TURBO SPRING-STAT MULTIVARIATE STATISTICS (MSDOS). Price: $479.95.

This is an MSDOS product. It will also run under Windows in a DOS window [click here for MSDOS NOTES].

SPRING-STAT is designed to meet the data analysis needs of business, social science, market research, economics, financial, engineering, and academic professionals. While these needs have been met in the past by mainframe packages (and more recently by downscaled or expensive versions of these packages), SPRING-STAT offers some unique features:

O It's comfortable to learn and use. The entire system is prompt/menu driven and interactive.

O The system is integrated and modular. This means that you need buy only those modules that you want. It's a low-risk way to familiarize yourself with the system. If you like what you've bought, you can always add more modules.

O The 300+ page manual covers all aspects of the system. Sample data and on-screen instructions are also included.

O Data is stored in ASCII format which means that it's relatively easy to exchange data with other sources. The SCREEN EDITOR MODULE can also read DIF files.

The following provisions are available in modules (see below) for entering and editing data:

The DATABASE MODULE allows up to 10,000 records with up to 500 variables (fields) per record. You can design custom databases interactively and add records, delete records, change fields within records, search and sort records, and design custom report formats. You can also append files and create sub-files that include only those variables that you specify, and only those records that meet the conditions you indicate. In fact, you can use this database module for non-statistical applications (mailing lists, record catalogs, etc.) because all the data is stored in ASCII format.

The TRANSFORMATION MODULE allows you to create one or more new variables using a BASIC-like interpreter that supports arithmetic, logical (IF, THEN, ELSE, AND, OR, inequalities and function (SQR, SIN, COS, TAN, ATN, INT, LOG, EXP, ABS, NEG) commands. You can include up to 26 existing variables as well as numeric constants when creating the formula for computing a new variable.

The TRANSFORMATION MODULE also allows you to recode variables (alphanumeric codes can be changed to numeric codes; continuous variables can be changed to categorical values representing intervals), and to transform variables to normal or standardized scores.

A spreadsheet-like SCREEN EDITOR MODULE is available for handling smaller jobs. This editor can read and write files that are compatible with the DATABASE MODULE, read/write files in its own unique format, and read DIF files. This module can be used to create and edit the data in a more convenient fashion than when using the DATABASE MODULE. This module is the only data entry/edit method acceptable to the MULTIDIMENSIONAL SCALING AND HIERARCHICAL CLUSTERING MODULE.

1. DATABASE MODULE

O Menu driven. No commands to remember.

O Maximum of 10,000 records, 500 variables/fields, 50 characters per variable/field.

O Add records.

O Edit fields within records.

O Delete records meeting user specifications.

O Sort on any variable/field.

O Search records by record number or range of record numbers.

O Search on up to 8 variables simultaneously with relational specifications inequalities, = [includes wild card option], AND, OR).

O Custom design database and report formats.

O Display records on screen or printer. Regular and compressed mode available for printer (IBM printer or compatible).

O Merge two files with identical field specs.

O Create sub-files that include only user-specified variables and observations.

O Reads and writes ASCII files.

O Access to directory and many DOS commands from within program.

2. TRANSFORM MODULE

O Requires DATABASE MODULE. Same record and variable specifications as DATABASE MODULE.

O Menu/prompt driven.

O Create up to 10 new variables.

O BASIC-like interpreter supports following operators: ARITHMETIC (exponent, *, /, +, -) LOGICAL (IF, THEN, ELSE, AND, OR, inequalities) FUNCTION (SQR, SIN, COS, TAN, ATN, INT, LOG, EXP, ABS, NEG)

O Use up to 26 existing variables and/or numeric constants to create formula for calculating each new variable.

O Categorize metric variables or convert alphanumeric to numeric variables.

O Compute standard or normal scores.

O Compute terms for polynomial regression.

O Compute dummy variables or interaction (e.g., cross product) terms.

O Compute predictions or residuals from regression.

3. SCREEN EDITOR MODULE

O Capacity: rows X columns up to 10,000.

O Menu/prompt driven.

O Reads DATABASE MODULE files. Can write DATABASE MODULE files with automatic creation of field labels and field specs.

O Reads/writes files in its own format. (MDS & CLUSTERING MODULE can read only files in this format.)

O Reads DIF files.

O Four-way arrow control for accessing any row or column in data matrix with highlighted pointer.

O Labels for both observations and variables.

O Spreadsheet-like data entry/edit format.

O Add/delete/change any row, column or label.

O Compute new column with same BASIC-like interpreter that comes with TRANSFORM MODULE.

O Display matrix on both screen and printer.

4. ELEMENTARY STATISTICS (UNIVARIATE ANALYSIS) MODULE

O Requires DATABASE MODULE. Capacity: up to 500 variables.

O Sum of X, Sum of X-squared, Sum of X-cubed.

O Maximum, Minimum.

O Range, Mid-range.

O Mean, Variance, Standard Deviation, Standard Error of Mean.

O Coefficient of skewness, Coefficient of variation.

5. CONTINGENCY TABLES (BI-VARIATE ANALYSIS) MODULE

O Requires DATABASE MODULE and ELEMENTARY STATISTICS MODULE. Capacity: up to 500 variables.

O Crosstabs of either categorical variables or continuous variables within computed intervals.

O Row, column, and cell percentages and totals.

O Chi Square and Contingency Coefficient.

O Pearson's R, Bi-variate regression coefficients (where applicable).

O Row and column means, T-test on means (where applic.).

6. CORRELATION MATRIX MODULE

O Requires DATABASE MODULE. Capacity up to 120 variables.

O Means and Standard Deviations.

O Sums of squares and cross products.

O Corrected sums of squares and cross products (Dispersion Matrix).

O Correlation Matrix.

O Bartlett's Sphericity Test.

O Multiple R-squared of each variable with others.

7. STEPWISE MULTIPLE REGRESSION MODULE

O Requires DATABASE MODULE and CORRElATION MODULE. Capacity up to 120 variables.

O Multiple R-squared, Standard Error of Estimate at each step.

O ANOVA for Regression vs. Residual at each step.

O F to remove, F to enter, Beta Coefficient, Error of each step.

O Provision to either enter or remove a variable at each step using the Sweep algorithm.

8. FACTOR ANALYSIS MODULE

O Requires: DATABASE MODULE and CORRELATION MODULE. Capacity up to 120 variables.

O Principal Components using HOW algorithm for eigen-structure computations.

O Eigenvalue, % Trace, Wilk's Lambda, Bartlett's test for each successively extracted eigenvalue/root.

O Unrotated Factor Structure, Commonalities, Factor Score Coefficients.

O Varimax (orthogonal) rotation.

O Promax (oblique) rotation with Factor Structure, Factor Pattern, and Inter-factor correlations.

9. CANONICAL CORRELATIONS MODULE

O Requires DATABASE MODULE and CORRELATION MATRIX MODULE. Capacity up to 120 variables.

O Eigenvalues, Canonical R's, and Bartlett's test for each Canonical root.

O Canonical Coefficients for Dependent and Independent variables.

O Structure Correlations for Dependent and Independent variables.

O Stewart & Love Redundancy measures.

10. MULTIVARIATE ANALYSIS OF VARIANCE MODULE

O Requires DATABASE MODULE and CORRELATION MATRIX MODULE. Capacity: up to 40 variables.

O Group Means and Standard Deviations for each variable.

O Univariate F-ratios and Correlation ratios (Etas) for each variable.

O Group sample sizes and determinants for each group's correlation matrix.

O Test for equivalence of dispersions (Box's M).

O Tests for equivalence of centroids (Wilk's Lambda and Rao's F).

O Multivariate Correlation Ratio (Eta).

11. STEPWISE MULTIPLE DISCRIMINANT ANALYSIS MODULE

O Requires DATABASE, CORRELATION, and MULTIVARIATE ANALYSIS OF VARIANCE MODULES. Capacity: up to 120 variables.

O Wilk's Lambda and Chi Square test on Lambda at each step.

O F to remove, F to enter for each variable at each step.

O Group Discriminant Coefficients for each variable at each step.

O Provision to either enter or remove a variable at each step using the Sweep algorithm.

12. CANONICAL MULTIPLE DISCRIMINANT ANALYSIS MODULE

O Requires DATABASE, CORRELATION MATRIX, and MULTIVARIATE ANALYSIS OF VARIANCE MODULES. Capacity: up to 120 variables.

OSquare Test on Lambda for each canonical root extracted.

O Discriminant function coefficients for each variable on each canonical root.

O Structure correlations for each variable on each root.

O Group centroids for each group on each root.

13. MULTIDIMENSIONAL SCAlING AND HIERARCHICAL CLUSTERING MODULE

O Requires SCREEN EDITOR MODULE. Capacity: up to 80 objects and 50 variables.

O Means, Standard Deviations, Correlation Matrix.

O Preliminary Solution using Principal Components.

O Euclidean distance matrix between objects.

O Multidimensional Scaling Solution using KYST.

O Hierarchical Clustering of objects based on distance matrix.

O Graphics (perceptual map of objects and variables; Dendrogram of clustering process).

O Ability to handle both OBJECT x OBJECT and OBJECT x VARIABLE matrices.

14. SYSTEM MANUAL

O The 300-page plus manual is soid separately and is required in order to use any of the above modules.

15. ANOVA MODULE

O Requires DATABASE MODULE. Capacity: 1 to 5 factors.

O One-way, Two-way, Randomized, Split-plot, Latin Square, and Randomized complete block designs.

O Post Hoc Tests (LSD test, Tukey HSD test, SHEFFE test).

O Complete ANOVA summary tables.

16. TIME SERIES MODULE

O Requires DATABASE MODULE. Capacity: 400 time periods.

O Spectral Analysis.

O Correlogram.

O Digital Filtering.

O Curve Plotting.

O Generalized Exponential Smoothing with adaptive parameters.

O Residual Analysis (Durbin-Watson; Serial Correlation).

O Double Precision Multiple Regression.

17. LOGIT MODEL MODULE

O Requires SCREEN EDITOR MODULE. Capacity: 40 treatments.

O Weighted Maximum likelihood fit.

O Displays Actual Points vs. Theoretical Curve.

O Full Range of Probabilities and 95% Fiducial Limits.

O Chi Square goodness-of-fit estimate.

18. NON-PARAMETRIC STATISTICS MODUlE

O Wilcoxon Signed Ranks test.

O Sign Test.

O Kendall's Tau.

O Spearman's Rho.

O Friedman's 2-way ANOVA.

O Kendall's Coefficient of Concordance.

O Mann-Whitney U Test.

O Kolmogorov-Smirnov 2 Sample test.

O Kruskal-Wallis One-way ANOVA.

19. GRAPHICS MODULE

O Requires SCREEN EDITOR MODULE, DATABASE MODULE, and ELSTATS MODULE. Capacity: 40 observations for labeled graphs; up to 10,000 observations for other graphs.

O Labeled 3-D Mapping.

O Labeled 2-D Mapping with interactive control for modifications.

O Scatterplot with associated statistics.

O Frequency Plot with associated statistics.

O Box Plot.

O Stem and Leaf Diagram.

O Uses Alphanumerics for plotting

20. EXPERT SYSTEM MODULE

O Requires: No other modules are required. Capacity: More than 50 rules and 100 conditions in Knowledge Base.

O Cursor-controlled pop-up menu interface. Just point and shoot.

O Expert system matches users data-handling or analytical needs to SPRING-STAT's resources through simple yes/no interactive dialog.

O User can query system at any point to see why a particular question is being asked.

O System is logic-based. Knowledge base consists of conditions or attributes, and rules for associating these conditions.

O Uses backward chaining and automatic backtracking.

O Knowledge base can be modified/customized using text editor or word processor.

21. TEXT EDITOR MODULE

O Requires: No other modules are required. Capacity: 64K buffer.

O Read/write ASCII text files.

O Local & extended cursor movement. Insert/Delete commands. Block commands. Search/Replace. Combine 2 files.

O Command-driven, but menu operation and on-line help are available.

22. HIGH RESOLUTION GRAPHICS MODULE

O Requires graphics capability. Also requires SCREEN EDITOR and GRAPHICS (#19) modules. Supports HP LaserJet and DeskJet printers, as well as FX-85 and other compatible dot matrix printers.

O Reads .DPM files generated by SPRING-STAT and MARKET SIMULATOR.

O Produces 32 graph types: Three types of 3-D graphs; graphs with optional semi-log and log-log scaling; polynomial (1-6 degree) curve fit with error plots; labeled bar and pie charts.

O Can be accessed from CROSSTAB module to produce 3-D representations.

NEW SCREEN EDITOR MODULE ENHANCEMENTS

(Included with regular version-free.)

O New SCREEN EDITOR features: up to 50K cell capacity; GoTo command for immediate row/col access; read comma delimited, space delimited or fixed-length field ASCII files; write DIF files; compute transpose, ranks, random numbers, lags, row or column sums. indexes; sort entire matrix on selected variable/column; save sub-set of matrix.

O Path support.

O Access cases, variables, and files using light bar (point & shoot).

OTHER SYSTEM ENHANCEMENTS:

O DOS shell; visit DOS from within system. O Faster screen refresh.

O Greater accuracy. All data is stored in IEEE format with more significant digits.

O Improved error trapping and recovery.

EXCERPTS FROM A REVIEW (by Bruce Brown in PC Magazine):

"Overall this is a powerful statistical analysis program with a nice data-management facility, a powerful editor, and an expert system to help you decide what statistical procedure to use. In addition, it includes a wide selection of statistical manipulation and analysis tools. "[SPRING-STAT is] much easier to use, much more affordable, and about as powerful as some of the heaviest-hitting stat programs on the market."

"SPRING-STAT's data manager, a stand-alone flat-file manager, the text editor, which uses WordStar control- key commands, and the transform module, with its variable creation, recoding, and transforming powers, combine to give you great flexibility in setting up new data files or working on existing files."

"Besides using color, the program uses both horizontal and pull-down menus. While you can activate options on the horizontal menus with single-key commands or by moving the cursor to the option, pull-down menu selections require you to move the cursor to the desired choice."

"SPRING-STAT's expert system asks a series of questions requiring yes or no answers. Each session is termed a consultation, and each consultation concludes with a suggested statistical procedure based on your answers to the questions."

"You don't have to spend a lot of money buying unnecessary parts of the program. We worked with the full system, but you can purchase selected modules for $39.95 each, saving both money and disk storage space (the full system takes up just over 3 MB). This salad-bar approach to buying software is a boon, especially for those whose needs are defined and relatively fixed."

Review from PC Magazine:

"Overall, this is a powerful statistical analysis program with a nice data-management facility, a powerful editor, and an expert system to help you decide what statistical procedure to use. In addition, it includes awide selection of statistical manipulation and analysis tools.

"[SPRING-STAT is] much easier to use, much more affordable, and about as powerful as some of the heaviest- hitting stat programs on the market.

"SPRING-STAT's data manager, a stand-alone flat-file manager, the text editor, which uses WordStar control- key commands, and the transform module, with its variable creation, recoding, and transforming powers, combine to give you great flexibility in setting up new data files or working on existing files.

"Besides using color, the program uses both horizontal and pull-down menus. While you can activate options on the horizontal menus with single-key commands or by moving the cursor to the option, pull-down menu selections require you to move the cursor to the desired choice.

"SPRING-STAT's expert system asks a series of questions requiring yes or no answers. Each session is termed a consultation, and each consultation concludes with a the questions.

"You don't have to spend a lot of money buying unnecessary parts of the program. We worked with the full system of 20 modules, which costs $469.95, but you can buy selected modules for $39.95 each, saving both money and disk storage space (the full system takes up just over 3 MB). This salad-bar approach to buying software is a boon, especially for those whose needs are defined and relatively fixed.