Abstract: Sequences are entities built by a limited number of elements that are ordered in a specific way. A typical example is human DNA, where the elements adenine, cytosine, guanine, and thymine (the organic bases) are ordered into a sequence. Other sequences are words, which are built by letters that appear in a specific order, or careers of employers, which are built by specific job positions ordered for a long time.

The user‐written SQ‐Ados, published in 2006, provide a number of tools to describe sequences and to measure the similarity between pairs of sequences. Since its first publication, the SQ‐Ados have been continuously updated with new functionalities.

This presentation gives an overview of existing programs and discusses the most recent additions, including functions for string comparisons, nearest neighbor identification, and a simple interface to the plugins for sequence analysis developed by Brandan Halpin.

Abstract: Although nonparametric regression is becoming a popular methodology in many disciplines and research contexts, a multivariate version of such an approach was not available in Stata, neither as a built-in command, nor as a user-written command. So far, in fact, StataCorp, as well as some users, has implemented only bivariate versions of local kernel regressions (like the popular lpoly command).

The proposed Stata routine mkern aims at filling the gap. This routine employs a "radial local weighted mean" approach by using various Kernel functions as weighting schemes (at the user’s choice).

By default, mkern also provides the "optimal bandwidth" by means of a (computational) cross-validation approach. However, users can also provide their own choice of the bandwidth, thus producing estimation for both oversmoothing and undersmoothing cases.

Finally, as an option, mkern offers a graphical plot of the row data against the predicted values to assess the goodness-of-fit of the provided estimation.

Abstract: SWire is a software interface enabling one to query Stata for the executing of basic operations like reading or writing data. The core of SWire is a Java plugin that functions as a server and that can be queried by using the HTTP protocol. SWire is in essence a wrapper around the Stata-Java API, and all the functions relating to this API can therefore be externally called by clients. SWire can be interrogated on a local network and it accepts AJAX queries, thus enabling the development of web applications that provide the user with graphical user interfaces. For example, a web application running on a browser can interact with Stata on a local network; this application could be a data entry form or an automatic report. By providing a communication protocol, SWire makes it possible to connect Stata with many other applications for data exchange and basic interaction, thereby extending the capabilities of Stata.

Abstract: Stata 14 introduced a suite of commands for performing Bayesian analysis directly in Stata. This presentation will be a hands‐on demonstration of these commands that will illustrate the full path of a Bayesian analysis from specification of prior distributions through use of MCMC chains to estimate the model to assessment of the mixing and convergence of the MCMC chain. Along the way, the flexibility of Bayesian models as well as the ways in that the analysis and interpretation differ from their frequentist counterparts will be illustrated.

Abstract: Quite often, observed income and survival data are incomplete because of left- or right-censoring or truncation. Measuring inequality, for instance, by the Gini index of concentration, from such incomplete data can produce biased results. We describe the Stata package GiniInc, which contains three independent programs to estimate the Gini concentration index under different conditions. First, survgini computes a test statistic for the comparison of two (survival) distributions based on the nonparametric restricted Gini index, using both asymptotic and permutation inference. Second, survbound computes nonparametric bounds for the unrestricted Gini index from censored data. Finally, survlsl applies maximum likelihood estimation for three commonly used parametric models to estimate the unrestricted Gini Index, both from censored and truncated data. We describe the package and illustrate its use through examples.

Abstract: The diffusion of international production networks has challenged the capability of traditional trade statistics to provide an adequate representation of supply and demand linkages among the economies. To address this issue, new statistical tools (the Inter-Country Input-Output tables) and new analytical frameworks have been developed. icio exploits the ICIO tables (for example, WIOD, OECD TiVA) to compute, display, and export i) standard measures of gross and value-added trade as in OECD.Stat; ii) Koopman, Wang, and Wei's (2014) standard decomposition; and iii) Borin and Mancini's (2015) bilateral decomposition, sink or source at the aggregate or sectoral level. Of special note is that the new command provides a user-friendly Stata dialog interface. In this paper, we first briefly describe the structure of ICIO tables as well as Koopman, Wang, and Wei's (2014) and Borin and Mancini (2015) methodologies. Then, a number of examples are used to illustrate the command features and its main options, focusing on the use of the dialog interface.

References:

Borin, A., and M. Mancini. 2015. Follow the value added: bilateral gross export accounting. Temi di discussione (Economic working papers), No. 1026, Bank of Italy, Economic Research and International Relations Area.

Koopman, R., Z. Wang, and S. Wei. 2014. Tracing value-added and double counting in gross exports. American Economic Review 104: 459–494.

Abstract: Item response theory (IRT) models enable us to investigate the relationship between an unobservable latent trait such as mathematical ability, attitude toward a social policy, or customer satisfaction and the test or questionnaire designed to measure that latent trait. Although IRT has a long history in education and psychometrics, it is now being adopted in many other fields such as sociology, business, and health research. This presentation provides an overview of IRT and demonstrates how to use the commands introduced in Stata 14 for fitting these models and for graphically investigating the relationships between test questions and latent traits. The session concludes with an evaluation of group differences in IRT models.