Jon Peck:Important SPSS Updates

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

For many years I have been interested in statistical selection problems.  This has produced the FUZZY extension command for case-control  matching (also PSM), the SPSSINC TOBIT REGR command for Tobit regression models, and the STATS HECKMAN REGR command for Heckman regression models.  Now I have uploaded the STATS RDD module for analysis of regression discontinuity designs, which have become popular in recent years.  STATS RDD estimates the effects of a treatment on an output in a nonexperimental context.  The treatment is applied based on whether an observed assignment variable (also known as a running variable) exceeds a known threshold.  There are many published examples in the areas of education, labor market analysis, political science, health, crime, environment, and others.

This extension requires the R Essentials.  It can be downloaded from the SPSS Community site (www.ibm.com/developerworks/spssdevcentral) in the Extension Commands collection or, for Statistics V22 users, it can be downloaded and installed from the Utilities menu.


Regards,

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：IMPORTANT updates update import Dates interested Community matching problems produced

A new extension command, STATS OPEN PROJECT has been uploaded to the SPSS Community website (www.ibm.com/developerworks/spssdevcentral) in the Extension Commands collection.

This command allows you to define a set of syntax and a list of data, syntax, and output files that constitute a project.  The syntax is run and the files are opened when you run this command.  You can set this as automatic behavior when Statistics starts.

You define a project in a simple text file that has one or more [RUN], [OPEN], and [PROJECT] sections.  RUN is equivalent to an INSERT command.  OPEN does a GET on SAV files and opens syntax and output files in the user interface.  PROJECT allows a project to open other projects, so if you have a core set of files such as a macro library you often use and then a variety of projects that depend on them, you can reference the core PROJECT in the other projects.  The dialog box help describes the file format in more detail.

This command provides a dialog box interface and standard Statistics syntax.  It requires the Python Essentials, which can be obtained via the SPSS Community website or the Statistics product materials, depending on the version.  As of V22, you can install OPEN PROJECT from the Statistics Utilities menu.  Otherwise you can download and install it from the website.

Regards,

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@us.ibm.com
phone: 720-342-5621

A new extension command, STATS GET TRIPLES, has been posted to the SPSS Community website (www.ibm.com/developerworks/spssdevcentral) in the Extension Commands collection (Downloads for SPSS Statistics; then Extension Commands).

This command reads data that is in Triple-S format.  A Triple-S dataset, typically a survey file, consists of an XML file that defines the metadata and a fixed format or csv file that contains the data.  The standard definition is at http://http://www.triple-s.org.

GET TRIPLES can read the data directly and can also create a standard syntax file that reads the data.  It includes functionality to handle multiple category and multiple response sets, and it can create custom attributes to deal with metadata that does not fit into the standard SPSS metadata.

This extension requires the Python Essentials.  It can be downloaded from the SPSS Community site or, for Statistics V22 users, it can be downloaded and installed from the Utilities menu.


Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@us.ibm.com
phone: 720-342-5621

Two new extension commands have been posted to the Extensions Collection on the SPSS Community website (www.ibm.com/developerworks/spssdevcentral).

STATS PROPOR REGR estimates regression models where the dependent variable is in (0,1).  It fits a beta distribution, and the regressors estimate the parameters of that distribution.

STATS PROPOR REGRPRED calculates predicted values for new data using a model produced by STATS PROPOR REGRPRED.

Users of Statistics 22 can download and install these commands using Utilities > Extension Bundles > Download and Install Extension Bundles.  Users on older versions of Statistics can get these from the website via Downloads for SPSS Statistics and then Extension Commands.

These commands require the R Essentials, which is also available via the website.

We hope you find these commands useful.


Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@us.ibm.com
phone: 720-342-5621

Two new extension commands have been added to the SPSS Community website (www.ibm.com/developerworks/spssdevcentral) in the Extension Commands collection.

STATS CATEGORY ORDER creates macros or multiple dichotomy sets with a specified variable order.  Items are created in sorted order of the count but allowing for special values such as "other" to be moved out of the sort order.  Inputs can be a collection of ordinary variables or one or more multiple dichotomy sets.  The macros can be used in the CATEGORIES subcommand of CTABLES.

STATS MCSET CONVERT converts a multiple category set to a multiple dichotomy set.  This may be useful in conjunction with STATS CATEGORY ORDER or when tabulating individual values or generating dummy variables for procedures such as REGRESSION.

Both of these require the Python plugin and work with Statistics V18 or later.

This brings the number of extension commands available from the SPSS Community site to 72.

Regards,

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@us.ibm.com
phone: 720-342-5621

A new version of the STATS LP extension command for SPSS Statistics has been uploaded to the SPSS Community site (www.ibm.com/developerworks/spssdevcentral) in the Extension Commands collection.  It requires the R Essentials.

The new version includes these improvements
- faster performance
- integer and mixed integer/linear programming
- support for bounds on the activity variables
- support for equality constraints.  (You can even solve sets of linear equations with it.)

Regards,

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@us.ibm.com
phone: 720-342-5621

Two new extension commands have been posted to the SPSS Community website (www.ibm.com/developerworks/spssdevcentral) in the Extension Commands collection on the Downloads for SPSS Statistics page.

STATS DBSCAN does density-based clustering with noise handling.  STATS DBPRED uses the results from STATS DBSCAN to predict the clusters for new data.  Density-based clustering depends on a maximum distance parameter and a minimum cluster size in addition to the variables.  It bears a relationship to hierarchical clustering with certain parameter settings, but it is reported to be much faster.

These procedures require the R Essentials.

Regards,

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@us.ibm.com
phone: 720-342-5621

A new extension command, STATS CANCORR, has been added to the SPSS Community website (www.ibm.com/developerworks/spssdevcentral) in the Extension Commands collection.  It computes canonical correlations and scores, and it can generate an SPSS Statistics syntax command file to score new data.

This command's functionality is similar to the old CANCORR macro, but it includes a dialog box interface, standard SPSS command syntax, and has pivot table output.  It requires the Python Essentials.

Regards,

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@us.ibm.com
phone: 720-342-5621

new extension command, STATS SURVREG has been posted to the Extension Commands collection of the SPSS Community website (www.ibm.com/developerworks/spssdevcentral).  Click on Downloads for SPSS Statistics and then Extension Commands to access it.  The procedure estimates survival regression for left, right, interval, and counting censoring for these distributions: Weibull, exponential, Gaussian, logistic, lognormal, loglogistic, and t.  It can do Tobit regression as well, although there is another extension command, SPSSINC TOBIT REGR, dedicated to that model.

It can save a dataset with a variety of residuals, predicted values, and diagnostic measures.

The package requires the R Essentials.  It has been tested mainly with Statistics 21 but is expected to work with versions back to 18.

We hope that you will find this package useful.

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@us.ibm.com
phone: 720-342-5621

A new extension command, STATS_EXRASCH, has been posted to the SPSS Community site (www.ibm.com/developerworks/spssdevcentral) in the Extension Commands Collection.  It estimates standard dichotomous Rasch, linear logistic test, rating scale, linear rating scale, partial credit, and linear partial credit models.  The output includes extensive tables, plots, and test statistics.  It requires the R Essentials, which are also available from the Community site.  The procedure has been tested mainly with Statistics 21, but it is expected to work with versions back to 18.

I would like to thank Ryan Black for his generous advice and review of the preliminary version of this procedure, which has resulted in a substantial increase in its quality.  He is, of course, not responsible for any defects that might remain.

Note:  Since the recent IBM developerWorks upgrade we have received some reports of malfunctioning links.  The dW team has traced this to a bug in Internet Explorer 8 which results in misinterpretation of some URLs used on the site.  If you encounter this problem, switching to Firefox or another browser should solve it.  The team is working on a solution.


Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
peck@us.ibm.com
phone: 720-342-5621