Scientific Computing
   Popular Searches:
lims, visualization, chemistry, statistics, hpc
DATA ANALYSIS



SITE SPONSORS
Home > Data Analysis > SYSTAT 13(2)

SYSTAT 13 
More Statistics, Less Effort 
John A. Wass
1 

Introduction
This month, your editor reviews another old friend. SYSTAT has long been a staple of the linear models crowd and now gets more powerful with every version. The present version includes Exact tests, Monte Carlo simulations, and the ever popular (and becoming more so every day) quality analysis section. As with most modern statistical software, it runs on 32- and 64-bit systems and loads trouble free (and I reviewed it on Vista!) from disk or internet. And they still offer perpetual licensure!

Upon first boot, the screen is refreshingly clean and easy to read. Tests and graphics are simple to find and most dialog boxes intuitive to understand. Even the first-time user will quickly master the nomenclature/use and will be up-and-running almost instantly. For those of us married to paper manuals, there are a series of these including 4 on statistics, and others on graphics, programming, data handling, and the specialized areas of Quality Analysis, Monte Carlo, and Exact Tests. While many appreciate menu-driven operations, for the diehard programming types there is the command line option. Table 1 lists the new features in version 13.

Selected New Features in SYSTAT13:

  • Error Variance with ARCH & GARCH
  • Best Predictors with Best Subset Regression
  • Fitness of Statistical Models Using Confirmatory Factor Analysis
  • Polynomial Regression
  • More 3D & 2D Graphics
  • ANOVA Upgrades
  • Basic Statistics Upgrades
  • Bootstrap Analysis Upgrades
  • Crosstabulation Analysis Upgrades
  • Hypothesis Testing Upgrades

Logistic Regression:

  • Simplified and easy user interface and command line structure to analyze binary, multinomial, conditional, and discrete choice models separately.
  • Option to specify the reference level for the binary and multinomial response models.
  • Simpler form of input data to analyze matched sample case-control studies with one case and any number of controls per set.
  • For Discrete choice models, SYSTAT 13 provides two data layouts: ‘Choice set’ and ‘By choice’ to model an individual’s choices in response to the characteristics of the choices.

Nonparametric Tests

  • Jonckeere-Terpstra test as an alternative to Mann-Whitney test: The test is used when the treatments are ordered in terms of the response. This test is based on the sum of the k(k-1)/2 Mann-Whitney counts (for k treatments).
  • Fligner-Wolfe test as an alternative to Mann-Whitney test: The test is used when one of the treatments acts as a control, to test equality of response to control vis-à-vis all other treatments, with a one-sided alternative. This is a Mann-Whitney test with two groups when the control is one group and all other treatments together form another.
  • Post hoc (multiple comparisons) tests of Dwass-Steel-Critchlow-Fligner and of Conover-Inman which can be used as a follow-up when a Kruskal-Wallis test shows significance.
  • A multiple comparison test due to Conover as a follow-up of Friedman test showing significance.

Now let’s look at a few of the specifics…

Screens/Statistics
Upon boot of the main system, the following screen appears:

2 
This Viewspace consists of four areas: Startpage, Output editor, Data/Variable Editor, and the Graph Editor. This allows for rapid access to any of the analytic areas that the user may be interested in seeing. The Startpage consists of Recent Files, Themes, Manuals, Tips, and a Scratchpad area for comments. When imported, data consists of rows that are cases, and columns that are variables. When the analyst right-clicks on a column and chooses ‘Variable Properties’, the variables may be defined as to type (unfortunately, but necessarily very important to modern analytic software). Once imported, the data can be examined for simple statistics, and transformed if the distribution appears problematic. As the more complete (read >2 or 3 variables) data sets are preferred in this column for examples, the Admire set (factors relating to childhood psychopathology) within the program was chosen for the descriptive statistics.
3 
A quick look at the data, both numeric and graphic, may be had by choosing Distributions/Basic Statistics and Distributions/Histograms from the main menu (only a portion of which is displayed below):
4 
For an analysis of variance, we looked at a data set on machine scores (desirability of some machine function) based on machine, operator and time:
 5 
We want a simple model for the score based upon the factors machine, operator, and time. Going to the main menu and choosing Analyze/Analysis of Variance/Estimate Model, and classifying the I/O as follows:
6 
we see that:
7 
8 
The above demonstrates graphically and numerically that the two main factors ‘Operator” and ‘Machine’ as well as their interaction are important to the score while ‘Time’ is not. To merely do an ANOVA on any factor is simple with SYSTAT’s Quick Access Menu:
9 
The main statistical menus are fairly extensive:
10 
11 
The graphics menu is also extensive and the present graphic quality is quite impressive compared to earlier versions:
12 
All of the usual options to modify these graphs are available both through the right-click and a separate graphics menu bar.
13 

Summary
Despite the power and depth in this version, the most impressive feature is the inherent simplicity. In the early versions the novice could easily get frustrated in deciphering what the dialog box was really asking for and feeding it the proper data types. This is to say nothing about approaching the level of sophistication required to fully understand the depth of the statistical results presented. Now everything seems so intuitive that as with many programs, the learning curve is tilted in favor of the novice (though it still requires some statistical understanding on the part of the analyst!).

While many programs containing the newer simulation and data mining software (SYSTAT 13 contains a highly efficient Classification & Regression Tree algorithm) come with astronomical price tags, this software is still quite reasonable. Interested readers are encouraged to download the 30-day trial, and the many intuitive operations should delight the novice.

Availability
SYSTAT13: $1,299 (Single license, full version with manuals), $999 (U.S. Government/Non-profit), $599 (Academic). Call for Site/multi-seat license pricing.

From
Systat Software, Inc.
225 W Washington St., Suite 425
Chicago, IL 60606
Phone: 312-220-0060
Fax: 312-220-0070
Email : info-usa@systat.com 
URL: www.systat.com 


Scientific Computing
Advantage Business Media
Rockaway NJ 07866

Email Article | Contact the Editor | Printer Friendly

Post to Del.icio.us | Digg This | Post to Slashdot
 






Bioscience Technology Chromatography Techniques Drug Discovery & Development Laboratory Equipment Pharmaceutical Processing R&D Scientific Computing
Advantage Business Media © 2012 Advantage Business Media
Privacy Policy | Terms & Conditions | Advertise with Us

Top Stories and Headlines
EVERY DAY!

FREE Email Newsletter