发帖
雷达卡
Here are some of the treats in Stata 14.1.
Bayesian analysis. The new Bayesian analysis features introduced in Stata 14 now have dramatically improved performance and capabilities when working with latent variables and with panel and multilevel models. And, there is more in Bayesian analysis. See help whatsnew after updating.
Multiple imputation. You can now add your own imputation methods to mi impute—Stata’s main engine for handling multiple imputation of missing data. It’s easy. After updating, type help mi impute usermethod for documentation and examples.
IRT. Those who use IRT (item response theory) will be glad to see the new diflogistic command that performs the logistic regression test for uniform and nonuniform differential item functioning (DIF). The Mantel–Haenszel test has improvements too. Type help difmh after updating.
Mixed models. Small-sample denominator-degree-of-freedom adjustments such as Kenward–Roger and Satterthwaite can now be performed when estimating contrasts via the contrast command after estimating linear mixed models.
MLEs of expressions. You can now predict expected means after mlexp—maximum likelihood estimation (MLE) of any expression you specify. This also implies that margins can be used after mlexp to estimate marginal and conditional effects, estimated marginal means, and predictive margins. mlexp also now supports linear constraints.
ICD-10. The ICD-10 facilities now work with all codes from 2003 to 2015.
Propensity-score matching. Propensity-score matching via teffects psmatch is now much faster for large datasets.
More CIs. You can now get confidence intervals (CIs) for variances and standard deviations from the ci commands.
Exporting results to Excel. The syntax for exporting estimation and other results to Excel spreadsheets has been dramatically simplified. You can also easily choose and export estimation and other results using the intuitive and functional dialog box.
High DPI support. Under Windows, high DPI monitors are now better supported. A host of icon sizings keep your icons looking good when working at higher resolutions.
Readme.txt
(48 Bytes, 需要: 5 个论坛币)
京公网安备 11010802022788号
