ZumaStat Has
Been Retired
Because of constant changes to the SPSS
programming architecture and difficulties of working with it and SPSS
consultants, ZumaStat has been retired and is no longer available. See the
support page for details, for how to unlock the password protection features
if you load the program onto a new computer (assuming you are an existing
user with a CD), for information if you recently ordered a copy, and for
future plans for the robust package.
ZumaStat
Statistical Programs
ZumaStat offers stand alone statistical
programs that can be integrated easily with the menu bars of SPSS and Excel.
The programs work with summary statistics (such as means, standard
deviations, percentages, and correlations) to perform a wide range of statistical tests
and power analyses (e.g., test of difference between two percentages, test of
difference between two correlations, single degree of freedom interaction
contrasts, t tests for means). ZumaStat also has a special SPSS
utility (called Z Plus) that makes using SPSS much easier and more flexible.
The ZumaStat programs have many useful features, including:
Z Plus for SPSS:
A special set of utilities to work
specifically with SPSS to make file manipulation and data manipulation
easier and more flexible. Create dummy variables and product terms
with the click of a few buttons. Interface with Excel graphics. Check out the SPSS Interface
section of this website to see just how much easier ZumaStat can make the
life of an SPSS user.
Emphasis on Confidence
Intervals: The ZumaStat
programs routinely report both significance tests and confidence intervals
for a wide range of statistical tests. Easily obtain
confidence intervals for percentages, correlations, means, standard
deviations, variance ratios, differences between
correlations, squared correlations, partial correlations, squared partial
correlations, squared multiple correlations, group differences in squared
multiple correlations, averages of correlations, percent of variance
accounted for statistics in ANOVA, single degree of freedom contrasts,
odds ratios, relative risks and a wide range of additional statistics.
Power Analysis and Analysis of Precision: ZumaStat offers programs to determine statistical
power. Most power programs conduct power on omnibus statistics
(e.g., an overall main effect in ANOVA or a multiple R in regression
analysis). However, rarely do researchers analyze data only at the
omnibus level. Instead, they do more focused tests, such as pairwise
contrasts or examining the significance of a regression coefficient
associated with a single predictor. A unique feature of ZumaStat is
that it does power analysis on these more focused tests. In
addition, ZumaStat offers a suite of utilities for determining optimal
sample sizes relative to minimizing margins of errors or the width of a
confidence interval. If you want to determine the sample size you
will need to have a confidence interval of a certain width, ZumaStat
provides perspectives on this. All of these programs are very simple
to use. ZumaStat offers an array of power programs that is far more
comprehensive than other power analysis software. For example, in addition
to standard statistical methods, ZumaStat offers power analysis utilities
for clustered designs (e.g., in educational settings when classrooms are
randomly assigned to treatment and control conditions). ZumaStat even offers
power analysis programs for growth curve models.
Meta-Analysis:
ZumaStat offers meta-analytic programs for analyzing raw means, d
statistics, correlations, percentage differences and odds ratios using
both fixed effects and random effects models. In addition, it
provides programs for conducting contrast analyses of different study
groupings to determine if the average effect size for one group of studies
differs from the average effect size for another group of studies.
ZumaStat provides utilities for weighted least squares
regression models predicting effect sizes from study characteristics that
are continuous or categorical in nature. ZumaStat also includes
programs to generate funnel plots and to explore issues of publication bias
using 'Trin and Fill' methods. Finally, ZumaStat offers
power analysis programs for meta-analysis so that one can determine if
non-significant results are adequately powered. The programs are easy to
use.
State-of-the-Art Missing Data Methods:
It is now commonly recognized that the traditional
strategies of pairwise and listwise deletion of missing data can be
suboptimal in many situations. An alternative approach based on Bayesian
and maximum likelihood imputations has evolved which is quite flexible and
has many desirable properties. However, the few available software programs
that implement these approaches are difficult to use. There is a free
computer program developed by researchers at Harvard that is reasonably
simple to use and that represents state-of-the-art imputation analysis. It
is called Amelia. ZumaStat has utilities to assist you in acquiring and
installing Amelia then interfacing it with SPSS. Amelia can be applied to
missing data scenarios in analysis of variance, OLS regression, logistic
regression, structural equation modeling and a wide range of other
statistical models. It can accommodate continuous, ordinal or nominal level
missing data. ZumaStat makes using Amelia much easier within the SPSS
environment.
Robust Statistics:
ZumaStat offers a special suite of programs for robust
statistical analyses based on the extensive functions written by Rand Wilcox
for the statistical package R. ZumaStat can be used form SPSS or Excel or
with an ascii file. It will pass data to R, execute the relevant Wilcox
function and then show you the output, all in a simple, intuitive,
point-and-click context. You do not need to know how to program R to use
these utilities. It is all reduced to simple point-and-click.
SEM Interface:
ZumaStat has utilities that will export data from SPSS
for input into LISREL or M Plus. For M Plus, ZumaStat also generates
syntax code for the initial parts if the program, so that all you need to do
is specify the model to be tested. Also, ZumaStat has utilities to make
running simulations on M Plus much easier.
Analysis of Change:
ZumaStat offers a suite of programs that provides
useful perspectives on the analysis of change between two time points.
Within Z Plus, it has a utility that will calculate six types of change
scores, (1) raw change, (2) the reliable change index (RCI) that is popular
in clinical psychology, (3) an ordinal version of the RCI, (4) a
residualized change score, (5) a backward residualized change score, and (6)
the Lord-McNemar true change score. ZumaStat interfaces with Excel to
generate Galton squeeze diagrams, a useful graphic for visualizing
regression to the mean. It also offers Pair-Link plots to better see
individual change. ZumaStat offers utilities to determine the reliability of
change scores from the reliability of its component parts, inherent
correlations between pretest and change scores by virtue of part-whole
dynamics, and the biasing effects of regression to the mean in intervention
studies where there are pre-existing differences on the pretest in the
treatment and control groups.
Requires
No Programming Skills: You
do not have to learn a programming language to
use ZumaStat programs. You simply type summary statistics into a
textbox (such as percentages and sample sizes) and press a button to get
the results.
Analyze Aggregate Level Statistics from Journal Reports:
ZumaStat allows you to input summary statistics typically reported in scientific
journals to replicate analyses, perform new analyses of your own, or to
calculate confidence intervals when the investigator failed to report
them.
Useful for Teaching Statistics: ZumaStat programs allow students to quickly see the results of
statistical tests when they vary such things as the sample size, the
standard deviation, the size of a correlation, and so on. By seeing
how these variations affect the significance tests and confidence
intervals, they gain a better intuitive feel for statistics.
Instructors can show demonstrations to students in the classroom using
ZumaStat with a projection screen. If students purchase the
programs, ZumaStat allows students to analyze data for interesting
examples with no need to learn a statistical computing package. All
the instructor needs to do is provide them with basic summary statistics.
Wide
Range of Statistical Tests:
The ZumaStat programs cover an extremely wide range of useful statistical
methods, even for seasoned data analysts. Take the time to look
through the links on Means and ANOVA, Frequencies, Regression, and Robust
Statistics.
Check
out the section on the List of Programs button to get a sense of the large
number of utilities that ZumaStat offers.
Types of Analyses
ZumaStat currently offers five categories
of statistical packages plus an SPSS utility package. A brief overview
of these programs follows:
 |
SPSS
Interface: ZumaStat offers a package of utilities that are
designed to enhance and simplify the data manipulation capabilities of SPSS.
ZumaStat allows you to create dummy variables and product terms with just a
few simple clicks on the mouse. It creates special dummy variables for
exploring bias due to missing data and can reorganize data sets for growth
curve modeling in other programs (such as HLM). ZumaStat easily
creates mean centered versions of variables (as well as versions that are
one standard deviation above and below the mean) and automates this
cumbersome process. It also mean centers data with a break variable
(or by strata), ZumaStat reformats large correlation matrices so
that you can see all of the entries of the matrix on a single screen.
ZumaStat provides simple point and click methods for saving subsets of
variables to a file, exporting data, and saving cases that meet the
requirements of the active case filter. It will also easily allow you
to drop variables from the active data file and reorder variables in an
active data file. ZumaStat will eliminate
missing data from a file based on multivariate patterns across many
variables. It provides extremely simple methods for adding or
subtracting constants from many variables, exchanging values on a variable
(e.g., making a 1 a zero and a zero a 1),
grouping scores on a variable, copying labels between variables, reverse scoring a variable,
calculating the number of days between two date variables and changing
column widths for large numbers of variables. ZumaStat
integrates SPSS with the graphing capabilities of Excel, allowing you to
easily pass data to Excel while automatically formatting a mean plot, a
factorial plot of means, frequency plots, plots of two way contingency
tables, scatterplots and smoothed scatterplots. It also permits the
easy addition of error bars. ZumaStat allows you to change the number
of decimals in a pivot table with one or two clicks as well as convert
scientific notation to standard notation.
ZumaStat also has utilities that facilitate the application of several
modified Bonferroni methods for controlling experimentwise error rates.
It has several routines for generating random numbers, an extensive package
for generating simulation data, and for analyzing binomial and normal distributions.
ZumaStat will randomly split data files into any number of groups for
purposes of cross-validation analysis. ZumaStat also allows you to apply
many features of the SPSS matrix language using simple point and click
instead of syntax. Finally, ZumaStat has a suite of programs for analyzing
change, including programs to provide you with perspectives on regression to
the mean. Browse the page on 'SPSS
Interface' for details.
|
|
Means and ANOVA:
This is a set of utilities that focus on the analysis of means using t
tests and analysis of variance frameworks. ZumaStat provides a wide
range of methods that facilitate follow-up analyses and that perform
multiple comparisons, single degree of freedom contrasts, interaction
contrasts, simple main effects analysis, effect size estimation,
equivalence tests, meta-analysis of raw means and d statistics (for both fixed effect
and random effect models) and power and precision analysis, including
power analysis for meta-analytic methods as well as clustered designs. ZumaStat has graphic features that
simplify the generation of factorial plots in Excel from SPSS.
ZumaStat also has utilities that facilitate the application of several
modified Bonferroni methods for controlling experimentwise error rates.
It includes techniques for analyzing outlier resistant measures of central
tendency, such as trimmed means. ZumaStat can generate an analysis of variance summary table from the typically
reported means, sample sizes and F ratio in a journal article.
Browse the page on 'Means and ANOVA' for details on the many programs in
this suite. |
|
Regression: This is a set of utilities
that focus on the analysis of correlation and regression. ZumaStat
provides methods for calculating confidence intervals for correlations,
squared correlations, squared multiple correlations and partial
correlations. ZumaStat performs significance tests and calculates
confidence intervals for tests of differences in correlations, differences
in squared multiple correlations, and differences in partial correlations.
ZumaStat offers a wide range of correlation and regression based programs
for power analysis and precision analysis
including power analysis for meta-analytic methods as well as clustered
designs. ZumaStat performs meta-analyses of correlations using both fixed effects
and random effects models. It performs equivalence tests for a wide
range of correlational statistics. ZumaStat provides a program for
computing all possible regression equations for up to 10 predictors, as an
alternative to stepwise regression. ZumaStat gives you r to Z
transforms and provides numerous utilities for the effective analysis of
interacti on terms and polynomial terms in multiple regression models.
ZumaStat facilitates the generation of scatterplots in Excel from SPSS.
ZumaStat allows you to enter a regression equation
and then calculate predicted values for different predictor profiles.
ZumaStat also has utilities that facilitate the application of several
modified Bonferroni methods for controlling experimentwise error rates.
ZumaStat does mediation analysis and has several programs for intricate
interaction analysis. It calculates correlations
corrected for attenuation. See the page on 'Regression' for details of the
many programs in this suite. |
|
Frequencies: This is a set of utilities focused on the analysis
of frequencies, percentages, relative risks, and odds ratios vis-a-vis
logistic regression. ZumaStat provides methods for calculating
confidence intervals and significance tests for percentages, percentage
differences, multiple comparisons of percentages and interaction analysis
of percentages. ZumaStat performs significance tests and calculates
confidence intervals for a relative risk, for odds ratios, and for
interaction analyses of odds ratios. ZumaStat performs meta-analyses
of both percentage differences and odds ratios using both fixed effects
and random effects. It performs equivalence tests for a wide
range of frequency based statistics. ZumaStat permits you to
calculate numerous values of Pseudo R squared in logistic regression and
has a descriptive and graphical extension of the Hosmer-Lemeshow fit
analysis to assist evaluation of model fit. ZumaStat provides an
"odds calculator" that converts log odds to odds and probabilities and
vice versa. ZumaStat allows you to enter a logistic regression
equation and then calculate predicted values for different predictor
profiles. ZumaStat also does power
and precision analysis for tests of percentages and logistic regression
problems including power analysis
for meta-analytic methods as well as clustered designs.
You can do a chi square test of independence for an R X C contingency
table where you enter the summary frequencies. ZumaStat also calculates
confidence intervals for Cramer's V. Browse the page on
'Frequencies' for details. |
|
Miscellaneous Utilities: ZumaStat
provides utilities for probability distributions that allow you to easily
calculate a p value or a critical value for a t distribution, a binomial
distribution, an F distribution, a chi square distribution and a normal
distribution. It includes utilties for the studentized range statistics.
ZumaStat also has utilities that facilitate the application of several
modified Bonferroni methods for controlling experimentwise error rates. It
calculates factorials, combinations and permutations. It estimates
regression to the mean effects in clinical trials. ZumaStat calculates the
reliability of a difference score from the reliability of its component
parts, calculates confidence intervals for an alpha coefficient, and will
provide perspectives on the number of items you need on a scale to achieve
an alpha coefficient of a pre=specified value. It calculates correlations
adjusted for unreliability. ZumaStat has easy to use random number
generation utilities that make it easy to randomly assign individuals to
groups and that permit you to generate random numbers in a specified range
either with or without replacement. ZumaStat will calculate confidence
intervals for a standard deviation and a variance ratio. It also links you
to useful statistical websites.
ZumaStat includes matrix utilities that allow you to compute the the
determinant, inverse, trace, rank or transpose of a matrix. You can also
add, subtract, multiply and divide two matrices. Finally, ZumaStat offers
a utility to calculate the reliable change index, a popular index of
clinical significance in clinical psychology and counseling. Browse the page on
'Miscellaneous Utilities' for details. |
|
Robust
Statistics: ZumaStat offers a wide range of modern analytic
methods based on robust statistical tests. These programs must be
used in conjunction with the computer program R as ZumaStat relies on
functions written for R. Using many of the powerful functions
written by Professor Rand Wilcox, ZumaStat offers an intuitive interface
for conducting a wide range of tests on robust measures of location, such
as the trimmed mean, the median, and M estimators. It performs
significance tests and calculates confidence intervals for
comparing two groups (for both independent groups and dependent groups),
three or more groups (corresponding to one way analysis of variance and
multiple comparisons) and you can conduct single degree of freedom
contrasts relevant to a wide range of factorial designs. ZumaStat
calculates robust estimates of variability, including the MAD statistic
and the interquartile range, among others. It provides for a comprehensive analysis of quantiles,
including quantile regression. ZumaStat offers several robust measures of correlation
and a wide variety of robust regression methods. It also provides
robust analysis of covariance algorithms. as well as extensive outlier
analysis ZumaStat offers access to almost all of the
functions in the Wilcox (2005) book. Browse the page on
'Robust Statistics' for more details.
|
