The Oceania Stata Users Group Meeting was Friday, 29 September 2017 at the Australian National University Commons, but you can view the program and presentation slides below.
Keynote 1: Incorporating Stata into reproducible documents
Abstract: Part of reproducible research is eliminating manual steps such as having to edit documents. Stata 15 introduces several commands that facilitate automated document production, including dyndoc for converting dynamic Markdown documents to webpages, putdocx for creating Word documents, and putpdf for creating PDF files.
These commands allow you to mix formatted text and Stata output and allow you to embed Stata graphs, in-line Stata results, and tables containing the output from selected Stata commands. We will show these commands in action, demonstrating how to automate the production of documents in various formats and how to include Stata results in those documents.
A graphic comparison of the Fieller and Delta intervals for ratios and estimates
Abstract: Fieller's 1954 proposal for the use of an inverse test to construct confidence intervals (Cis) for the ratio of normally distributed statistics has been shown to be superior to the application of the Delta method in a number of applications. In this presentation, I demonstrate how a simple graphic exposition can be used to illustrate the relationship between the Delta and the Fieller Cis. The advantage of the graphical presentation over the numeric result available in the Fieller Stata ado (Coveney 2004) is that it may indicate how the level of significance can be changed to result in finite upper and lower bounds.
In addition, I also demonstrate how this method can be used to draw inferences for the turning points in high order and fractile polynomials by the definition of the implied CIs for the first derivative function.
A number of examples are provided in the area of the structural coefficient in the exactly identified two-stage least squares estimator, the inflexion point in an environmental Kuznets curve and the 50% dose in a dose response model. These examples can all be shown with a simple Stata do-file for the plots of a series of generated values.
University of Melbourne
Stata: A key strategic statistical tool of choice in major impact evaluations of socioeconomic programs
Abstract: This presentation outlines the program logic of any major international impact program evaluation; discusses strategic considerations in such quantitative and qualitative evaluations, particularly the key attributes of a strategic analytical tool; and finds that Stata is internationally a highly regarded, state-of-the-art software used in data analysis of impact evaluations of socioeconomic programs.
Modeling technology adoption decisions among small-holder cassava producers in East Africa
Abstract: Cassava is the second most important food crop in Africa after maize. It is a major staple crop for more than 200 million people in East and Central Africa, most of them living in poverty in rural areas. However, its production is undermined by several factors, particularly the problem of emerging pests and diseases. The whitefly (Bemisia tabaci) is the most serious pest of cassava, causing significant yield losses through direct feeding damage and as a carrier of virus diseases. However, there are few empirical assessments of the economic impacts of the whitefly on small-holder producers. Some of these constraints are also associated with socioeconomic factors, while others have to do with environmental and institutional factors. We conducted a comprehensive socioeconomic study covering Uganda, Tanzania, and Malawi to determine the status of cassava production with the following specific objectives: (1) What is the present status of cassava production and productivity in Uganda, Tanzania, and Malawi? (2) What is the current adoption rate of improved cassava production technologies in the study countries? (3) What is the economic impact of B. tabaci complex on small-holder farmers?
The primary data for this study were collected from cassava farmers in Uganda, Tanzania, and Malawi—using a pretested survey questionnaire that was orally administered to individual farmers. A total of 1,200 respondents were selected and interviewed using a multistage random-sampling technique. Using the mvprobit routine in Stata, we employ a multivariate probit regression to model simultaneous adoption of interrelated technology among small-holder cassava farmers. Here I present preliminary results and discuss the implications.
Keynote 2: Data.gov.au and NationalMap: What's up with that
Abstract: In December 2015, the Australian government released its Public Data Policy Statement, which required agencies to make open data discoverable through data.gov.au. Since the release of the Statement, the amount of open data discoverable through the platform has increased considerably.
In this presentation, I'll introduce you to the data.gov.au and NationalMap platforms, and share some tips and tricks about using them. I'll cover things like searching for data, the data.gov.au API, and NationalMap's ability to turn tables into spatial data. Prepare to be pretty hands-on.
In June 2016, we started working with Data61 to create the next generation of the data.gov.au platform. To conclude, I'll talk about the work done to date and introduce you to our prototypes and design concepts that showcase what the future of data.gov.au could look like.
The Department of Finance
Digital Transformation Agency
Extracting metadata from Stata datasets
Abstract: Data processed using Stata are often stored in proprietary Stata (*.dta) files. This is practical and useful for the life of a project but creates an obstacle for anyone who wishes to use the data and doesn't have Stata. The Stata command export writes data from a Stata dataset to a text file, a format that ensures portability and long-term accessibility. However, the variable-level information in Stata data files, such as data types, variable labels, and value labels, is lost. Ideally, these metadata would also be extracted into a text file for long term preservation. In addition, the metadata could be imported into data capture software such as REDCap. My presentation will introduce an ado-program that extracts variable-level metadata in a CSV format and illustrate how easily the metadata can be used to create either an empty or a populated REDCap database.
Murdoch Childrens Research Institute (MCRI)
Adolescent interest in science careers in Europe: Trends between 2006 and 2015: Example of Stata analysis
Abstract: In this project, I investigate trends in youth vocational interests related to STEM (science, technology, engineering and mathematics) in the OECD's Programme for International Student Assessment data pooled for 2006 and 2015 data from 26 European countries. The focus is on ascertaining whether EU member states have succeeded in maintaining (or increasing) students' interest in STEM careers in the decade and on exploring whether gender segregation within career expectations related to STEM has changed over time and across the EU countries.
The data we use represent the so-called Large-Scale Assessment Studies, and the presentation will focus mainly on the challenges of effective visual presentation of key results using Stata.
xtbreak: Structural change tests in heterogenous panels with known or unknown breaks
Abstract: xtbreak describes a statistical test for structural change in panel-data models, wherein the break may affect some, but not all, cross-sectional units in the panel dataset. The test can be applied for cases of known and unknown break dates. The test is robust to nonnormal, heteroskedastic and autocorrelated errors, and end-of-sample structural change.
The University of Sydney
Standardized categorization of maternal alcohol consumption throughout pregnancy
Abstract: The AQUA (Asking Questions about Alcohol in Pregnancy) study assesses the impact of alcohol consumption on the unborn child. Data have been collected on nearly 1,600 pregnant women across four waves: the three months pre-pregnancy and at each trimester of pregnancy.
The quantity and frequency of alcohol consumption for each woman and wave were collapsed into one of four categories (none, low, moderate, or high), representing her average weekly alcohol consumption. A single occasion where 50 grams or more of alcohol were consumed was defined as a binge episode. Both average weekly consumption and number of binge episodes are taken into account in my analysis. The same survey questions on alcohol consumption were asked at each wave.
I will present a standardized system for coding these questions and also the algorithm that calculates a woman's overall level of alcohol consumption at each wave. My ado-file can be adapted to other surveys aiming to quantify alcohol consumption.
Full presentation coming soon
Keynote 3: What's new in Stata 15
Abstract: A brief overview of the new features of Stata 15. I will be discussing the newest features in what StataCorp President, Bill Gould, calls "our most remarkable release yet."
Wishes and grumbles
Stata User Group Workshops: 28 September 2017
Panel Data Using Stata, by Vasilis Sarafidis, offers a comprehensive treatment of the analysis of panel data that will allow participants to pragmatically deal with fundamental issues, such as controlling for individual heterogeneity, reducing collinearity among regressors, addressing statistical hypotheses, and identifying effects that are simply not detectable in pure cross-section or time-series data.
Survival Analysis Using Stata, by Rory Wolfe, assumes familiarity with basic statistical concepts such as medians, confidence intervals, and p-values. It does not assume any familiarity with survival analysis methods. Participants should be familiar with the basics of performing statistical analysis in Stata.