Stata Data Management and Analysis Training Course

Participants can import common source files and assess their initial quality.

• Importing Excel worksheets with import excel and sheet selection
• Reading CSV and delimited files using import delimited options
• Loading Stata datasets, preserving labels and stored value formats
• Checking observations, variables, storage types and missing-value patterns
• Identifying duplicate identifiers through isid, duplicates report and tag
• Converting string and numeric variables without losing source meaning

Participants can create analysis-ready datasets through controlled cleaning and recoding.

• Missing-value codes distinguished from valid zero and blank responses
• Recoding categorical values with recode, replace and generate
• String functions for trimming, standardising and extracting text components
• Date conversion using date(), daily() and display format controls
• Variable and value labels maintained through label commands
• Data cleaning decisions recorded in comments and transformation logs

Participants can generate derived measures and prepare records for analysis.

• Generating ratios, rates, percentages and indexed performance measures
• Conditional logic using if, inrange(), cond() and missing()
• Using egen for grouped totals, means, ranks and identifiers
• Standardising variables with z-scores and centring transformations
• Creating indicator variables from categorical and threshold-based conditions
• Validation checks comparing derived values with source business rules

Participants can combine related datasets and resolve matching problems.

• Appending monthly extracts with append and source-file identifiers
• One-to-one merges using stable keys and merge result codes
• Many-to-one merges for transactions, reference tables and master records
• Resolving unmatched observations through identifier and date investigations
• Reshaping wide and long data with reshape long and reshape wide
• Collapsing transaction records into customer, branch and period summaries

Participants can automate repeated tasks using Stata programming features.

• Local macros for file names, variable lists and parameters
• foreach loops applied to repeated recoding and reporting tasks
• forvalues loops for numbered variables and period-based processing
• Temporary files and preserve restore for controlled intermediate steps
• Script sections organised around import, cleaning, analysis and outputs
• Debugging code with trace, assertions and incremental execution

Participants can summarise distributions, relationships and group differences accurately.

• Summary statistics using summarize, detail and tabstat commands
• Frequency tables, proportions and missingness using tabulate options
• Cross-tabulations with row, column and cell percentage displays
• Grouped summaries using bysort, collapse and table commands
• Distribution inspection through centiles, skewness and percentile comparisons
• Outlier review using box plots, scatterplots and logical thresholds

Participants can create clear Stata graphics that communicate analytical findings.

• Histograms and kernel density plots for distribution assessment
• Bar charts comparing categories, proportions and grouped summary values
• Scatterplots with fitted lines for continuous-variable relationships
• Line charts showing monthly trends, seasonality and intervention timing
• Graph titles, legends, axis scales and colour choices
• Exporting graphics in PNG, SVG and publication-quality formats

Participants can estimate uncertainty and test defined analytical questions.

• Sampling distributions, standard errors and confidence interval interpretation
• One-sample and two-sample t-tests for mean comparisons
• Paired t-tests for before-and-after measurement designs
• Proportion tests for rates, uptake and compliance measures
• Chi-square tests for association between categorical variables
• Statistical significance, practical significance and decision thresholds

Participants can estimate, diagnose and explain linear regression models.

• Ordinary least squares assumptions and fitted-value interpretation
• Specifying regressors using factor variables and interaction terms
• Interpreting coefficients, confidence intervals, p-values and adjusted R-squared
• Residual plots for linearity, variance and influential observation checks
• Multicollinearity assessment using variance inflation factor diagnostics
• Heteroskedasticity-consistent standard errors using vce(robust) options

Participants can model binary and count outcomes using appropriate regression methods.

• Logistic regression for default, attrition, compliance and treatment outcomes
• Odds ratios, predicted probabilities and marginal effects interpretation
• Classification tables, sensitivity and specificity for binary models
• Poisson regression for incident counts and event frequencies
• Negative binomial models for overdispersed count data
• Model comparison using likelihood statistics and information criteria

Participants can analyse repeated observations across entities and time periods.

• Panel structures defined by entity identifiers and time variables
• Declaring panels with xtset and checking temporal gaps
• Fixed-effects estimation for unobserved time-invariant entity characteristics
• Random-effects estimation and assumptions behind variance component models
• Hausman testing for fixed-effects and random-effects model selection
• Interpreting within, between and overall variation in panel output

Participants can prepare time-series data and build basic forecasting models.

• Time variables converted, formatted and declared using tsset
• Lag, lead and difference operators for temporal transformations
• Trend, seasonality and structural breaks assessed through time plots
• Autocorrelation functions and partial autocorrelation function interpretation
• ARIMA model components for stationary univariate series
• Forecast evaluation using holdout periods and forecast error measures

Participants can analyse survey data while accounting for sample design.

• Survey weights, strata and primary sampling unit concepts
• Declaring survey designs using svyset command specifications
• Weighted proportions, means and totals using svy estimation
• Design-based standard errors compared with unweighted estimation
• Subpopulation analysis using subpop() rather than restrictive filtering
• Survey result tables suitable for policy and programme reporting

Participants can quality-check Stata work and prepare transparent analytical outputs.

• Assertions and cross-checks for record counts and valid ranges
• Data signatures and checksums for source-file change detection
• Code review criteria covering clarity, dependencies and reproducibility
• Estimation result storage using estimates store and restore
• Regression tables exported with esttab and estout formatting options
• Analytical notes documenting assumptions, exclusions and interpretation limits

Participants can translate a decision question into a documented Stata analysis plan.

• Decision questions translated into outcomes, predictors and comparison groups
• Analysis plans defining datasets, inclusion rules and statistical methods
• Data dictionaries linking business terms to variable definitions
• Results narratives separating findings, limitations and recommended actions
• Chart selection matched to comparison, trend, distribution and relationship questions
• Executive summaries built from evidence, implications and next steps

Participants can build, review and present a complete Stata analysis from source data to recommendations.

• Analytical brief scoped around a defined operational or policy question
• Source data imported, cleaned and documented within a do-file
• Derived dataset constructed with validation checks and reproducible transformations
• Statistical model selected, estimated and tested against stated assumptions
• Findings presented through tables, graphics and concise interpretation
• Peer review addresses code quality, evidence and decision recommendations