MicrOsiris

MicrOsiris Statistical and Data Management Software System

MicrOsiris is a comprehensive statistical and data management package for Windows based microcomputer systems. It is derived from OSIRIS IV, a statistical and data management package developed and used at the University of Michigan.

Developed for serious survey analysis using moderate to large data sets, MicrOsiris's speed and minimal disk storage requirement, coupled with the interactive dialogs and on-line help systems, make it also suitable for college or graduate level courses in statistics and applied survey research methods. Further, the interactive statistical decision tree for choosing appropriate statistical techniques is a comprehensive decision tool---not just for using MicrOsiris.

MicrOsiris accepts as many cases as you can get on your system. There are interactive dialogs for managing MicrOsiris sessions and an on-line help facility for aid in running a particular command. SAS and SPSS files can be easily imported

Key Benefits

Handles any size data set.
Interactive dialogs and batched runfiles.
Data entry with Excel
Imports and exports to SAS, SPSS, STATA
Special Data Mining techniques for market analysis (SEARCH --very fast for large datasets).
Interactive statistical decision tree for selecting appropriate statistics.
Special nominal and ordinal statistical analysis techniques (MCA, MNA).
Can read OSIRIS data sets from the Institute for Political and Social Research (ICPSR).
Reads UNESCO IDAMS datasets.
Online manual
Decision Tree for choosing statistics and commands to compute them
Online context sensitive help system with additional links to manual.
Requires only 6MB when fully loaded; additional memory allocated as needed depending on number of variables used.
Less than 12MB on disk--including manual and Statistical Decision Tree.

Features

Statistics	Data Management	Commands
Requirements		Download

Data Management

Preparing Data for Input	General Purpose Utilities
Checking and Correcting Datasets	Transforming Datasets

MicrOsiris has the ability to handle weighted data, a powerful general purpose RECODE, commands to check and validate "wild codes" and perform consistency checking among numerous variables, and the ability to aggregate data to a new level of analysis. Commands and directions are provided for importing data into the system and for reformatting MicrOsiris datafiles for use with other systems, including spreadsheet, SAS, and database programs.

Statistics

Analysis of Variance and Hypothesis Testing	Factor and Cluster Analysis	Multivariate Analysis using Ordinal and Nominal Predictors
Frequency Distributions and Associated Statistics		Index Reliability and Survival Analysis

In addition to the usual basic statistics and functions (hypothesis testing, correlation analysis and classical regression--including stepwise and dummy variable), MicrOsiris provides multivariate analysis of nominal- and ordinal-scaled data. It can produce scatterplots, generate univariate and bivariate frequency distributions and related statistics (including lambdas, gammas, kappa, taus, chi-square, and Gini coefficients), non-parametric statistics and rank-order statistics, analysis of variance, and multiple classification analysis (multiple regression using categorical predictors). MicrOsiris also provides for factor analysis with varimax and oblimin rotations, and a variety of cluster analysis techniques.

Of special interest is SEARCH, a binary segmentation command used to develop predictive models for dependent variables. It is an elaboration of the OSIRIS AID and THAID programs developed at the Institute for Social Research. SEARCH divides a sample, through a series of binary splits, into mutually exclusive series of subgroups. The groups are chosen so that at each step the two new groups account for more of the variance or distribution than any other pair of subgroups. Predictor variables may be ordinal or nominally scaled. The dependent variable may be continuous or categorical.

Commands in MicrOsiris


*Preparing Data for Input*		Analysis of Variance and Hypothesis Testing
ADDDATA	Add records to a dataset	ANOVA	One-way analysis of variance.
DICTIONARY	Create and modify dictionaries describing datafiles.	TTEST	Means testing.
IMPORT	Import files from SPSS, SAS, STATA, and other systems.	MANOVA	Multivariate analysis of variance.
MATRANS	Transform or invert matrices.	Correlation and Regression Analysis
MATRIX	Matrix input routine.	LOGLINEAR	Dichotomous dependent variable regression analysis using linear or logit model.
MERGE	Combine two datasets.	CORRELATIONS	Pearsonian correlation analysis, allowing unequal n’s for each pair of variables.
*General-Purpose Utilities*		PARTIALS	Partial correlation analysis.
LISTDATA	List MicrOsiris datasets.	REGRESSSION	Multiple regression analysis and dummy variable regression.
LISTDICT	List MicrOsiris dictionaries.	Multivariate Analysis using Ordinal and Nominal Predictors
SORTDATA	Sort datasets.	MCA	Multiple classification analysis.
Checking and Correcting Datasets		MNA	Multivariate analysis of nominal data.
CONCHECK	Perform consistency checks between variables.	SEARCH	Binary segmentation.
FIXDATA	Correct individual variables case by case.	Factor and Cluster Analysis
		CAP	Spatial configuration analysis.
WILDCODE	Check and document invalid codes.	COMPARE	Compare configuration matrices.
Transforming Datasets		CLUSTER	Cluster analysis.
AGGREG	Aggregate individual records across subsets defined by the user. Compute summary descriptive statistics.	HICLUSTER	Hierarchical cluster analysis.
EXPORT	Export datasets as ASCII files for other systems such as SAS, SPSS, database and spreadsheet programs.	FACTAN	Factor analysis.
RECODE	Recode variables.	MINISSA	Guttman-Lingoes Smallest Space Analysis (Nonmetric multidimensional scaling).
TRANS	Reformat and translate datasets; produces permanently recoded datasets when used with RECODE.	Index Reliability and Survival Analysis
Frequency Distributions and Associated Statistics		ITAN	Item analysis.
SCATTERPLOT	Bivariate scatter plot diagrams	LIFETABLE	Life table analysis.
USTATS	Produces univariate statistics.	REL	Internal consistency of composite measures.
TABLES	Univariate and bivariate tables, rank-order statistics Chi-Square, Cochran's Q, Fisher, Friedman, Kendall's concordance, Kolmogorov-Smirnov two-sample, Kruskal-Wallis H, Spearman Rho, Kappa, Cramer’s V, Lambdas, Leik-Gove, Kendall’s Tau, Gamma, Mann-Whitney, Sign test, Wilcoxon).

Hardware Requirements

Windows XP or above. MicrOsiris is optimized for a Pentium processor or better, but does not require one. It can use as much memory as you have; permitting large numbers of 2- and 3-way cross tabulations, univariate statistics, and means testing to be run in one pass of the data.

Small footprint: Approximately 12MB on disk--including manual and Statistical Decision Tree.

MicrOsiris requires a variable amount of free workspace, depending upon the command used and size of job.