Why use a data analysis package like SPSS when you could use Excel? I’ve just come across an interesting marketing piece from SPSS that goes into benefits one gets from using a dedicated data analysis package instead of trying to do all of your analysis in a spreadsheet. While it would be fair to expect that this isn’t necessary an unbiased comparison, it might offer some food for thought to those of you trying to figure out why you should bother to upgrade.

Key reasons offered by “Discover Secrets Your Spreadsheet Can’t Tell You”:

• Easy access to descriptive statistics and frequencies: True. while you can do descriptives in Excel using some of the built-in functions and the data analysis add-in, it is a lot easier and faster in SPSS.
• Wider variety of charts & graphs: True, although I tend to find Excel easier to manage
• Better, more flexible pivot tables  Sortof true. That is, true if you have SPSS Tables. If you don’t, then in my opinion Excel pivot tables are easier to work with. SPSS Tables, on the other hand, is extremely easy to use and lets you do a lot more things that you can’t do with Excel Pivot Tables.
• Full set of statistical tests: True. While it is definitely possible to run statistical tests in Microsoft Excel, they’re much harder to find and work with compared to SPSS, where they pretty much come “free” with every function you run.
• Easy to run similar reports and graphics for subsets: True. Using the “Split” function in SPSS, it is relatively easy to create tables and charts for subsets without doing any extra work. Or you can create syntax (SPSS’s macro language) that lets you reuse your tables and codes over and over again.
• Labels instead of codes in your reports: I love this feature. Just because your survey software makes Male=1 and Female=2 doesn’t mean you want to see lots of 1s and 2s in your reports. And while it isn’t difficult to use search/replace in Excel to change all of your 1s to Male and your 2s to Female, SPSS lets you keep your values intact.
• Accurate results when some data is missing: Sortof true. For this item, they point to the benefits you get from using the SPSS Missing Value Analysis add-on module (an extra $800 or so). This tells you whether the questions that were skipped by your respondents will impact your analysis, and will even estimate what these values should have been. Obviously Excel can’t do anything like that, but keep in mind you need to buy the extra module to get it to work.
•  Helps you spot data-entry errors or unusual data points: Certainly SPSS can help with this one, but I think you can get these types of results pretty easily in Excel.
• Easy import functions: I’m not sure that I completely agree with this one. It is true that it is easy to bring in text files. And they do provide functionality to bring in ODBC databases, including Excel spreadhsheets, Access tables and SPSS databases. But the interface for doing so is a little funky and the experience isn’t as clean or smooth as it is with Excel.
• Unlimited rows: This point describes how SPSS can handle an unlimited number of rows while Excel can only handle 65,000. Microsoft Excel 2007 can handle unlimited rows too, but SPSS’s assertion may have been true when the article was published.
• Using SPSS saves time and increases productivity: I suppose that really depends on what it is that you’re trying to do. There are a lot of analysis that I find easier to do in Excel. But certainly if you’re doing statistical analysis it is easier and faster in SPSS.
• SPSS makes it easy to understand statistical results. SPSS has added a lot of extra help files and tutorials that explain how you can/should interpret a lot of the statistical jargon that the software spits out. Excel obviously does not.

A few reasons why I still do a lot of stuff in Excel:

• For most people, the learning curve is much less steep with Excel: Learning SPSS was initially an unpleasant experience. It has a lot of options that don’t make sense until after you’ve spent a lot of time with the program. Once you’ve learned the software you’ll be amazed that you ever lived without it (or some other data analysis package) but until then you’ll spend a lot of time cursing it.
• It’s expensive. Especially if you already have Excel. Expect to spend over $1,700 for a copy.
• Charts are easier to manage/control in Excel: In my opinion. While SPSS has a lot of neat charting features, they aren’t as dynamic as Excel’s chart functionality — that is, when creating a presentation, I often need to go back in and tweak the charts or rearrange the data or rearrange the bars. In Excel, this is as easy as editing the underlying spreadsheet, which would automatically update the Powerpoint. In SPSS, you have to recreate the chart and recopy it into the presentation.
• More flexible use of functions: Excel has a lot more functions than SPSS and gives you more flexibility in how you use them.

Read “Discover Secrets your Spreadsheet Can’t Tell You”

Bertolt Meyer has written a not-so-happy review of SPSS for the Mac 16.0. His general thesis is that it is the “most insulting piece of software” he has ever come across. He felt that it didn’t look nor act like an Macintosh application; it isn’t properly internationalized; and more than a dozen bugs. (more…)

Let’s say you’re a market researcher, you have an extra $1000 lying around, and you’re looking for an easier way to improve the look, feel and efficiency of your cross-tabs. What do you buy? If you’re me, you buy the the Tables add-on for SPSS. While the text below certainly isn’t a detailed tutorial on how to use SPSS Tables, it should give you an idea of the features it makes available to help you decide whether it is worth the money. (more…)

SPSS announced recently to its Mac-based customers that SPSS 17.0 for Mac would not be released for the PowerPC based Mac, effectively discontinuing SPSS development for the PowerPC. PowerPC Mac users presently represent about 3% of all online computer users, down from 4.2% back in 2006 when the Intel based Mac was released. This based on the following letter which was recently e-mailed to SPSS for Mac users: (more…)

I spend a lot of time working in SPSS, and occasionally I need answers about various techniques and methods that aren’t readily available in the included documentation. Fortunately, there is a tremendous amount of free SPSS information and training materials scattered all over the web. Here are a few of my favorites. (more…)

Anderson Analytics and SPSS is offering a free webcast on March 20, 2008 at Noon EST in which Senior Consultant Jesse Chen will offer creative tips and tricks for analyzing unstructured (text) data. The webcast will last about an hour and will feature Chen using a variety of real-world case studies (probably integrating SPSS Text Analysis for Surveys).

• SPSS Event Info Page
• PR Web Press Release for the Event

SPSS has announced the release of SPSS Dimensions 5.0, an upgrade to last year’s version 4.5. Although in my opinion this feels more like an incremental release, it does have some notable new features including the SPSS Desktop Author, a new customizable question wizard, better integration with SPSS Predictive Enterprise Repository and SPSS Report. (more…)

…which I suppose is fine if you’re trying to run the software on Linex, or on a Macintosh, or on some other kind of platform where you only received new updates intermittently and always wished you could use the latest version.

However, if you’re on a Windows PC and are used to using a version of SPSS that is fully integrated into the Windows operating system, then the new version is a little jarring — and perhaps even a little annoying.

All of the same functionality is there as before — SPSS gets credit for that one. They must have put a lot of work into recreating every function in Java for SPSS 16. Most of the dialog boxes even look similar. All of the menus now have spiffy little icons which I’m sure will be helpful when it comes to quickly locating the option that you’re looking for.

Furthermore, all of SPSS 16 dialog boxes are resizable. All of them. Sometimes the way in which it resizes makes more sense than others, but it is especially handly to be able to resize some of the smaller dialog boxes where it was hard to see what you were doing.

Buy as a Windows user who is used to all of the Windows controls — you know, all of the standard drop down fields and the standard buttons, and the standard checkboxes — the switch to Java is a little wierd. Things in SPSS 16.0 don’t have the same feel as they used to. Things feel a little sluggish. Everything feels a bit like one of those online calculators or games that are written in Java.

But like I said before, everything seems to work. All of the buttons are in the same place they used to be. And if it speeds up the development time for new features — if it makes it possible for SPSS to meet my needs better in the future — then I’m willing to give it a shot and hope that I get used to what feels like a step backwards.

The jury is still out.

SPSS recently put out a news release (which has been picked up by at least two news sites CRM Today and TMCnet) whose sole purpose appears to be to scare companies away from using the plethora of survey research tools in favor of their multi-modal survey system. How can you begin to know what your customers are thinking, reasons the release, if you only ask those who are online when hardly anyone is even using the Internet these days?

"Web-based surveys may appear to be less intrusive and easier to conduct, but without pen and paper or a good ‘old-fashioned’ telephone, organizations miss the opinions of many, including those without a computer, the forever and selectively computer illiterate and a large part of the senior population that simply missed the tech revolution."

The release goes on to remind everyone that paper and phone survey are "in many cases essential, if organizations expect to present clients with the most accurate and complete view of customer attitutes and opinions." 

It then proves its point by referring to a recent Pew report:

"In fact, the Pew Internet & American Life Project recently found 49 percent of Americans only occasionally use modern gadgetry and many others bristle at electronic connectivity — the Internet."

Wow. Who would have thought that in 2007 more than half of the US population either doesn’t use and/or extremely dislikes the Internet. We all may need to rethink our online programs and go back to the phone banks, door-to-door solicitors and shopping malls many of us have mostly abandoned.

…but before we do, here are some stats not included in the SPSS release:

• Total US population is about 300 million people with 225 million of them over the age of 17. (US Census)
• There are 178.8 million web users in the US (comScore, June 2007)
• 71% of all adults are online (Pew)
• 87% of 18-24 year olds, 83% of 30-49 year olds, 65% of those 50-64 and 32% of those over 65 are online. (Pew)
• 73% of white, 62% of black, and 78% of English-speaking Hispanic are online. (Pew)
• 73% of people living in Urban/Surburban Environments and 60% living in rural areas are online. (Pew)
• 93% of those earning $75K+, 82% of those earning $50K-$74K, 69% of those earning $30K-$49K and 55% of those earning less than $30K are online. (Pew)
• Total number of households is 105.4 million (US Census)
• Almost 70% of US households have Internet access at home.  (Leichtman Research Group Q1 2007)
• 53% of US households have high-speed access (Leichtman Research Group Q1 2007)

The Pew study that SPSS refers to in their release is called "A Typology of Information and Communication Technology User." The study measures not whether or not people have internet access (as implied by SPSS) but instead tries to categorize people by the degree to which information and communication technologies are utilized and enjoyed.

According to the report, only 15% of the population can be characterized as "Off the Network" — that is, individuals with neither cell phones nor internet connectivity. They tend to be in their mid-60s, nearly three-fifths are women. Only 7% have college degrees (vs. the US average of 27%) and only 4% earn over $75K a year (vs. the US average of 22%). They are the heaviest users of "old media" such as radio and TV but do not have the inclination to try new information and communication technology.

Obviously, not everyone is online and if you’re looking for a particularly special group you may want to revert to paper, pencils and phones. However, I’m thinking that for most purposes you’re going to be able to find who you’re looking for online.

But the implication of the SPSS release is that unless you use (expensive) multi-channel research techniques (provided by them?) you will be collecting bad information and misleading your clients. This isn’t true and is in fact extremely misleading.

• Read the SPSS news release
• Read Pew Report "Typology of Information and Communication Technology Users"

In the upcoming SPSS Directions User Conference in Prague (May 16) Product Management Director Kyle Weeks will discuss some of the new features in SPSS 16. These include:

• SPSS 16 has a new Java interface allowing for Windows, Mac, and Linux versions of SPSS, a searchable Output Viewer, resizable dialogs and more;
• Improved data editor (adds find and replace capabilities to both variable view and data view). Also unicode support, import/export of Excel 2007 data, and an improved data editor;
• Syntax to change string length and data types; ability to set a permanent default working directory; elimination of short/long string distinction; ability to suppress the number of active datasets.
• More powerful statistics, including a new Neural Networks add-on module, a new Partial Least Squares algorithm, a new Cox Regression for Complex Samples module, support for algorithms written in R and improvements to Generalized Linear Models and General Estimating Equations;
• Latent Class Analysis in Amos 16
• SPSS 16.0 has improved programmability (see below)
• More integration with SPSS Predictive Enterprise Services, allowing you to store/retrieve and query to/from the Predictive Enterprise Repository via both the user interface and syntax
• Multi-threaded procedures for improved performance and scalability.

Other sources also report that SPSS 16.0 for Windows will use a new syntax editor. We can also assume that it will support Vista (since they still haven’t released a patch for SPSS 15).

SPSS has also indicated that some of the original functionality of SPSS Trends and SPSS Tables that has since been superceded by newer functionality will be eliminated. In SPSS Trends 16, the Exponential Smoothing, Autoregression, and ARIMA dialogs will be removed, while the more flexible Create Models; Apply Models; Seasonal Decomposition and Spectral Analysis dialogs will remain.

In SPSS Tables 16, Basic Tables, General Tables, Multiple Response Tables, and Tables of Frequencies will be removed, while the more flexible Custom Tables and Multiple Response Sets will remain.

It is worth noting that all of the functionality offered by the removed dialog boxes will continue to be available through syntax.

Details on the new Programmability of SPSS 16.0:

• EXTENSION command for user procedures with SPSS syntax
• Dataset features for complex data management
• New dataset class extends Python transformation program capabilities to multiple datasets
• Similar to INPUT PROGRAM but can read and write datasets
• Multiple input and output datasets
• User code written in Python
• Ability to use R procedures within SPSS through R Plug-In
• Provides ability to run R code within SPSS
• Use to take advantage of statistical capabilities in R
• Access active SPSS datset
• Write results to SPSS Viewer
• Improved implimentation of User Procedures
• Can be written in Python but specified using SPSS traditional syntax
• User never writes or sees Python code
• Used as if a built-in SPSS command
• Python module called with syntax already checked and processed by SPSS
• More general PLS module
• Dialog box interface tools coming in SPSS 17

I still wish it would be nice if they added the ability to organize variables in folders…maybe in SPSS 17?

• View the "See It in SPSS 16.0" Presentation (both Powerpoint and live demonstration)
• List of free SPSS demonstrations and events online
• Read the full SPSS Dimensions agenda
• Visit the SPSS home page
• Read the letter from SPSS regarding Tables 16 and Trends 16

Found an interesting comparison between the features of SPSS and STATA (two statistical analysis packages), as provided by several statisticians on Windows Live Spaces:

SPSS Advantages:

• Slightly more user friendly in making complex tables & graphs
• Nice routines for testing interactions in logistic regression models
• Friendly ANOVA commands
• Generally easier to use
• Sophisticated survival analysis

STATA Advantages:

• Much easier to run a probit
• Much better documentation
• Can do a lot more procedures than SPSS
• Great company support, friendly user base
• Multiple pooled cross sectional time series routines
• Count procedures (poisson, negative binomial and zero routines)
• Maximum likelihood estimators (Tobit, multinomial logit, ordinal logit, ordinal probit)
• Huber-White correction for heteroskedascity
• More comprehensive ANOVA routines
• Cox regression
• Duration analysis procedures
• Capability to estimate models for complex surveys
• Better weighting capability (pweights vs. aweights and iweights)
• Ability to take clustering into account
• Lots of user written solutions
• Much better handling of longitudinal panel data
• Event history analysis capabilities
• Panel data analysis capabilities
• Faster development than SPSS
• Better leasing arrangement

What this all means to market researchers I cannot say — I generally in my day to day life do not use many of the statistical procedures they describe and I’ve never tried STATA.

• Read the full comments

University of Missouri graduate student Mitch Hardin recently posted a note on his blog about how after spending a lot of quality time with SPSS he switched to R, a free software environment for statistical computing and graphics that runs on a variety of platforms (Windows, MacOS, Unix).

Although R is “almost entirely command-line driven,” Mitch likes the fact that it offers more information about what is going on and there are a lot of user-defined function. Personally, I can’t imagine using a command-line to do my stats processing, but then I’m not a big one for getting things done in SPSS syntax either. I couldn’t find much evidence of people using R for marketing research, but the software is free and I can’t imagine why it wouldn’t work if you needed a powerful statistical package but didn’t want to spend a lot of money.

• Mitch Harden’s post on Somatopsychic
• R Project for Statistical Computing

I came across a post on comp.soft-sys.stat.spss in which the purchaser of a new Dell system with Vista Home Premium edition was unable to install SPSS 13, 14, or 15. They indicate that they spoke to SPSS support, who told them that none of the versions which require activation will install on Vista.

In a follow up message, it was noted that SPSS was working on a patch and that the estimated date of release is the end of April, 2007.

• Read the thread of messages at comp.soft-sys.stat.spss
• SPSS product support matrix

Followup: SPSS has released a hotfix to address the problem. It can be downloaded from the SPSS Support Website (login required — you can use user: guest/password: guest). In addition to the hotfix, the page also identifies the procedure for installing SPSS 15 on Visita.

Not to go too much into detail yet, but we’ve been spending some time with SPSS Desktop Reporter 4.0. Wow. Despite the price, it’s a great software package with a lot of extremely useful features and worth every penny.

What it allows you to do is to take an SPSS file (or any variety of other data files), create crosstab tables and charts (doing all sorts of fancy analysis along the way) and then save structure of what you’ve created so that you can then rerun all of the tables and charts against an expanded dataset later on. Makes it great for surveys that feature multiple waves. But it’s also great for ad-hocs. If you’re a fan of the SPSS Tables module you’ll absolutely love SPSS Destop Reporter, which is kind of like what you’d get if you built an entire application around SPSS Tables.

Like most SPSS products, you definitely pay for what you get and you’re probably looking at spending around $2,800 for one copy (the price goes down for additional copies). If you have the budget available, it’s definitely worth taking a look at.

Learn more at the SPSS web site.

I recently received a request for information about the available data tabbing (cross-tab, table generating) software that is available out there and thought I might share what I know. Personally I was a little surprised that there aren’t more software packages out there that do this kind of work — the ones that are available are fairly expensive.

The idea of this type of software is that you can take a data set (perhaps in SPSS format or some other standard or not so standard format) and almost automatically generate tables for each of the questions in the dataset. Usually each table would feature a number of columns representing different market segments (the "banner") while the various responses to each question would be shown in each row.

Such software is usually capable of automatically doing various statistical tests, sometimes including but not necessarily limited to t-tests, z-test, ANOVA and Chi-Square. Some programs can also weight and sort the data, as well as merge data sets, provide sample balancing, generate charts to go with each table, and provide output into a number of different formats ranging from text files to Excel spreadsheets to Powerpoint templates.

WinCross is perhaps the most well known of the bunch, and has been around a long time (it is presently at version 7). It is entirely compatible with SPSS 15.0, allows for 32,000 data columns and up to 3,000 rows per table; offers a number of different statistical tests; and can important export in a number of formats. My biggest complaint is that the user interface seems like it was developed about 10 years ago, and that it costs $1,995 per copy (the more copies you buy, the less it costs). Learn more at the WinCross web site.

Version 2 of GMI Research Analyzer was recently released by Global Market Insite. It will also take your data set, crosstab it, create reports, and export it into a number of different formats — but it also allows you to do drag-and-drop analysis of you data and has a much friendlier interface. It can be purchased from the GMI web site for $995 per copy.

MarketSight is a completely online tool for analyzing your data and is sold as a subscription service for $995 per year. It offers much of the same data analysis and reporting capabilities as the WinCross or GMI tool, and it has the added advantage of being accessible from anywhere you havea web connection. There are some limits to the size of the files you can upload yourself (only 50,000 records) but if you’re looking for a fairly easy to use program that doesn’t require a lot of effort to learn, visit the MarketSight web site.

SPSS Desktop Reporter was recently released as an integrated member of the SPSS Dimensions suite of products. Based on my limited experience with the program it appears to be a very user-friendly tool that can be used independently of the other Dimensions products. If you’ve ever used mrTables — well, this is like mrTables, except that you have all of the data available on your desktop and there are many more options. SPSS Desktop Reporter has all of the features of the other software packages, plus especially easy integration with other SPSS products. It sells for about $2,800 per copy. More information is available at SPSS.com.

If you are on a "budget" (and I use that term loosely) and already have SPSS, you may be able to get SPSS Tables (a module available for SPSS) to do many of the things you need, albeit with the requirement that you do a lot of it manually. It sells for only $800 a copy and can be found on the SPSS web site.

Another package worth considering if you are on a budget is Memphis Market Intelligence’s Survey Explorer. Different versions of the software are available which offer the ability to work with differently sized datasets — at different prices. So if you are dealing with a medium sized data set (50 questions, with no more than 2500 respondents) you could get the "Personal" edition of the software for only $519, the "Professional" edition for $579, or the "Enterprise" edition for $749. There most advanced version of the software allows for an unlimited number of questions, an unlimited number of waves and up to 20,000 respondents and can be purchased for $1,439. Although I haven’t had a chance to play with it, it appears to offer many of the features of the more expensive packages and is worth looking into at the Memphis Software web site.

If there are any other packages that you would recommend, please don’t hesitate to send them to me.

• WinCross
• GMI Research Analyzer
• MarketSight
• SPSS Desktop Reporter
• SPSS Custom Tables Module for SPSS
• Memphis Software’s Survey Explorer