Mac version, compared with Windows, are being one-half version behind (Windows is up to 11.5); missing features include not being able to directly import SAS files, the loss of the Method subcommand of Crosstabs, and three subcommands of nonparametric tests (both of the latter are missing from the optional Exact Tests module).

New features include direct publication to the SmartViewer Web Server, an extra-cost option that allows quick and easy dissemination of results throughout an organization (or to the world). Data can now be restructured so multiple cases can form a single case, or vice versa, which will save some researchers incredible amounts of time and nuisance. A database wizard can recode string values to numeric variables, obtain random samples, and do other clever, time-saving tasks that once required workarounds or other software. Other clever features for quickly rearranging data, such as the innovative pivot controls, remain.


The optional and very handy Reports module now produces modern output, with tab-delimited tables instead of the old monospaced fonts. It’s still only camera-ready if you’re not very particular, but it’s much better than in the past, and individual tables can quickly be modified for acceptable results. The Reports module can save a great deal of time when, for example, analyzing survey data. This and the Advanced Statistics modules are almost “must-haves.”

The OLAP cubes facility has been upgraded to allow calculation of differences between variables. Those who hate scientific notation will be glad to know it can now be suppressed in all output. This upgrade does not just deal with system software and appearance issues; a number of statistical upgrades were also made. These can be explored fully on SPSS’ Web site, and may affect a surprisingly large number of users.


Most people will probably export SPSS tables and charts to other programs for publication, usually by copying and pasting to the other application, but SPSS 11’s tweak to output tables—adjusting the column width to match output—makes it easier to simply save output as PDF (or HTML) and use it without modification. This may be the greatest time-saver of the new version for some people.

While SPSS seemed stable on my system, an early blue and white Power Mac with only 448 MB of RAM and a 400 MHz G4 upgrade, I did experience a corrupted preferences folder which required re-installation (resetting permissions did not help), and the program seemed to lock the Command key down at times. Syntax pasted in from the Classic Scrapbook did not work properly and had to be retyped, though syntax from older versions ran well when opened normally. In extended testing, SPSS sometimes crashed when being closed, and some parts of the output sometimes refused to use certain fonts.

SPSS 11 makes it easy to take heavy-duty statistics into the world of OS X. The fact that SPSS 10 and 11 are mutually exclusive also makes it easy to decide on upgrading—if you stay in OS 9, you cannot use SPSS 11, and if you move to X, you have to.

Because SPSS has such a slow interface, I strongly recommend that you do not use it with older machines or less powerful laptops. A 700 MHz G4 or above is just about required for tolerable speed, and a faster processor is really needed to make the program enjoyable in everyday use. If that means staying with OS 9, it may just be worth it to keep the relatively snappiness of SPSS 10. Version 11 does not add enough new features to justify an upgrade for many users (except, as noted, for OS X compatibility, which is downright needed for those with machines that won’t boot into 9).

SPSS 11 is capable, fairly user-friendly, and easier to work with than previous versions. The number crunching is very fast, and the reports increasingly easy to use in presentations and articles. It is no small wonder that SPSS 11 is still the market leader in professional statistical analysis.

Reader Comments (69)

anonymous · May 7, 2003 - 11:18 EST #1
You said, "It is no small wonder that SPSS 11 is still the market leader in professional statistical analysis."

Hardly. One word: SAS. SAS has yet to come to Mac OS X. It should. SPSS is good, but SAS is ubiquitous and is the standard in academia.
David Zatz · May 7, 2003 - 11:59 EST #2
Frankly, I'd like to try SAS. They treat their employees very, very well and seem to treat their customers well, too. I do face a very steep learning curve, of course, since I can handle SPSS as second nature. However, I believe that SPSS is the market leader, with market leader defined by number of sales. SAS is popular but, the last time I saw metrics, SPSS had them beat in statistical analysis. I think SAS probably leads in corporate intelligence and data mining by now, but that's another beast--based as though it is on the core SAS package! (Not unlike the PT Cruiser being based on the core Neon package--but being totally different in just about every meaningful way.)

It would be good if SAS was to simply toss their source code over the wall to MacKiev. After a year or so, we'd have a real choice in MacLand.

I don't know if these still exist but there is also MacANOVA and StatView. Neither is in the same class as SPSS or SAS.

I wonder if there are open source stats packages for OS X?

Incidentally, I actually prefer SPSS 10 under OS 9 to SPSS 11 under X, but that's largely a function of my computer. If I had a current model, I'd probably opt for 11 due to those small incremental improvements which make everyday use much easier.
Tracy Lightcap · May 7, 2003 - 14:05 EST #3
I use STATA as my main statistics engine on my 400 MHz G4 running OS X 10.2.4 with 636 MB of RAM. It is fast, ridiculously complete (13 volumes of documentation), programmable, and relatively cheap. Unlike SPSS and SAS, The STATA developers have kept some perspective on their pricing and there is an active users' community that constantly adds new routines. This, no doubt, is one reason why the application is growing in popularity.

For open source work in OS X, the choice is R. This is an open source version of S, done by pretty much the same development team. R has just been selected as the recommended stats software at the Interuniversity Consortium for Political and Social Research at Michigan, so expect it to be much more widely used in the future. I haven't downloaded it yet because it needs to be recompiled and, frankly, I don't need it. It's there and free for those who do, however.
Brian · May 7, 2003 - 16:03 EST #4
I saw the following on the SPSS list recently. An update is coming to address the speed problems. Yeah, v11.0.1 has a beta feel to it, but hopefully, this will fix it, and it beats running v10 in Classic. If you want to run v10 in OS X, just update to 10.0.8a (free on the SPSS site) then alias the actual application out of the application package and its "stuff" folder. Drag the actual Classic application to the Dock. See, it really was "designed with 10 in mind," just not quite enough to actually run wtihout modification, even in Classic. :-) v11 is very nice though. I can't wait for the update! --Brian
From: "Peck, Jon" {}
Date: Thu Apr 24, 2003 10:50:06 AM America/Indianapolis
Subject: SPSS 11 for Mac OS X

SPSS is pleased to announce that SPSS 11.0.2 for Mac OS X was sent to manufacturing today. This update release includes significant performance improvements and bug fixes. All owners of SPSS 11.0.1 that are registered with SPSS Inc. will be sent a copy of the 11.0.2 CD. This is an automated process and is expected to be complete in 28 days.

SPSS 11.0 for Mac OS X is the first version of SPSS for the Macintosh to offer linear mixed models, which ensure accurate predictive models when working with nested data. For example, with the flexibility of this procedure, analysts can formulate a wide variety of models and work with repeated measure designs, such as incomplete repeated measurements in which the name of variables differs across subjects. The mixed models procedure can be used to more accurately model student performance within classrooms, consumer satisfaction within families and in a variety of other research applications.

SPSS 11.0.2 for Mac OS X was beta tested by a set of users who were chosen because they had had problems with the original release, and the feedback was very positive. Here are some sample reactions.

"I downloaded and installed it immediately and checked a few things, that bothered me most with the 11.0.1 version and I was surprised to see, that everything seems to be perfect now."

"I am very happy to report that it works flawlessly."

"Thank you VERY much!"

If you have questions, please contact your SPSS sales representative.
Howard · May 7, 2003 - 21:59 EST #5
SAS makes JMP for OS X and it is much easier to use for statistical work than SPSS. I beta-tested earlier versions of SPSS and JMP. Once you use JMP, you will never go back. It is a rich, visual environment for data analysis. Previous versions integrated with Applescript and JMP's own native scripting language so that you could, for example, automatically export your Filemaker data, have it imported into JMP, run survival curves, and export the raw survival data or the pictures with a single click.
David Zatz · May 8, 2003 - 08:46 EST #6
JMP does some things very well, but there are other things it does not do well at all. I have contacted Stata and will try to get a copy for review. I hope ATPM will print the comparison. I am also installing the 10.0.8 revision to try it out and am eagerly awaiting 11.02.
David Zatz · May 8, 2003 - 14:06 EST #7
I have contacted SPSS regarding the issues raised here. First, they did have a number of bug reports on v11 which they were generally not able to repeat, but they did, as stated here, come out with v11.02 which is being sent on CD to registered customers soon. As far as faster speed is concerned, this may only apply to certain operations, not to overall responsiveness, but we will see when it arrives.

With regard to running v10 under OS X, they do NOT advise it. They said it could be made to launch, but would be buggy and that they would not stand by any results you may get. So I don't recommend it either! :)

I will take this opportunity to add something I left out this time around--namely that SPSS support is usually very good!
anonymous · May 12, 2003 - 09:39 EST #8
Not making patches (e.g. the license renewal utility) and upgrades (e.g. to v11.02) publicly available via the company web support area does not attest to good support.

The syntax editor in v11.01 has several bugs, e.g. after selecting text with the keyboard and typing Command-R to execute, normal editing functions (like cursor movements) often do not work until the shift key is pressed again, and copying and pasting using the keyboard (Command-C/-R) often do not work. Instead, the selected text is replaced with a c or v.

SPSS is really a memory eater on the Macintosh (and has been at least since v10), much more so than on Windows machines. Because v10 and v11.01 for Mac both install a large number of *.dll files, my guess is that an extra layer has been added on top of Cocoa or Carbon that, more or less, maps the Win API.
David Zatz · May 12, 2003 - 10:10 EST #9
My guess is that you are correct about the extra layer. This would explain the long startup time. As far as posting the 11.02 patch, I was told that it was prohibitively large in size, which is why they are sending it out on CD. I can understand that. Bandwidth is an issue, though the cost of SPSS is high enough that I'd hope they'd still post it. I suspect there are many people out there who download the patches for pirated versions, so maybe it's a bigger issue than you or I would think.

I did not experience those problems, but do not doubt that you did. Indeed, as time goes on, it seems that SPSS may be very sensitive to the environment it's in, indicating the need for greater testing in future. I believe the SPSS support staff when they tell me they did not experience any problems, though.

As I believe I implied, stick with version 10 if you can :) and I'll try ringing Stata for a comparison again. A few years ago, I did compare SPSS with Minitab back when Minitab was still being kept current and I can say that it's not so bad for Minitab to have fallen by the wayside.
Dana Leighton · June 1, 2003 - 07:24 EST #10
I have been looking for a good OS X native stats package to replace my venerable StatView, and believe I will settle on Stata (primarily because of completeness and user community). Aabel looks promising, but the stats routines (especially for inferential stats) are woefully incomplete so far.
Brendon Dellar · June 29, 2003 - 04:35 EST #11
SPSS v.11 for Mac OS X is too slow for machines with less than a 1GHz G4. You'd end up insanely frustrated if you intend on running this on a G3 of any description.
David Zatz · July 15, 2003 - 08:25 EST #12
I don't completely agree. I've been testing it with a blue and white G3 with a G4 400 accelerator and it's slow, but not intolerable. I will say that, for extended use, a dual CPU or perhaps 700 MHz would be almost necessary. But, there are some decently fast G3s out there that could handle it OK.

I did get the SPSS upgrade and it neatly solved that problem of non-starting with a message about deleting the preferences.
Dave McDowell · July 30, 2003 - 11:51 EST #13
I find that I still run into a problem with SPSS freezing up even after the update to 11.0.2

I'm using a dual G4 (867) with OS X 10.2.6.

Does anyone have a problem like this? Any suggestions?
Julie McBeth · August 8, 2003 - 06:37 EST #14
I agree with Brendon. The Alivec code in SPSS makes a world of difference in the performance of SPSS. A G4 is definitely much better than a G3. I've compared both and I wouldn't use a G3 unless I absolutely had to.
David Zatz · February 22, 2004 - 15:17 EST #15
I must note an additional SPSS bug: on both the G3 and a dual G4, in OS X 10.2 and 10.3, I find it keeps activating the command key. Anyone else? I've never been able to upgrade to the 11.0.2 version. It would be nice if they just posted the upgrades to their site instead of sending CDs.
David Zatz · May 14, 2004 - 09:06 EST #16
I've been able to make preliminary tests of SPSS 11.02.

While 11.01 took about 20 seconds to launch on a dual G4 (1.2 GHz), 11.02 took about 5 seconds. That's a major improvement. Once launched, it seemed more responsive. I will work with it for a while to see if an update to this review is needed.
zaiti noreen rosli · September 22, 2004 - 10:43 EST #17
how can i download the spss trial version 11.0?
Dave · September 22, 2004 - 11:06 EST #18
There is apparently no trial version available at this time - and the Windows version is up to 13. I suspect SPSS has dropped Mac support yet again. I have inquired.
Dave · September 22, 2004 - 13:40 EST #19
OK...word from SPSS is that there will be a new Mac release in mid-to-late 2005. They aren't sure if it's 12 or 13. Turns out 13 hasn't been released for PC yet either - it's "coming soon." 13 contains the long-sought (by some) long text fields, 32,000 instead of 255 characters.
anonymous · October 15, 2004 - 19:09 EST #20
11.0.2 doesn't work with G3's and Panther 10.3.3 or later. This includes the late model G3 based iBooks (800Mhz - 900Mhz) prevalent in education. SPSS aren't clear on whether they'll fix it or not.
David Zatz · October 15, 2004 - 20:41 EST #21
I am unimpressed with their current commitment to the Mac. It'd probably be better to have customers deal directly wiht MacKiev...
Mike Abrams · November 1, 2004 - 19:54 EST #22
I was mailed the 11.0.2 version against my will. I had purchased the upgrade over the phone and received sudden notice and apologies from SPSS sales that it would not work with my G5 and Panther.

They still sent it!

I don't know how SPSS continues to do business with pitifully inept customer service like this.
Tim Dowling · November 2, 2004 - 14:25 EST #23
"anonymous," I don't think that's right. I have used SPSS 11.0.2 without a hitch on several G3s running the latest Panther releases.
Chris Boyd · November 7, 2004 - 23:08 EST #24
I am purchasing a new mac laptop and will be running SPSS, since I also use SPSS for Windows at work and would like to avoid learning a new program (though I might be conviced otherwise). How much memory and speed do you reccomend that I have to have to have SPSS run smoothly. I will be managing data sets that generally have several hundred thousand cases, but can be several million.
David Zatz · November 8, 2004 - 08:16 EST #25
It does seem as though, despite the wonder of two versions being ported in a row, SPSS may be neglecting us again. Or not - I understand a new version is in the works. Of course it will be $$$, but at least it will be there.

Regarding SPSS on a laptop, 11.02 works fine on a G4. Not sure about how much memory you'll need. Speed...well, "need" is hard to say. I'd want to have at *least* 512 MB of RAM to run OS X regardless. SPSS is a RAM pig for the user interface but the actual calculation space seems more reasonable, so you could always tart there and build up later if needed. I will point out that it will be a bit sluggish on a single-processor G4 laptop, though still usable - again, calculations will be faster than the uesr interface suggests! - and that RAM costs more when purchased from Apple.
chrismcr · November 20, 2004 - 14:21 EST #26
After fresh install, SPSS 11.0.2 does NOT launch on on G3 iBook 700mhz nor a B/W G3 350 mhz, both running running 10.3.6 . I see some, but not all, have had success with G4 machines. This update was originally made avaiable in 2003 and it is unthinkable how they cannot provide a solution in the course of a YEAR. Our university has a site license for both Windows and Mac SPSS and I will be making a reccomendation to the administration that we look elsewhere (STATA, JMP).
David Zatz · November 20, 2004 - 15:18 EST #27
It *is* hard to stay where we're not wanted. SPSS has been unresponsive since version 4 and though this most recent flurry seemed to bode well (10 and then 11), they are glacial in their updates. Version 12 or 13 is coming soon, I'm told, FWIW since the upgrade will cost almost as much as just switching to Stata.
Dave Rahn · December 1, 2004 - 18:53 EST #28
Wow. I have been spending the afternoon trying to decide which stats program to purchase. In an earlier Mac I was very satisfied with Statview, but my research needs are a bit more ambitious and my OS X G4 laptop with 1.33 Ghz has been whining for a challenge. At the top of the page I was ready to commit to SPSS. The support issues make me wonder if Stata is not a better investment just now.
David Zatz · December 1, 2004 - 19:17 EST #29
It would be very good to know when the next SPSS update will be published. Still, Stata does seem to have better Mac far as I can tell...I've never seen it!
Paul Redfern · December 31, 2004 - 07:45 EST #30
I am convinced that what anonymous (1) said about SPSS 11 for Mac mapping the Win API is true. I have experienced real frustration in using this product: problems which replicate known issues in moving from a Unix to a Windows environment. E.g., you can only save syntax once onto a Windows formatted drive. If you then open it in any Windows text editor, including SPSS syntax editor, all the line feeds have been stripped out. If you try saving an edited syntax file a second time, you have to 'save as', upon which command it is in fact saved as a '.' file, so invisible to Unix systems. Imports of syntax originally written in Windows arrives with lines missing (like this!). The application is not scriptable.

Support has been poor to non-existent (denial of bug, followed much later by admission of problem, but _no_ attempt/offer to rectify it). I paid £1500 plus £300 annual "maintenance" fees. What am I getting for this?

Despite your remarks about SPSS 6 for Mac, SPSS 6 for Unix (Solaris) was a stable, if spartan, application. Why couldn't/can't they simply recompile that app for the IBM chip?

p.s. I did put it to SPSS (UK) tech support that their product was simply a Windows application wrapped in a Mac emulator - they denied it point blank. I suspect also that what anonymous(1) needs to do to copy and paste is use CTRL-C, not cmd-C (rant, rant)
David Zatz · December 31, 2004 - 08:54 EST #31
I have never run SPSS in Unix-Unix, only in OS X, but I will say that there are lots of ways to convert the syntax files - all sorts of "line feed change" utilities, opening in BBEdit Light and using the line feed search-replace, opening in Word, etc. (When moving things to Windows, I often open Mac files in Wordpad, which will convert the line feeds automatically; ditto moving to Unix where I use nano, which is similar to pico, but detects and "fixes" Mac line feeds.)

A note to SPSS tech support on the line feed issue is called for. However, this can hardly be blamed on "mapping the Win API." It's simply a programming oversight which is common to some other programs. The good news is that SPSS data files are easy to move. I've never had the "file starting with ." problem.

It IS shameful that on the Mac you often have to drop stuff off into BBEdit before moving to Excel.

I agree on the costs being far too high given the total lack of bug fixes.
Jeff Nailen · January 22, 2005 - 19:36 EST #32
Great review and discussion Dave. I know the Mac versions haven't kept up with the Windows versions, but I'm glad to have a Mac version available at all given our low market share/niche status that will hopefully change with the new Mac mini and iPod craze.

I've been shopping around for a multidimensional database that runs on OS X and the one that powers SPSS is the only one I've found. Do you know of any others?

Their customer service has been great for me so far. I emailed them regarding their next Mac version before I had found your review. Good news: version 13 for Mac OS X is the version that will be available sometime this year, they're skipping over version 12:

Hello Jeff,

We're in the process of scheduling a new Mac release for later this year. It will be SPSS 13. We have no official release date yet.


Marcus Hearne
Product Manager
Survey Applications Business Center
Phone: 1.312.651.3731 / Fax: 1.312.264.3731
Jeff Nailen · January 22, 2005 - 21:44 EST #33
By the way, I checked out STATA and JMP, and while they both look good, neither have a multidimensional database and all you can do with that: OLAP functions, datacubes, pivot tables, 3D graphics, etc.
David Zatz · January 23, 2005 - 12:08 EST #34
When you say "multidimensional," what exactly do you mean? Like relational? Because the database that powers SPSS is pure flat-file. Perhaps I'm showing my ignorance or perhaps we have a wording difference. If you like OLAPs, datacubes, pivot tables, etc., you may want to look under "report writers" instead, since that's where most of that would be. Surprised Stata can't do that sort of thing...JMP of course is a visualization tool at heart.

Glad they're shipping 13 soon! Though I sure hope they've made some improvements to the core engines rather than just adding new statistical features. I long for the days when you could get a new release that contained speed boosts, bug fixes, and user interface updates rather than new "checklist" options...
Paul Redfern · February 10, 2005 - 07:45 EST #35
(Apologies for delay in replying - my Mac processor fried a month or so back).
1) Line feeds: I have written/helped write a couple of line feed utilities in Perl to replicate the 'dos2unix' utility, which I will happily send/post if you wish.

My point in mentioning this isssue was in support of my belief that SPSS 11 for Mac is in fact a Windows program in a Mac wrapper; namely because the text outputs from its syntax editor do not 'behave' like text outputs from native Mac apps. (And yes, I have mentioned this to SPSS tech support, as I have all the other issues I cited here and in my original post. Their response? I shouldn't be trying to export/save/write syntax on a Windows formatted partition. After much complaining, they finally accepted that these _were_ bugs, but other than reporting them to Chicago, they have done nothing about them)

Line feeds are however a minor issue. Much more important are missing lines and hiding syntax files.

2) Missing lines: e.g: what was originally, say,

compute hunts=3.
if (x ge 300) hunts=1.
if (x lt 300) hunts=2.
val labs hunts 1 'In favour' 2 'opposed' 3 'No

becomes, when opened in SPSS for Mac:

compute hunts=3.
if (x ge 300) hunts=1.
if xal labs hunts 1 'In favour' 2 'opposed' 3 'No

A similar problem can happen at the end of the file: according to the syntax editor, you've reached the end of the file (the EOF marker), according to the processor, the end comes a couple of lines earlier/later - really aggravating when the file in question is an 'include' file. To repeat, this only happens if the file is opened in SPSS for Mac. If opened in any other Mac-based text editor, it displays exactly as it did originally.

When you are trying to import 500+ line syntax files, this is a serious labour/time resource issue. You've debugged them once, now you have to debug them again.

3) Inability to save syntax in a Windows partition: Again, this creates serious labour/time resource issues if you are trying to evaluate/de-bug syntax on a cross-platform network. To repeat, using SPSS for Mac, you can save syntax once and once-only on a Windows partition. If you try to save it a second time, it won't let you (the file's become 'read-only'). If you try to 'save as' the file is saved as with a name like '.12344xxcsq1', and is hidden (on Windows as well as on Mac OS's) to boot. The original file disappears altogether.

Why do I appear to be trying to do things the hard way?
Well, we used to run SPSS for UNIX, exporting the output files into Word for Windows, to WP them for mailing to clients. Over the years we've built up quite a library of syntax files. We had to move away from the UNIX enviroment (long story), so ported everything to Windows (no problems, apart from the shift from relative to absolute file referencing). The downside of this move was that we lost the ability to edit batches of syntax files using utilities such as UNIX 'sed' stream editor.
The UNIX-based MacOSX gave us the opportunity to return our SPSS operations to a familiar UNIX-style environment. And so we invested a lot of money, and now time, in switching to an OSX-based operation. Except, we can't take advantage of any of the facilities offered by a UNIX-style environment because (IMO) SPSS for Mac is not a native MacOSX app, but is simply the Windows app inserted into a MaxOSX emulator.
Yet it would be so simple to create a native MacOSX app. All you would have to do is recompile the UNIX app (presently compiled for Sun's Solaris chip) for the IBM chip used in Apple products. The output wouldn't be as fancy, but at the end of the day, an average is an average, sum(x)/n(x), no matter how you dress it up.

Maybe SPSS 13 will be that app. But I bet that, like all their releases in recent years, SPSS 13.0 will be the beta version, and 13.0.5/13.1, the de-bugged (by us) alpha version. And they charge _us_ maintenance fees?
David Zatz · February 10, 2005 - 08:08 EST #36
Your troubleshooting brought things up that mine did not - very good and disturbing points! One possible solution would be to store the syntax on a cheap ($400) network-attached storage box formatted for UNIX or MacOS.

I agree that 13.0 will most likely be full of bugs and even slower, with key outstanding issues untouched. SPSS has been moving inexorably away from statistics and into "business intelligence" where the standards are presumably lower.
mathieu Roy · April 17, 2005 - 14:58 EST #37
any news on the spss 13 front?
David Zatz · July 6, 2005 - 16:03 EST #38
No news so far as far as I can see. Maybe it's being delayed for Mactel ;)

By the way, I've revived (with Joel West's help and blessing) the MacStats site. It's now at and has a BUNCH of SPSS alternatives. I plan to re-review SPSS in light of this at some point.

By the way, Paul, have you tried creating a sparse image and saving THAT on the Windows network? Then you could save the Mac files to the PC server in a Mac filesystem.

I agree on the beta versions, maintenance, etc. I sure wish Stata et al would provide an SPSS syntax conversion utility!
Mathieu Roy · August 4, 2005 - 09:32 EST #39
Got word from SPSS.

version 13... pushed to Q1 2006.

at that time, an announcement about the plans for the transitiion to intel will be revealed (could go either way, I guess?).

version 11.0.4 will be out in the next couple of weeks to fix tiger compatibility and some bugs (should have been out last week of july). Should also take better account of dual processor systems.

There you go....
Dave · August 4, 2005 - 09:35 EST #40
Hmmmm. Well, it does seem likely that they'll make version 13 MacTel only, or at least MacTel compatible.

Version 11.04 has been promised next week for two weeks...

I did hear from a reader recently that you CAN run SPSS 10 under Mac OS X - even Tiger, 10.4. See for instructions!
kallur · August 10, 2005 - 12:30 EST #41
I have notices that sometimes when I switch between some windows, the menu bar does not adapt to the window. For instance, after switsing from a syntax editor window to a data editor window, instead of the date edior windows menu bar, the syntax editor windows menu bar is displayed.
Mathieu Roy · August 17, 2005 - 09:24 EST #42
oh well... I've ordered a copy of the SPSS 11 Grad pack to see how it runs on my current machine (iBook600G3), and tide me over until version 13 comes around "next summer" i presume, without support for Mactels until 2008. This is specualtion of course... but I have a feeling this is how it will play out.

I'm certainly pissed off at the quality of support SPSS gives it's mac products. Unfortunatly my lab is currently all windows, except for my trusty iBook.
anonymous · September 2, 2005 - 11:01 EST #43
11.0.4 is out.
David Zatz · September 2, 2005 - 11:10 EST #44
SPSS 11.04 is a WONDERFUL update that negates much of my criticism. Suddenly SPSS is responsive and quick; the bugs I found are seemingly gone; I can instantly copy and paste from SPSS to Excel and BBEdit - directly, too. The menu bar switches instantly from one window to the next. It's an amazing feat. I don't know if it fixed Paul Redfern's issues but at least it plays nicely with datasets stored on Windows servers, according to the company.

Release 13 will be produced, SPSS says, next year; we don't know if MacTel support will be included but unless they're insane, it will be.
David Zatz · September 2, 2005 - 11:11 EST #45
One more thing... a reader pointed out that you CAN run SPSS 10 under Classic. Instructions and details are at
Cesar · September 6, 2005 - 19:43 EST #46
Does anyone knows about this..I am new to SPSS

Question I was trying to get answered is this -- in this version of SPSS (11 for OSX) is there no longer a restriction of only 8 characters for each variable name? I’ve heard this is true for the latest windows version, but want to know for the MAC.

thanks in advance

David Zatz · September 6, 2005 - 20:43 EST #47
It may be true for Release 13 - I don't know - but not for Release 11 on any platform to my knowledge. Release 13 *is* coming to the Mac according to SPSS but not until next year. It's not a huge hassle given the variable label facility but it is a bit of an annoyance.
Mathieu Roy · September 18, 2005 - 20:37 EST #48
I don't know about other users' experiences but 11.0.4 on my 600 mhz iBook G3 (640MB RAM), running 10.3.9 is really smooth for me.

I think I will be happy to use it on my somewhat outdated machine.
David Kalisch · October 8, 2005 - 17:37 EST #49
Hi there!
At first, sorry for my bad english!
For a clinical study i have to use SPSS. Everybody in my hospital is windows user except of me. So my question is: what kind of problems, wil I have to face, while running spss 11. on my mac and the others using version 12 on ther PC's.
thankful for every help!!
David Zatz · October 8, 2005 - 17:54 EST #50
As far as I know, there should be no problems. I've never had problems with different versions before ... even going from 4 to 11 and back.
David Kalisch · October 17, 2005 - 18:02 EST #51
Thanks a lot!
anonymous · January 12, 2006 - 08:18 EST #52
Anyone hear anyhting about MacOS X86 compatibility?

I mean running spss 11.04 on the new macbook for example?
David Zatz · January 12, 2006 - 08:39 EST #53
Well...on the one hand, I'd like to think there should be no problems at all. On the other, we have not acquired a MacBook and do not plan to until the iBook equivalents are out; our first MacTel will probably be the Mini or whatever the PowerMac version will be (Mac Pro?). If II hear anything, I'll post it.
Mathieu Roy · January 16, 2006 - 18:06 EST #54
from spss:

"We've yet to get our in-house Intel-based iMacs in order to do some thorough testing. Preliminary testing by Apple and others has shown SPSS 11.0.4 to work without error, but until we do a complete suite of procedures in-house we cannot know for sure.

A more public announcement on the expected delivery of a version of SPSS Mac with universal binaries will come out at a later date. "

looks good...
David Zatz · April 10, 2006 - 09:49 EST #55
Gotta love SPSS' commitment to the Mac. ;) They'll be the last to come up with the goods... I guess SPSS is great for those who always buy two-year-old machines!

See for a list of stats programs with compatibility...
Paul Redfern · April 20, 2006 - 12:04 EST #56
Hello David,

Sorry for the long delay in responding. I think that there is something funny going on with ATPM's comments service. Although I obviously selected the 'Email me new comments' checkbox, nothing came through until last week, when I started getting comments on 'what a great game' this was.

The icy sarcasm drew me back to your site, but no, it appears they really were talking about some computer game, and I'm down to receive responses from that site, not this!

I'm afraid I don't know what a 'sparse' image is, although I appreciate the suggestion. However, unless my Windows-based clients can access the data/syntax contained there it won't solve the problem.

Anyway, to take up your invitation to comment on 11.04, I'm afraid it made no difference whatsoever to the issues I raised.

Did I mention the line feed issue when reading in data inputted in a Windows environment? If I look at the data in a regular (Mac) text editor, it looks absolutely fine, exactly the same as it did when it was input on the Windows machine

Yet if a client sends me rawdata inputted in a Windows environment, SPSS for Mac, 11.04 included, discovers linefeeds in random places, such that the first 40 or so cases will be read in correctly, but then I have to add a couple of blank lines into the dataset to enable the next 30 or so cases to be read in, then another 3.5 lines for the next 15, and so on and on.

I actually did at one time have the head of UK SPSS support on the line to me, and beyond him telling me of the great new update (11.04), he could promise me nothing. If Chicago don't/won't address the issue, Guilford certainly won't.

I'm now refusing to pay any more to SPSS. I reckon they're in breach of the both the Trades Descriptions Act (it's not SPSS for the Mac, it's 'passing off' something that looks like the real thing for the real thing) and the Sale of Goods Act (it's not as described, its not of mercantible quality and its not fit for the purpose).

Give me my money back (not you of course)!
David Zatz · April 20, 2006 - 12:07 EST #57
I am not objecting to what you say which I'm sure is true. However you can filter the line feeds using TextWrangler, WordWrapper, or other utilities - a nuisance but not TOO bad.

A sparse image is a type of disc image, a file that can be mounted by the file system and looks like a hard drive or CD drive.

SPSS has no idea of what is in SPSS Mac, it's written by MacKiev using SPSS mainstream code. I have been told that we can expect SPSS 12 or 13 or 14, and believe it will be Universal when it comes.
Paul Redfern · April 20, 2006 - 12:34 EST #58
Hello David,

Yes I know it's a nuisance and it can be done, but let me make it clear that you can't see which lines are problematic until you run it through SPSS. So you can't develop any general utility to spot the problematic lines, so in the example I just gave, I'd have to try reading in the data at least 3 times. This is to say nothing of the issues about reading and writing syntax we've discussed before.

I hear also what you're saying about MacKiev being responsible. If I'd known about all this before I'd bought the programme, I'd have bought a PC instead and installed the Windows version on that. If they were to give me my money back, that's what I'd do. But I'm stuck with this situation now. Hopefully, my experience will prevent someone else making the same expensive mistake.

As Deng Xiao-Peng famously said: What does it matter if a cat is black or if it is white, so long as it catches the mice? I love my Mac, but this programme, on which my business depends, catches no mice.
David Zatz · April 20, 2006 - 13:06 EST #59
Well, you can at least buy Parallels now for the Mac ($39) and run SPSS for PC on that...under Windows...when you need it. If you have a MacTel which fortunately I do - and htat was the prime reason for buying it, to run the odd Windows POS.

You can indeed spot the problematic lines using TextWrangler or a similar utility. They'll look like spaces but be an invisible character. Easy to spot and script out.

Shame you can't switch to Stata or something but if you need to share SPSS files... the GOOD news is that they do seem to be getting more serious about Macs at SPSS and I'd expect that sooner or later they'll start taking it in-house. Or MacKiev will start getting better.
Atul Sharma · May 30, 2006 - 13:44 EST #60
I have used SAS and SPSS for many years, usually depending on which was offering a more up-to-date Mac product at any given time, with neither winning awards in that department. Nevertheless, I have recently switched on both platforms to R, a powerful language, with an active user community, solid on-line help (especially for an open-source project), and user-group Mac support on par with any commercial competitor. It's only deficiency in my mind is the 1970s-vintage console interface, but there are some full-featured command editors available (TINN-R, Emacs ESS, JGR, etc) and even some GUI's suitable for teaching introductory statistics (RCmdr, PMG, RPad). I don't think I'll ever look back, and hope that more people will adopt it and work to improve the Mac port, with a GUI interface being much needed for teaching and intermittent users. Unfortunately, the one reasonably comprehensive GUI (R Cmdr by John Fox) requires the Tcl toolkit, which meant it would run only under X11 on Mac OS X. However, Aqua TCl libraries were supposed to ship with 10.4 Tiger (sorry, I'm still using Panther), which should go a long way to make installation of R Cmdr more of a point and click endeavour suitable for introducing statistics at the undergraduate level.

Marit Hauger · January 26, 2007 - 09:21 EST #61
First of all- I do notknow anything about computers... I bought a MacBook 13 with an socalled Intel Core Duo prossessor a couple of weeks ago. Now I am going to write my thesis with use of SPSS.I can not install the version I have- SPSS for MacOSX 13. Which version should I purchase instead? And where do I buy/install it? How much do I have to pay? Thanks, Ma
David Zatz · January 26, 2007 - 09:31 EST #62
SPSS 14 or SPSS 15 - whichever is next - will support the Intel. All you can do now is (a) use a G4 based laptop or other non-Intel Mac with 13, or (b) install Parallels Workstation and run SPSS for Windows, or (c) switch to Stata. SPSS may or may not give you a refund; they did say in advance it would not be Intel-compatible, though their insanely slow speed in bringing out Mac versions is lamentable and makes one question their commitment (as if they have one; the software is really done by MacKiev).

Stata has stood by the Mac for a while now with full, 100% support. Unfortunately I find SPSS much easier to use!
Mathieu Roy · January 26, 2007 - 09:35 EST #63
Hello Marit,
Did you try instaling it (ver. 13) or are you not installing it because it says it does not support the new intel processors?
I don't have ver. 13, but ver. 11 for mac works fine on my coreduo macintosh (a little slow, since it runs inside a translation layer, but faster than my ibook g3 ever was). SPSS are reportedly working on a new intel compatible build but it's not easy to say when that might be done.

So, my first question is have you tried to instal it? if it doesn't I could give you a few options.

Sorry to hear about your troubles.
Mathieu Roy · January 26, 2007 - 10:32 EST #64
Good info here:
David Zatz · January 26, 2007 - 10:42 EST #65
I'm blushing.

Seriously, if version 13 is installed, and working, chances are very high that you can run your stats wihtout problem. However, it is POSSIBLE that some of your results could be wrong esp. if you do a lot of iterative analyses. Simple frequencies, sig testing, etc. SHOULD be OK - I haven't seen any differences in my tests, but they weren't as extensive as those SPSS might have used. If indeed they do test the Mac versions, and didn't just post a warning to avoid legal liability, or to get us all to buy version 14 (or will it be 15 before we get Intel nativity?)

Can you tell I'm frustrated with SPSS? Really, besides their own record on the Mac - with the excellent Release 4, awful Release 6, and mediocre releases since then, and no real support after you buy - there's the companies they bought that had reasonably priced Mac products, then quickly had no reasonably priced products AND no Mac products. It almost seemed like they hunted down good Mac software and killed it (not unlike Computer Associates). Then they have lots of promises but their delivery is oh, so slow...

I think SPSS survives because Stata is still stuck in the mainframe view of the world (no simple cut and paste into Word or Excel tables), Statistica seems too far out of the mainstream (I can't say from experience), R is a typical Linux/open-source project (great if you spend a few years learning it, I guess), and SAS just isn't on the Mac at all.

Stats programs remind me of web design programs... I can't see why either one is so hard to use, so frequently lacking obvious, minor changes that would make them much better, and so demanding of system resources (except of course Stata!)... and for that matter SO EXPENSIVE. I think we can all confidently say SPSS doesn't put all THAT much work into each new version to justify their pricing.
bernadette · August 7, 2007 - 01:20 EST #66
hi mathieu

i have the same problem as marit -- only i've tried to install spss v13 but will not install. is there a way to get v13 installed? i know someone said that some stats are inaccurate, but i'll take that risk...

Mathieu Roy · August 7, 2007 - 10:15 EST #67
I don't have version 13 myself. I have been using 11...
Sorry I can't help you more.

Anyone hear anything about the next version coming out soonish?
dave clancy · August 26, 2010 - 09:05 EST #68
what are the things spps does that jmp doesn't do well. I have never come across an easier program to use than jmp. And for survival analysis it is unbeatable.
David Zatz · August 26, 2010 - 12:52 EST #69
There's a possibility that I'm biased by being used to PASW, but I have to say, dealing with survey data is one area where I found JMP very difficult in comparison. Blindingly fast, but not really suited to path analysis using regression, generating summary tables, and such. Haven't played with simple stuff like t-tests yet or with really complex stuff. I need to spend more time in JMP, I admit. I have had some stability issues in JMP when dealing with imported data.

Add A Comment

 E-mail me new comments on this article