Code, interpretation and assumptions
Read the do-file alongside its explanation, then connect the commands to the browser labs. Each example keeps its complete downloadable source.
Reading here does not run the script. Full Stata is required to execute these files. They clear the active dataset, so save existing work before running one. Use a separate folder for any exported figures.
Begin with guided browser practice →After the beginner path
Construct a coffee-sales dataset and read coefficients, fitted values and uncertainty in their stated units.
Intermediate
Compare a misleading association with an explicitly constructed instrument in a training example.
Intermediate
Examine OLS and IV in one known process, then change the strength of confounding.
Intermediate
Inspect probability scales, control timing, ratio outcomes, pooled comparisons and instrument relevance.
Intermediate
Inspect OLS, panel and IV postestimation commands, their maintained assumptions and their limits.
Intermediate to advanced
Use factor variables, marginal effects and prediction grids to examine two- and three-way interactions.
Advanced
Separate within- and between-firm variation, compare covariance assumptions and inspect a distinct dynamic-panel example.
Compare outcome models, proposed designs and inference assumptions. The printable references describe their conditions without changing Stata data.
Full method reference (.do) · Compact reference (.do)Full Stata is required to run this file. It clears the active dataset. Run the complete file first; sections share data and local macros.
Computed in native Stata from the bundled file and its fixed seeds. These are reference results, not a model running in this page.