This toolkit aims to guide researchers interested in investigating the economics of tobacco control, especially in low- and middle-income countries such as Mexico. It presents theoretical background on the economics of tobacco and provides step-by-step tools developed in Stata to estimate own- and cross-price elasticities of tobacco products and the crowding-out and impoverishing effects using household expenditure surveys (HES). It deals with standard issues with HES and provides tips for data management and analysis in Stata. These assessments are basic inputs for designing better fiscal policies, which are the most effective measures to reduce tobacco use. The tools included could also be applied to other harmful products such as alcoholic or sugar-sweetened beverages, which are also major risk factors for NCDs. Case studies from low- and middle-income countries implementing ad hoc and replicable Stata do-files are also provided in the toolkit. The policy discussions and rationale of different economic concepts in tobacco control and interpretation of results could also benefit policy makers, analysts in government, and civil society organizations engaged in tobacco control activities.

Two approaches to economic theory are used: Neoclassical and evolutionary. Two models are used: fixed-effects data panel (PE) with instrumental variables (IV) and the quantile model for data panel. Porter's hypothesis (greater environmental regulation leads to greater innovation and therefore greater competitiveness) has been a controversial topic since its appearance in the 90s. Studies have tried to prove or disprove it with different results: depending on factors such as the available data, the variables, the methodology used, and the level at which the analysis is done (macro or micro). The sequence where environmental regulation influences first innovation, and then productivity deals with endogeneity and possible problems of bias and asymmetry. By using total factor productivity (TFP) and the EPS environmental stringency index created by the OECD, over a 20-year period the results show evidence in favor of Porter.

When the USMCA entered into force, the volume of trade has grown considerably between the North American partners because it ought to strengthen the economic relationship of these nations. However, some countries present a high commercial concentration due to the exports and imports carried out between countries to satisfy the demands of the commercial partners. By using Stata, the Herfindahl–Hirschmann index (HHI) is computed (Ansari 2012) with the command version hhi5 by Yujun and Lian (2016) and concentration indexes of exports from Mexico and the United States of America to perform an analysis of exports in the states of the mentioned countries, to identify the position of the key states for cross-border trade through commercial corridors established by the USMCA, where 70% of North American trade moves. The most important corridors are the West Coast Corridor, the Canamex Corridor, and the North American Superhighway Corridor.

Most empirical applications consider a fixed unknown underlying data-generating model (DGM) that researchers try to find, based on a particular theoretical framework that is combined with the data associated with the variables involved in the selected model specification. Bayesian model averaging provides an approach, where instead of focusing the estimation on the search for that unique unknown model, researchers can incorporate the uncertainty about the DMG to obtain probabilities associated with relevant predictors, measurements about complementary or substitutable predictors across different model candidates, and also predictions that incorporate uncertainty about the model and the parameters. In this presentation, I will use the new suite of bma commands to illustrate those and other aspects that can be derived using Bayesian model averaging.

Several strategies have been proposed in the literature to respond to these difficulties, the most popular of which is to report average marginal effects (on the probability scale) rather than odds ratios. Average marginal effects (AMEs) have many desirable properties but at least in part they throw the baby out with the bathwater. The size of an AME strongly depends on the marginal distribution of the dependent variable; for events that are very likely or very unlikely the AME necessarily has to be small because the probability space is bounded. Logistic regression, in contrast, estimates odds ratios which are free from such flooring and ceiling effects. Hence, odds ratios may be more appropriate than AMEs for comparison of effect sizes in many applications. Yet, logistic regression estimates conditional odds ratios, which are not comparable across different specifications. In this presentation, I aim to remedy the declining popularity of the odds ratio by introducing an estimand termed the “marginal odds ratio”; that is, logit coefficients that have properties similar to AMEs, but which retain the odds ratio interpretation. I define the marginal odds ratio theoretically in terms of potential outcomes, both for binary and continuous treatments, I discuss estimation methods using three different approaches (G-computation, inverse probability weighting, RIF regression), and I present Stata software implementing these methods.

DASP commands provide short avenues for estimating and producing graphic material to analyze easily economic inequality. This presentation focuses on the clorenz, cdensity, igini, and digini commands, which are programmed as ado-files; in addition, examples of its application by using microsimulated databases are shown through the Household Income and Expenditure Survey (ENIGH), 2020. The exercises compare the set of syntax with standard calculation Stata language, necessary for the calculation of inequality, Lorenz curves, and kernel density curves, and in parallel, the output is replicated with the DASP commands mentioned.

The construction of the index was based on the manual for measuring education and financial inclusion proposed by the OECD/INFE and 14 questions from the ENIF (National Survey of Financial Inclusion), while specific Stata commands were used to calculate the decomposition. To estimate high and low probabilities, an ordered multinomial probit probabilistic model was generated. The data were obtained from the four microdata sources of the 2021 National Survey of Financial Inclusion (ENIF), published by INEGI (National Institute of Statistics and Geography). The results confirm that the inequality in financial literacy is a consequence of a social structure problem, which contributes to new empirical evidence. Finally, exercises of this nature, using Stata, allow for the argumentation of new ways to create and evaluate more efficient variables for econometric model estimation.

The objective is to analyze the trend of the prevalence of VP in men and women and identify the associated factors in Mexico. The data used come from the National Survey of Addictions 2011 and the National Survey of Drug, Alcohol and Tobacco Consumption 2016; a sample of 44,963 individuals was selected. By using Poisson models with Stata, we show that prevalence of PV was 15.58% in 2011 and 14.90% in 2016. The associated factors were being a woman (RR=1.09, IC95%0.99–1.19), alcohol consumption by the partner (RR=1.68, CI95% 1.54–1.84) and drug use by the partner (RR=2.80, CI95% 2.46–3.18). Single marital status (RR=0.66, 95% CI 0.56–0.78); having previous partners (RR=0.60, 95% CI 0.55–0.66); more years of living with a partner (RR=1.81, 95% CI (1.47–2.23), living in an urban area (RR=1.18, 95% CI 1.05–1.33). Main conclusions display how prevalence of IPV has decreased mainly in the population that has higher family income. Factors associated with VP are similar in both sexes, so actions aimed at preventing this problem should include men and women.

There are generalized mixed linear models, which include the Poisson models and their variants. The use of the quasi-Poisson variant is extremely attractive when there is overdispersion in the distribution of the data because it allows generation of association models based on the Poisson distribution. This presentation presents the criteria and procedures for the choice and generation of a quasi-Poisson model in Stata, using as an example an association model with data from an entomoviral surveillance study.

For this purpose, a survey was designed with 76 questions, including the open-ended question: As a result of the COVID-19 health contingency, how has your teaching changed? The survey was answered by 750 professors, of whom 627 responded to the question. We present the analysis method applied with Stata 17 to analyze the 627 responses. Coincidence analysis was performed, a research technique that analyzes texts, documents, or responses by extracting keywords to obtain structured information and identify possible response patterns. Text analysis tools (txttool, precoin, and coin) were used to identify the most frequent words and possible relationships between them. Implementing these tools made it possible to obtain information on the modifications made by the statistics professor in his teaching due to the COVID-19 health contingency.

The adjustment model was developed using various lifestyle factors, such as physical activity, tobacco use, and family history of cardiovascular disease, in addition to ultra-processed food variables. The adjustment model was created using Stata through a series of steps, beginning with exploratory analysis, moving on to univariate analysis, and concluding with stepwise analysis. The resultant model was assessed for interaction, multicollinearity, and linear regression hypotheses. Data from 230 freshman university students enrolled at the Instituto Tecnológico de Sonora (ITSON) were examined.

Updated data are automatically imported from various sources, such as the Timely Statistics of the Ministry of Finance (for example, import delimited https://...), the Economic Information System of the Institute of Statistics (INEGI), and the Censuses and household surveys. After cleaning and saving the databases (sysdir_site) in a “hosting” of a tax simulator, a sysprofile.do file is elaborated to link all the Stata programs in the office to a shared folder. This process allows access to 78 programmed do- and ado-files as well as preprocessed databases. With this, anyone can easily request income, expenditures, both financing and indebtedness specific to a given year, and desired concepts in a coordinated work. In addition, an internal command is introduced to automatically integrate Stata values into LaTeX documents, which facilitates the generation of reports and documents with accurate and up-to-date information.

The interventions are to 363 cases and 587 controls who underwent the one-step test with oral overload of 75 grams of glucose with baseline determination at one hour and two hours to determine the presence or not of gestational diabetes mellitus. From the electronic file, sociodemographic, anthropometric, gynecoobstetric, pathological and nonpathological antecedents were collected. The measurement was performed with Stata 17 with a univariate exploratory analysis using the sample mean and standard deviation to determine the centrality and dispersion. Subsequently, a bivariate analysis was carried out to determine the association and correlation of the variables of interest with the presence or absence of gestational diabetes. Finally, a comprehensive logistic model with the study factors was used to determine their effect and statistical significance. The results are that women with gestational diabetes mellitus have greater age, weight, and obstetric risk, and the main risk factors associated with gestational diabetes were age and obesity.

Variables identified with the potential to be modifiable can in turn be used in preventive programs. The objective of this research was to analyze the association between behavioral variables related to cardiovascular health with anthropometric indicators of obesity in freshman university students enrolled at the Technological Institute of Sonora. The response variable was body fat, and the predictor variables were food and nutrient groups and physical activity, according to the criteria of the American Heart Association. Potential association analyses were used, and multiple models were built by stepwise forward selection (p≤0.05 and biological plausibility) with data from 230 university adolescents using the Stata software.

The analysis considered the design variables necessary for the adequate handling of the information. Using the svy command, the prevalence of previous diagnosis of type-2 diabetes (PDT2D) was estimated in a representative sample of Yaqui indigenous adults (n=351), inhabitants of the traditional towns of the ethnic group in Sonora. In the same way, the means and proportions of the possible factors associated with PDT2D were calculated, and the number of individuals of the indigenous group that presented the variable of interest was known.

Methods to determine blood lead levels seek to find the quantitative concentration in ug/dL of blood lead when statistically analyzed on population samples rarely seen with normal distribution. We will discuss three ways to analyze blood lead levels by describing the differences and achievements of each type of analysis using Stata and the National Survey of Health and Nutrition (ENSANUT) 2018 open data. The study consists of a cross-sectional analysis of the capillary blood samples obtained in the survey, measured in ug/dL of blood, with three ways of statistical processing: with the logarithmic transformation for the analysis with linear regression, when analyzing the data obtained naturally with robust regression, and with categorical analysis with cutoff points referred to in international regulations with logistic regression; multivariate models were compared with the same adjustment variables. The strategies for the selection of the multivariate analysis are made not only because they are new or novel but also to maintain consistency with the results of other studies that are internationally comparable.

Think about, for example, the effect of job training programs on earnings or the effectiveness of COVID vaccines. To capture this heterogeneity, Stata 18 introduces two commands that estimate treatment effects specific to each cohort and time period. For repeated cross-sectional data, we have hdidregress. For panel data, we have xthdidregress. Both commands let you aggregate treatment effects by cohort and exposure to treatment and visualize these effects graphically. Tests of pretreatment parallel trends are also available. This presentation will illustrate how both commands work and briefly discuss the theory underlying them.

We say the “unit of allocation” is the cluster, or the group. One of the main consequences of a cluster design is that participants within any one cluster often tend to respond similarly and thus their data can no longer be assumed to be independent. It is important that the analysis of a cluster-randomized trial accounts for this issue. In our Meta Salud Diabetes cluster randomized controlled study, summary statistics were used to describe the clusters and participants at baseline. Differences between the intervention and control arms at 3 (T3) and 12 months (T12) were estimated from linear mixed models. We used the repeated measures for mixed models, which uses an unstructured time and covariance matrix. Unadjusted models included fixed effects of strata (North, Central, South), the outcome at baseline (also known as analysis of covariance), time, treatment and time x treatment. A random effect for cluster was included and the Kenward–Roger adjustment to the denominator degrees of freedom was used.

These models account for the correlation due to repeated measures on participants nested within clusters, allow for testing at specific time points as well as the difference in pattern of change over time between arms, and give valid results when data are missing completely at random and at random (if the models are specified correctly and include covariates associated with missingness). Binary outcomes were tested with generalized linear mixed models using the logit link and binomial distribution. To account for potential selection bias or imbalance of key covariates in individual participants within clinics, adjusted models were also fit and included the baseline variables of age, gender, SES score, any CVD medication use (blood pressure, heart or stroke medication), smoking status and time in the GAM (just joined, <1 year, >1 year). We also performed two sensitivity analyses. We first replaced the lipid levels that were beyond the limit of detection with missing values. The second sensitivity analysis used multiple imputation to fill in missing data in the primary analyses. Variables associated with missingness, the outcome, or both were included in the imputation model, as well as all variables from the analysis model, including center. Multiple imputation with chained equations using 100 imputations was used, and the imputation was performed separately by arm.

the effects that can produce one relevant benefit in México named Pensión para el bienestar de adultos mayores on any other member of the household, such as the preference for working less with fewer number of hours related to the age of the household occupied members. I employ Stata to capture the cross-section impacts of this policy with a Bayesian probit regression model with sample selection (BPSS) by using microsimulated data from MEXMOD fed with Encuesta Nacional de Ingresos y Gastos de los Hogares in 2014 and 2020 (ENIGH).

Two examples of regional microeconomic spillovers interest us: 2010–2022 changes in enrollment and graduated at higher education in social sciences with the ANUIES dataset and 2005–2022 Mincer schooling returns distribution with the ENOE and INEGI microdata. The syntax, matrix results, and templates that are presented show the versatility of Stata and Mata as ideal tools in the management and analysis of large volumes of data with a focus on statistical and econometric analysis with strategies based on learning and teaching that require the use of real and recent information to be replicated, summarized, and analyzed with algorithms, procedures, and structured code.

solar radiation, deforestation, among others) obtained from the National Water Commission and social variables (birth rate, population density, population, educational level, among others) obtained from population censuses. Said projection will be based on four moments: 1. Obtaining and merging bases of data from different sources to build a single database with both variable types. 2. From generating regressions to understand the type and degree of conclusions based on the same model for these variables. 3. With the coefficients of realization, establishment of projections at the level, country, states, regions in the states and if possible, even at municipal levels. 4. We will use Stata for the pilots of the program. We present the results of the pilot and his way of doing it in Stata.

We explore the effect of sex, skin tone and hair color, face symmetry—as a proxy for beauty or attractiveness—and country of origin. Net of a explicit productivity measure, we find a discriminatory employment penalty of 11% from Central and South American workers as well as a penalty for workers with asymmetric faces of 9% that is present only when operatives take firing decisions—when managers take firing decisions, no “beauty effect” is present. For wages, we find only weak evidence that migrants from Central and South America are offered lower wages than native workers in the Mexican labor market. Finally, we find strong evidence of a sex wage penalty: women are offered wages that are about 6.6% lower than those offered to men.

The objective is to evaluate the effect of such a policy in urban communities in Mexico through Stata, using microdata from the National Household Income and Expenditure Survey (ENIGH). A working do-file is shown to clean, classify, and describe the variables, regrouping the types of food and using deciles (xtile) according to socioeconomic attributes. A difference-in-differences econometric model is designed to isolate the effect at the regional and temporal levels between control and treatment groups. With loops, 5 regions and 19 selected products are interacted verifying MCO linearity assumptions through the tests: VIF, estat imtest, estat hettest, sktest error, and swilk error, kdensity error, normal. Tables and graphs edited with the outreg2 command are reported, and it is observed that labeling is effective in reducing the expenditure of certain foods and is differentiated according to the region, locality, product, and year.