Home  /  Stata Conferences  /  2024 Italy

Proceedings

9:00–9:45
Session I: Invited speaker

Structural equation modeling with partial least squares using Stata Abstract:
(Read more)
Structural equation modeling (SEM) is a multivariate statistical framework that can model both observed and unobserved (latent) variables through complex relationships. In this talk, we present plssem, a community-contributed Stata package for partial least-squares SEM, an approach to SEM which has attracted a lot of interest in the last 20 years from an increasing number of researchers and practitioners from many fields such as marketing, information systems, economics, psychology, and others. After introducing the topic to the audience, we will illustrate the current architecture of the package and its main features.

Contributor:
Mehmet Mehmetoglu
Norwegian University of Science and Technology

(Read less)

Additional information:
Italy24_.Venturini.pdf

Sergio Venturini
Univeristà Cattolica del Sacro Cuore Cremona
9:45–11:30
Session II: Community contributed, I

Optimal policy learning using Stata Abstract:
(Read more)
This presentation introduces the Stata package opl for optimal policy learning, facilitating ex ante policy impact evaluation within the Stata environment. Despite theoretical progress, practical implementations of policy-learning algorithms are still poor within popular statistical software. To address this limitation, the package implements three popular policy learning algorithms in Stata (threshold-based, linear-combination, and fixed-depth decision tree), and provides practical demonstrations of them using a real database. Also, I present a policy scenario development proposing a menu strategy, which is particularly useful when selection variables are affected by welfare monotonicity. Overall, the package contributes to bridging the gap between theoretical advancements and practical applications of policy learning.

(Read less)

Additional information:
Italy24_Cerulli1.pdf

Giovanni Cerulli
IRcRES

Using marginal effects for interpretation in item response theory and in tests of differential item functioning: Introducing Stata commands irt_me and irt_dif Abstract:
(Read more)
The field of categorical data analysis has largely shifted from the limitations of coefficient interpretations to the more flexible and powerful possibilities afforded by marginal effects, spurred by Stata’s widespread implementation of the margins command. Despite item response theory being the latent variable corollary of categorical data analysis, a similar transformation in interpretation tools and practices has yet to emerge. I propose using tests of marginal effects for interpretation in item response theory models, demonstrating the advantages to this strategy over focusing on coefficients. Further, I show how to solve several issues when translating the idea of marginal effects to a latent variable model. A new command, irt_me, automates the estimation of marginal effects after any item response theory model (irt) in Stata, including models with binary, ordinal, nominal, and count items (or a mix).

Differential item functioning is a method of detecting item bias that has traditionally relied on tests of interaction terms in the item response theory coefficients. However, it is well established in the categorical data analysis realm that coefficients are inappropriate for tests of interaction: tests of the equality of marginal effects are instead the recommended approach. A new command, irt_dif, provides tests of differential item functioning by testing the equality of marginal effects from item response theory models fit across separate groups.

(Read less)

Additional information:
Italy24_Mize.pdf

Trenton D. Mize
Purdue University

Too much or too little? New tools for the CCE estimator Abstract:
(Read more)
This talk will cover new developments in the literature of common correlated effects (CCE) and their implementation in Stata. First, I will discuss regularized CCE. CCE is known to be sensitive to the selection of the number of cross-section averages. CCE overcomes the problem by regularizing the cross-section averages. Secondly, I will discuss the test for the rank condition based on DeVos, Everaert, and Sarafidis (2024, Econometrics Reviews). If the rank condition fails, CCE will be inconsistent and therefore testing the condition is key for any empirical application. Finally, I will show the selection of cross-section averages using the information criteria from Karabiyik, Urbain and Westerlund (2019, Journal of Applied Econometrics) and Margaritella and Westerlund (2023, The Econometrics Journal).

(Read less)

Additional information:
Italy24_Ditzen1.pdf

Jan Ditzen
Free University of Bozen
11:30–12:30
Session III: Exploiting the potential of Stata 18, I

Causal mediation analysis with Stata Abstract:
(Read more)
Causal inference is an essential goal in many research areas and aims at identifying and quantifying causal effects. By decomposing causal effects into direct and indirect effects, causal mediation provides further insight into underlying mechanisms through which causal effects operate. This talk presents the basic theoretical framework for causal mediation analysis and discusses a variety of examples using Stata’s mediate command. Examples will include linear and generalized linear models using a variety of outcome and mediator variables as well as different types of treatments.

(Read less)

Additional information:
Italy24_Luedicke.pdf

Joerg Luedicke
StataCorp
1:30–2:30
Session IV: Stata tips and tricks

nnls: Nonnegative least squares using Stata Abstract:
(Read more)
The nnls command enables users to carry out nonnegative least squares using Stata calling Python in the background. A simple application of the nnls Stata command on real data will be provided.

(Read less)

Additional information:
Italy24_Cerulli2.pdf

Giovanni Cerulli
IRcRES

htmltab2stata: Converting HTML tables into a Stata dataset Abstract:
(Read more)
htmltab2stata parses HTML code from websites. It detects tables enclosed with the HTML <table> environment and transforms the table into a Stata dataset.

(Read less)

Additional information:
Italy24_Ditzen2.pdf

Jan Ditzen
Free University of Bozen

Implementing groupwise – heteroskedasticity – robust – variance – covariance estimators in fixed-effects panel-data regression with Stata Abstract:
(Read more)
Stock and Watson (2008) prove that the plain White heteroskedasticity-robust VCE is generally inconsistent for fixed T , N -> ∞ in fixed-effects panel-data regression. Bruno (2024) proves that the aforementioned VCE is (fixed T , N _> ∞) consistent under groupwise heteroskedasticity (GH), that is, when the conditional variance of the idiosyncratic error is time-invariant but can vary across individuals. As is well known, the vce(robust) option of xtreg in Stata implements the cluster–robust VCE, not the White VCE. In this talk, I show simple Stata procedures to implement the White VCE and a second GH-robust VCE in fixed-effects panel-data regression. Monte Carlo experiments prove that both VCEs, under GH, have good finite-sample properties, compared to the bias-adjusted VCE by Stock and Watson and the cluster–robust VCE.

(Read less)

Additional information:
Italy24_Bruno.pdf

Giovanni Bruno
Bocconi University
2:30–4:00
Session V: Exploiting the potential of Stata 18, II

Bayesian model averaging Abstract:
(Read more)
Are you unsure which predictors to include in your model? Rather than choosing one model, aggregate results across all candidate models to account for model uncertainty with Bayesian model averaging (BMA). Which predictors are important given the observed data? Which models are more plausible? How do predictors relate to each other across different models? BMA can answer these questions and many more.

Stata 18 introduced the bma suite of commands to perform BMA in linear regression models. In this talk, you will learn how to explore influential models, make inferences, and obtain better predictions with BMA. I will demonstrate the utility of BMA for any researcher—Bayesian, frequentist, and everyone in between! No prior knowledge of the Bayesian framework is required.

(Read less)

Additional information:
Italy24_Cain.pdf

Meghan Cain
StataCorp

Text mining in economics and health economics using Stata Abstract:
(Read more)
Within the more relevant data science topics, text mining is an important and active research area that offers various ways to extract information and insights from text data. Its continued use and improvement could drive innovation in several areas and improve our ability to interpret, evaluate, and utilize the vast amounts of unstructured text produced in the digital age. Extracting insightful information from text data through text mining in healthcare and business holds great promise. Text mining in business can provide insightful information by analyzing large amounts of text data, including research papers, news, and financial reports. It can help analyze market sentiment, identify emerging trends, and more accurately predict economic indicators by economists. For example, economists can find terms or phrases that reflect investment behavior and sentiment changes by applying text-mining methods to financial news. Text mining can provide essential insights into health economics by examining various textual data, including patient surveys, clinical trials, medical records, and health policy. Researchers and policymakers can use it to understand healthcare utilization patterns better, identify the variables that influence patient outcomes, and evaluate the effectiveness of different healthcare treatments. Text mining can examine electronic health data and identify trends in disease incidence, treatment effectiveness, and healthcare utilization. In this presentation, I will illustrate the instruments currently available in Stata to facilitate several text-mining methods.

(Read less)

Additional information:
Italy24_Drago.pdf

Carlo Drago
University Niccolò Cusano
4:15–5:30
Session VI: Community-contributed, II

geoplot: A new command to draw maps Abstract:
(Read more)
geoplot is a new command for drawing maps from shapefiles and other datasets. Multiple layers of elements such as regions, borders, lakes, roads, labels, and symbols can be freely combined, and the look of elements (for example, color) can be varied depending on the values of variables. Compared with previous solutions in Stata, geoplot provides more user convenience, more functionality, and more flexibility. In this talk, I will introduce the basic components of the command and illustrate its use with examples.

(Read less)

Additional information:
Italy24_Jann.pdf

Ben Jann
University of Bern

Fitting spatial autoregressive logit and probit models using Stata: The spatbinary command Abstract:
(Read more)
Spatial regressions can be estimated in Stata using the spregress, spxtregress, and spivregress commands. These commands allow users to fit spatial autoregressive models in cross-sectional and panel data. They are designed to estimate regressions with continuous dependent variables. The spatbinary command now allows Stata users to fit spatial logit and probit models, which are important models in applied econometrics.

(Read less)

Additional information:
Italy24_Spinelli.pdf

Daniele Spinelli
University of Milano-Bicocca
5:30–5:50
Session VII: Application study using Stata

Do alternative work arrangements substitute standard employment? Evidence from worker-level data Abstract:
(Read more)
This study analyzes the impact of an alternative work arrangement (AWA) called “voucher” on earnings of atypical workers and on their alternative income sources using Italian administrative data. Specifically, I investigate whether this form of very flexible work substitutes income from more standard labor contracts and welfare transfers related to employment insurance (sick and parental leave and unemployment benefits). I estimate cross-income elasticities using fixed effects and difference-in-differences specifications that correct for sample selection of individuals in the labor market. Results show that vouchers increase overall labor income, but they also substitute earnings derived from other labor contracts. I do not find relevant associations between vouchers and welfare transfers. The positive effect of vouchers on total income is smaller in specifications that correct for sample selection bias, and the substitution effect with other labor income sources is substantially larger. Overall, my findings show that AWAs tend to substitute standard employment, with small positive net effects on earnings, which are larger for intensive users of vouchers, and in geographic regions with a more sizable informal sector.

(Read less)

Filippo Passerini
University of Bologna and LABORatorio R. Revelli
5:50–6:15 Open panel discussion with Stata developers
Contribute to the Stata community by sharing your feedback with StataCorp's developers. From feature improvements to bug fixes and new ways to analyze data, we want to hear how Stata can be made better for our users.

Workshop: Introduction to partial least-squares structural equation modeling (PLS-SEM) using Stata

Instructors

Sergio Venturini (Università Cattolica del Sacro Cuore–Cremona)
Mehmet Mehmetoglu (Norges teknisk-naturvitenskapelige universitet)

Date

10 May 2024

Description

PLS-SEM, also referred to as partial least-squares path modeling, is an alternative approach to SEM that is being increasingly used in social sciences, psychology, business administration, and marketing. PLS-SEM can be viewed as a component-based SEM alternative to the covariance-based structural equation modeling, which can be described as a factor-based SEM technique. As such, the PLS-SEM approach provides researchers with a multivariate statistical technique that can readily be used to fit exploratory and complex SEM models. Although there are several standalone specialized PLS-SEM software packages available, this course introduces participants to the PLS-SEM methodology through the community-contributed Stata package plssem, which was developed by the course instructors.

The course is of particular interest to researchers and professionals working in social sciences, psychology, business administration, marketing and management. Because of its introductory nature, it is also accessible to individuals, regardless of their respective disciplines or fields, who need to acquire the requisite toolset to apply the PLS-SEM methodology to their own data. During the course, theoretical concepts are reinforced by applied case study examples in which the course instructors discuss current research issues, highlighting potential pitfalls and the advantages of individual techniques.

  • Session I: Structural equation modeling (SEM) and different approaches to SEM; PLS-SEM versus CB-SEM
  • Session II: Specification of a PLS-SEM model; reflective and formative constructs
  • Session III: Assessment of the results of a PLS-SEM analysis; construct and discriminant validity; goodness of fit; path coefficients interpretation
  • Session IV: Advanced issues in PLS-SEM; mediation analysis; multigroup analysis; higher-order constructs

Prerequisites

Participants are expected to have knowledge in basic statistics. More specifically, a working knowledge of linear regression analysis is required. Previous exposure to Stata or other statistical software packages would also be an advantage.


Scientific committee

Una-Louise Bell
TStat – TStat Training

Rino Bellocco
University of Milano-Bicocca
Giovanni Capelli
Istituto Superiore di Sanità
Maurizio Pisati
University of Milano-Bicocca
Giovanni Cerulli
IRCRES-CNR

Logistics organizer

The logistics organizer for the 2024 Italian Stata Conference is TStat S.r.l., the distributor of Stata for Italy, Albania, Bosnia and Herzegovina, Greece, Kosovo, North Macedonia, Malta, Montenegro, Serbia, Slovakia, and Slovenia.

View the proceedings of previous Stata Conferences and Users Group meetings.