whatsnew6to7.hlp

是一个经济学管理应用软件 很难找的 但是经济学学生又必须用到

{p 4 4}
{cmd:xtregar} estimates cross-sectional time-series models in which epsilon_it
is assumed to follow an AR(1) process.  {cmd:xtregar} reports the within
estimator and a GLS random-effects estimator.  {cmd:xtregar} can handle
unequally spaced observations and exogenously unbalanced panels.
{cmd:xtregar} uniquely reports the modified Bhargava et al. Durbin--Watson
statistic and the Baltagi--Wu locally best invariant test statistic for
autocorrelation.  See help {help xtregar}.

{p 4 4}
{cmd:xtivreg} estimates cross-sectional time-series regressions with
(generalized) instrumental variables, or, said differently, estimates
two-stage least squares time-series cross-sectional models.  {cmd:xtivreg} can
estimate such models using the between-2SLS estimator, the within-2SLS
estimator, the first-differenced 2SLS estimator, the
Balestra--Varadharajan--Krishnakumar G2SLS estimator, or the Baltagi EC2SLS
estimator.  All the estimators allow use of balanced or (exogenously)
unbalanced panels.  See help {help xtivreg}.

{p 4 4}
{cmd:xtpcse} produces panel-corrected standard errors (PCSE) for linear
cross-sectional time-series models where the parameters are estimated by OLS
or Prais--Winsten regression.  When computing the standard errors and the
variance--covariance estimates, the disturbances are, by default, assumed to
be heteroskedastic and contemporaneously correlated across panels.  See help
{help xtpcse}.


    {title:Survival analysis (st)}

{p 4 4}
{cmd:stcox} will now estimate proportional hazard models with continuously
time-varying covariates, and you do not need to modify your data to obtain the
estimates.  See the {cmd:tvc()} and {cmd:texp()} options in help {help stcox}.

{p 4 4}
{cmd:streg} can now estimate parametric survival models with individual-level
frailty (unobserved heterogeneity).  Two forms of the frailty distribution are
allowed:  gamma and inverse gaussian.  Frailty is allowed with all the
parametric distributions currently available.  See help {help streg}.  (New
commands {cmd:weibullhet}, {cmd:ereghet}, etc., allow users to estimate these
models outside of the st system; see help {help weibull}.)

{p 4 4}
{cmd:streg} has also been modified to allow estimation of stratified models,
meaning that the distributional parameters (the ancillary parameters and
intercept) are allowed to differ across strata.  See the {cmd:strata()} option
in help {help streg}.

{p 4 4}
{cmd:streg} has also been modified to allow you to specify any
linear-in-the-parameters equation for any of the distributional parameters,
which allows you to create various forms of stratification, as well as
allowing distributional parameters to be linear functions of other covariates.
See the {cmd:ancillary()} option in help {help streg}.

{p 4 4}
{cmd:stptime} calculates person-time (person-years) and incidence rates and
implements computation of the standardized mortality/morbidity ratios (SMR).
See help {help stptime}.

{p 4 4}
{cmd:sts test} has been modified to include additional tests for comparing
survivor distributions, including the Tarone--Ware test, the
Fleming--Harrington test, and the Peto--Peto--Prentice test.  Also new is a
test for trend.  See help {help sts}.

{p 4 4}
{cmd:stci} calculates and reports the level and confidence intervals of the
survivor function, as well as computing and reporting the mean survival time
and confidence interval.  See help {help stci}.

{p 4 4}
{cmd:stsplit} is now much faster and now allows for splitting on failure
times, as well as providing some additional convenience options.  See help
{help stsplit}, but remember that {cmd:stcox} can now estimate with continuous
time-varying covariates without you having to {cmd:stsplit} the data
beforehand.

{p 4 4}
{cmd:stcurve} has a new {cmd:outfile} option.  See help {help streg}.


    {title:Commands for epidemiologists}

{p 4 4}
Five new commands are provided for the analysis of Receiver Operating
Characteristic (ROC) curves.

{p 4 4}
{cmd:roctab} is used to perform nonparametric ROC analyses.  By default,
{cmd:roctab} calculates the area under the curve.  Optionally, {cmd:roctab}
can plot the ROC curve, display the data in tabular form, and produce
Lorenz-like plots.  See help {help roctab}.

{p 4 4}
{cmd:rocfit} estimates maximum-likelihood ROC models assuming a binormal
distribution of the latent variable.  {cmd:rocplot} may be used after
{cmd:rocfit} to plot the fitted ROC curve and simultaneous confidence bands.
See help {help rocfit}.

{p 4 4}
{cmd:roccomp} tests the equality of two or more ROC areas obtained from
applying two or more test modalities to the same sample or to independent
samples.  See help {help roccomp}.

{p 4 4}
{cmd:rocgold} independently tests the equality of the ROC area of each of
several test modalities against a "gold" standard ROC curve.  For each
comparison, {cmd:rocgold} reports the raw and the Bonferroni adjusted
significance probability.  Optionally, Sidak's adjustment for multiple
comparisons can be obtained.  See help {help rocgold}

{p 4 4}
{cmd:binreg} estimates generalized linear models for the binomial family and
various links.  It may be used with either individual-level or grouped data.
Each of the link functions offers a distinct, epidemiological interpretation
of the estimated parameters.  See help {help binreg}.

{p 4 4}
{cmd:cc} and {cmd:cci} now, by default, compute exact confidence intervals
    for the odds ratio.  See help {help cc}.

{p 4 4}
{cmd:icd9} and {cmd:icd9p} assist when you are working with ICD-9-CM
diagnostic and procedure codes.  These commands allow the cleaning up,
verification, labeling, and selection of ICD-9 values.  See help {help icd9}.


    {title:Marginal effects}

{p 4 4}
{cmd:mfx} reports marginal effects after estimation of any model.  Marginal
effects refers to df()/dx_i evaluated at x, where f() is any function of the
data and the model's estimated parameters, x are the model's covariates, and
x_i is one of the covariates.  For instance, the model might be probit and f()
the cumulative normal distribution, in which case df()/dx_i = the change in
the probability of a positive outcome with respect to a change in one of the
covariates.  x might be specified as the mean, so that the change would be
evaluated at the mean.

{p 4 4}
{cmd:dprobit} would already do that for the probit model, and there have been
other commands published in the STB that would do this for other particular
models, such as {cmd:dtobit} for performing tobit estimation.

{p 4 4}
{cmd:mfx} works after estimation of any model in Stata and is capable of
producing marginal effects for anything {cmd:predict} can produce.  For
instance, after {cmd:tobit}, you could get the marginal effect of the
probability of an outcome being uncensored, or the expected value of the
uncensored outcome, or the expected value of the censored outcome.

{p 4 4}
{cmd:mfx} can compute results as derivatives or elasticities.  See help
{help mfx}


    {title:Cluster analysis}

{p 4 4}
{cmd:cluster} performs partitioning and hierarchical cluster analysis using a
variety of methods.  Two partitioning cluster methods are provided -- kmeans
and kmedians -- and three hierarchical-cluster methods are provided -- single
linkage, average linkage, and complete linkage.  Included are 14 binary
similarity measures and 7 different continuous measures (counting things such
as the Minkowski distance {it:#} as one).

{p 4 4}
The result is to add various characteristics to the dataset, including
variables reflecting cluster membership.  {cmd:cluster} can then can display
results in various ways.

{p 4 4}
More than one result can be saved simultaneously, so that the results of
different analyses may be compared.  {cmd:cluster} allows adding notes to
analyses and, of course, the dropping of analyses.  {cmd:cluster} also
provides post-clustering commands that can, for instance, display the
dendrogram (clustering tree) from a hierarchical analysis or produce new
grouping variables based on the analysis.

{p 4 4}
{cmd:cluster} has been designed to be extended.  Users may program extensions
for new cluster methods, new cluster management routines, and new
post-analysis summary methods.

{p 4 4}
See help {help cluster} and, if you are interested in programming extensions,
see help {help clprog}.


    {title:Pharmacokinetics}

{p 4 4}
There are four new estimation commands and two new utilities intended for the
analysis of pharmacokinetic data; see help {help pk}.

{p 4 4}
{cmd:pkexamine} calculates pharmacokinetic measures from
time-and-concentration subject-level data. {cmd:pkexamine} computes and
displays the maximum measured concentration, the time at the maximum measured
concentration, the time of the last measurement, the elimination rate, the
half-life, and the area under the concentration-time curve (AUC).  See help
{help pkexamine}.

{p 4 4}
{cmd:pksumm} obtains the first four moments from the empirical distribution of
each pharmacokinetic measurement and tests the null hypothesis that the
measurement is normally distributed.  See help {help pksumm}.

{p 4 4}
{cmd:pkcross} analyzes data from a crossover design experiment.  When
analyzing pharmaceutical trial data, if the treatment, carryover, and sequence
variables are known, the omnibus test for separability of the treatment and
carryover effects is calculated.  See help {help pkcross}.

{p 4 4}
{cmd:pkequiv} performs bioequivalence testing for two treatments.  By default,
{cmd:pkequiv} calculates a standard confidence interval symmetric about the
difference between the two treatment means.  Optionally, {cmd:pkequiv}
calculates confidence intervals symmetric about zero and intervals based on
Fieller's theorem.  Additionally, {cmd:pkequiv} can perform interval
hypothesis tests for bioequivalence.  See help {help pkequiv}.

{p 4 4}
{cmd:pkshape} and {cmd:pkcollapse} help in reshaping the data into the form
that the above commands need; see help {help pkshape} and {help pkcollapse}.


    {title:Other statistical commands}

{p 4 4}
{cmd:jknife} performs jackknife estimation, which is (1) an alternative,
first-order unbiased estimator for a statistic; (2) a data-dependent way to
calculate the standard error of the statistic and to obtain significance
levels and confidence intervals; and (3) a way of producing measures
reflecting the observation's influence on the overall statistic.  See help
{help jknife}.

{p 4 4}
{cmd:lfit}, {cmd:lroc}, {cmd:lsens}, and {cmd:lstat} now work after
{cmd:probit} just as they do after {cmd:logit} or {cmd:logistic}.

{p 4 4}
{cmd:drawnorm} draws random samples from a multivariate normal distribution
with specified means and covariance matrix.  See help {help drawnorm}.

{p 4 4}
{cmd:corr2data} creates fictional datasets with the specified means and
covariance matrix (correlation structure).  Thus, you can take published
results and duplicate and modify them if the estimator is solely a function of
the first two moments of the data, such as {cmd:regress}, {cmd:ivreg},
{cmd:anova}, or {cmd:factor}.  See help {help corr2data}.

{p 4 4}
{cmd:median} performs a nonparametric test that K samples were drawn from
populations with the same median.  See help {help median}.

{p 4 4}
{cmd:tabstat} displays tables of summary statistics, possibly broken down
(conditioned) on another variable.  See help {help tabstat}.

{p 4 4}
The command {cmd:avplot} now works after estimation using the {cmd:robust} or
{cmd:cluster()} options.  See help {help avplot}.

{p 4 4}
{cmd:ml} can now perform estimation with linear constraints.  All that is
required is that you specify the {cmd:constraint()} option on the {cmd:ml}
{cmd:maximize} command.  See help {help ml}.


    {title:Distribution functions}

{p 4 4}
Stata's density and distribution functions have been renamed.  First, all the
old names continue to work, even when not documented in the manual, at least
under version control.  The new standard, however, is, if {it:X} is the name
of a distribution, then

{p 8 26}{it:X}{cmd:den()}{space 8}is its density{p_end}
{p 8 26}{it:X}{cmd:()}{space 11}is its cumulative distribution{p_end}
{p 8 26}{cmd:inv}{it:X}{cmd:()}{space 8}is its inverse cumulative{p_end}
{p 8 26}{it:X}{cmd:tail()}{space 7}is its reverse cumulative{p_end}
{p 8 26}{cmd:inv}{it:X}{cmd:tail()}{space 4}is its inverse reverse
	cumulative{p_end}

{p 4 4}
Not all functions necessarily exist and, if they do not, that is not solely
due to laziness on our part.  In particular, concerning the choice between
{it:X}{cmd:()} and {it:X}{cmd:tail()}, the functions exist that we have
accurately implemented.  In theory, you only need one because
{bind:{it:X}{cmd:tail()} = 1 - {it:X}{cmd:()}}, but in practice, the one-minus
subtraction wipes out lots of accuracy.  If one really wants an accurate
right-tail or left-tail probability, one needs a separately written
{it:X}{cmd:tail()} or {it:X}{cmd:()} routine, written from the ground up.

{p 4 4}
Anyway, forget everything you ever knew about Stata's distribution functions.
Here is the new set:

{p 8 31}{cmd:normden()}{space 8}same as old {cmd:normd()}{p_end}
{p 8 31}{cmd:norm()}{space 11}same as old {cmd:normprob()}{p_end}
{p 8 31}{cmd:invnorm()}{space 8}same as old {cmd:invnorm()}{p_end}

{p 8 31}{cmd:chi2()}{space 11}related to old {cmd:chiprob()}; see below{p_end}
{p 8 31}{cmd:invchi2()}{space 8}related to old {cmd:invchi()}; see below{p_end}
{p 8 31}{cmd:chi2tail()}{space 7}related to old {cmd:chiprob()}{p_end}
{p 8 31}{cmd:invchi2tail()}{space 4}related to old {cmd:invchi()}{p_end}

{p 8 31}{cmd:F()}{space 14}related to old {cmd:fprob()}{p_end}
{p 8 31}{cmd:invF()}{space 11}related to old {cmd:invfprob()}{p_end}
{p 8 31}{cmd:Ftail()}{space 10}same as old {cmd:fprob()}{p_end}
{p 8 31}{cmd:invFtail()}{space 7}equal to old {cmd:invfprob()}{p_end}

{p 8 31}{cmd:ttail()}{space 10}related to old {cmd:tprob()}; see below{p_end}
{p 8 31}{cmd:invttail()}{space 7}related to old {cmd:invt()}; see below{p_end}

{p 8 31}{cmd:nchi2()}{space 10}equal to old {cmd:nchi()}{p_end}
{p 8 31}{cmd:invnchi2()}{space 7}equal to old {cmd:invnchi()}{p_end}
{p 8 31}{cmd:npnchi2()}{space 8}equal to old {cmd:npnchi()}{p_end}

{p 4 4}
We want to emphasize that if a function exists, it is calculated accurately.
To wit, {cmd:F()} accurately calculates left tails, and {cmd:Ftail()}
accurately calculates right tails; {cmd:Ftail()} is far more accurate than
{bind:1 - {cmd:F()}}.

{p 4 4}
There is no {cmd:normtail()} function.  The accurate way to calculate
left-tail probabilities (z<0) is {cmd:norm(z)}.  The accurate way to calculate
right-tail probabilities (z>0) is {cmd:norm(-z)}.