home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Shareware Overload Trio 2
/
Shareware Overload Trio Volume 2 (Chestnut CD-ROM).ISO
/
dir28
/
st-size.zip
/
TOOLS.DOC
< prev
Wrap
Text File
|
1993-01-03
|
13KB
|
217 lines
STATOOLS(tm)
STATOOLS(tm) are stand-alone programs designed to fill in
some of the gaps left by major statistical packages. They are
available on 5-1/4" disks for the IBM PC and compatibles
running DOS 2.0 or later versions. The disks contain
executable files and user's guides. FORTRAN source code is
included where indicated. The programs may be obtained from
Gerard E. Dallal
54 High Plain Road
Andover, MA 01810 USA
Please do not send purchase orders.
The PC-SIZE disk contains two programs: 'PC-SIZE' and
'PC-SIZE: Consultant'. PC-SIZE was written in 1985 for people
trained in the techniques of sample size determination who
wanted a program to automate the calculations. Consultant,
written in late 1988, was intended to be a more user-friendly
successor to the original PC-SIZE. Because some users have
found it useful to have both programs, they are now
distributed as a set. ($15; instructors may duplicate a
registered copy for a user's fee of $5 per student)
PC-SIZE: Consultant is a prompt driven program that
determines the sample size requirement for comparing
population means from independent or paired samples and
population proportions from independent samples. Mean
differences can be stated in absolute or relative (percent)
terms. Variability can be stated in terms of standard
deviations or intervals likely to contain most of the data.
PC-SIZE: Consultant can calculate the power of specific
sample sizes and determine the sample size needed to
achieve specific power. It can also calculate the sample
size needed to insure, with a specified amount of
probability, that a confidence interval for a single
population mean or for the difference between two
population means does not exceed a specified length. The
program can generate written reports for inclusion in
manuscripts and proposals. PC-SIZE: Consultant is suitable
for the occasional user with limited statistical training.
PC-SIZE determines the sample size requirements for single
factor experiments, two factor experiments, randomized
blocks designs, and paired t-tests. In generic mode, PC-
SIZE can determine sample sizes for any experiment in which
the power at the alternative is given by a non-central F
distribution with fixed numerator degrees of freedom,
denominator degrees of freedom that are linear in the
sample size, and a non-centrality parameter that is
proportional to the sample size. PC-SIZE can determine the
sample size needed to detect a non-zero population
correlation coefficient when sampling from a bivariate
normal distribution. It can also be used to obtain the
common sample size required to test the equality of two
proportions. PC-SIZE can calculate the power of specific
sample sizes as well as determine the sample size needed to
achieve specific power.
PC-PLAN generates randomization plans involving up to 20
treatments and 400 subjects. A plan can include a blocking
variable (such as sex) that is not subject to randomization.
When an equal number of subjects receives each treatment, the
randomization may be carried out within subblocks to insure
against serious imbalance should the study be terminated
prematurely. PC-PLAN can construct plans in which the order
of up to 5 treatments is randomized for each subject. ($10
for a disk containing PC-PLAN, RANDOM, and PC-EMS. Source
code for PC-PLAN is included.)
RANDOM generates lists of random integers. It can produce
integers from the set 0(1)N or 1(1)N with replacement, as well
as many random permutations of K multiples of 0(1)N or 1(1)N.
($10 for a disk containing PC-PLAN, RANDOM, and PC-EMS.
Source code for RANDOM is included.)
PC-EMS produces tables of expected mean squares for balanced
experiments by using the Cornfield-Tukey algorithm. ($10 for
a disk containing PC-PLAN, RANDOM, and PC-EMS.)
PC-AIP fits additive-in-the-probits models to two-dimensional
contingency tables with ordered column classifications. ($10
for a disk containing PC-AIP and STAT-SAK. Source code for
both programs included.)
STAT-SAK is the STATistician's Swiss Army Knife. While not
the custom tool for any particular job, it carries out a wide
variety of miscellaneous tasks that are not easily performed
by most large statistical packages. These include probability
distributions (normal, t, chi-square, F, binomial, Poisson,
studentized range); exact confidence intervals for binomial
p's; Fisher's exact test for 2 by 2 tables; exact confidence
intervals for an odds ratio from a 2 by 2 table; Mantel-
Haenszel test and approximate 95% confidence interval for the
common odds ratio; Bartholomew's test for increasing
proportions; Bartholomew's test for increasing normal means;
one and two sample t-tests from summary statistics;
correlation coefficients: test that a population value is 0,
confidence interval for a single coefficient, test of equality
of two independent coefficients, test of equality of two
dependent coefficients (coefficients from the same correlation
matrix). ($10 for a disk containing PC-AIP and STAT-SAK.
Source code for both programs included.)
OCTA (One man's Contingency Table Analysis): Interactive log-
linear analysis of multi-dimensional contingency tables by use
of the Deming-Stephan iterative proportional fitting
procedure. Quasi-independence models; Brown's tests of
partial and multiple association. Tables of up to 1000 cells
containing no more than 7 dimensions and 20 categories per
dimension. ($10)
RCJOIN identifies sources of interaction in two-way tables of
counts (Computers and Biomedical Research, 21(1988), 129-136)
($5; source code included)
FMT: FORTRAN FORMAT Statement Generator. Use a word
processor or text editor to create a text file looking like
the output wanted from a FORTRAN program. ^'s specify
numerical fields, ~'s specify character fields. FMT turns the
text into FORTRAN FORMAT statements. ($7)
--------------------------------------------------------------
The following programs work with MYSTAT/SYSTAT system files.
They can be used independently of MYSTAT and SYSTAT by using
an accompanying utility, a2s, to convert ASCII files of
numerical data to MYSTAT/SYSTAT system files.
cLOGISTIC: Conditional logistic regression for stratified
case-control data. Up to 500 strata with no more than 10
cases plus controls per stratum. The product of the number of
subjects and the number of variables must be no greater than
6000. (Accompanies LOGISTIC at no additional charge.)
CONTENTS displays the contents--variable names and number of
records--in MYSTAT/SYSTAT files. Can be run from the DOS
prompt. ($5)
LMSMVE: Least median of squares regression and robust
distances from minimum volume ellipsoids. Built around the
programs PROGRESS by Rousseeuw and Leroy and MINVOL by
Rousseeuw and van Zomeren. Includes high resolution residual
plots. ($10)
LOGISTIC: Logistic regression. Maximum likelihood estimates
of the regression coefficients are obtained by using a Newton-
Raphson iterative procedure. Interactions may be specified in
the model. The model may contain up to 20 terms. ($10)
MUDIFT: Multivariate distribution-free comparison of growth
curves. (Guo S, Simon R, Talmadge JE, Klabansky RL. Computers
and Biomedical Research, 20(1987), 37-48.) MUDIFT was written
with Guo, Simon, Talmadge, and Klabansky. ($5)
ODDJOB: Miscellaneous statistical procedures, including tests
of normality (Lilliefors, Wilk-Shapiro, D'Agostino-Pearson);
the Hartigans' DIP test of unimodality; isotone regression;
two-phase regression; structural relations; Breiman and
Friedman's alternating conditional expectations (ACE) for
estimating optimal transformations for multiple regression;
Cleveland's locally weighted scatterplot smoother (LOWESS);
test of equality of two dependent correlation coefficients
(coefficients from the same correlation matrix); Johnson-
Neyman analysis of heterogeneous regression slopes in two
group, one covariate ANCOVA; comparing means from incomplete
paired data. ($10)
PARTIAL: Partial correlation coefficients, partial regression
residual (leverage) plots, and partial regression plots
(Larsen-McCleary). ($5)
PCT: Percentiles. Percentiles may be obtained within groups
defined by a designated grouping variable. All five
definitions of percentiles given in SAS's PROC UNIVARIATE are
available. ($5)
PITMAN: calculates observed significance levels by using
recursive relationships to obtain a randomization
(permutation) distribution without having to directly
examining all possible permutations of the data. It performs
one and two sample randomization tests and rank tests in the
presence of an arbitrary number of ties in the data. ($10;
source code included)
TRACK: Foulkes-Davis tracking index, the probability that two
individuals chosen at random will have measurement curves over
time that do not cross. (Foulkes and Davis. Biometrics,
37(1981), 439-446) Fits linear, quadratic, or cubic
polynomials to each individual's data and judges the adequacy
of the fit. Produces a command file that allows SYGRAPH
(Systat, Inc.) to draw the measurements curves. ($5; source
code included)
BUMP: Dose-response curves in the presence of a downturn.
(Simpson and Margolin, Biometrika, 73(1986), 589-596). BUMP
was written with Douglas G. Simpson and is available ONLY FROM
HIM at Department of Statistics, University of Illinois at
Urbana-Champaign, 101 Illini Hall, 725 South Wright Street,
Champaign, IL 61820. (SEND TO SIMPSON: $5 OR formatted 5-
1/4" floppy diskette in a self-addressed stamped mailer;
source code included.)
GENERIC: FORTRAN source code that can be used as a template
for writing programs that analyze data in MYSTAT/SYSTAT system
files. ($5)
--------------------------------------------------------------
LOGISTIC and (c)LOGISTIC are now available in a version that
operates directly on Epi Info files ($10). When ordering,
please specify the SYSTAT/MYSTAT version or the Epi Info
version. If neither is specified, the SYSTAT/MYSTAT version
will be sent.