Educational App
Economic Dataset Generator
Generate fictitious cross-sectional or longitudinal datasets for economics exercises.
1. Dataset structure
2. Select variables
Search and select the variables to include. Selected variables will appear below with individual settings.
When enabled, the app selects all relevant variables, locks individual settings, and generates more plausible economic relationships.
Selected variables
| Variable | Type | Missing % | Distribution | |
|---|---|---|---|---|
| No variables selected yet. | ||||
3. Global settings
Maximum: 7000 observations.
Allowed range: 2 to 4 periods.
CSV is compatible with Excel, Stata, R, and Python.
4. Visual evidence of assumptions
This section uses a fixed dataset generated by this tool: cross-sectional data, people as the unit of analysis, 7,000 observations, no missing values, and relationships based on economic evidence.
a) Unbiasedness
An estimator is unbiased when, across repeated samples, its expected value is equal to the true parameter. In other words, it does not systematically overestimate or underestimate the parameter.
E[θ̂] = θ
The visual intuition uses the bivariate specification:
Yᵢ = β₀ + β₁Xᵢ + uᵢ
- Yᵢ: Annual income
- Xᵢ: Education years
The app draws 500 bootstrap samples with replacement, each with n = 200. The green curve shows the probability density of β₁ estimates. The black vertical line is the β₁ estimated with the full dataset of 7,000 observations. The red vertical line is the average β₁ across bootstrap samples.
b) Consistency
An estimator is consistent when it converges in probability to the true parameter as the sample size increases.
plim θ̂ₙ = θ as n → ∞
The visual intuition uses the same bivariate specification and compares bootstrap distributions for different sample sizes: n = 200, n = 400, and n = 600. If the estimator is consistent, the distribution should become more concentrated around the full-sample estimate as n increases.
c) Efficiency
An estimator is efficient when it has the smallest variance among a defined class of unbiased estimators. Under the Gauss-Markov assumptions, OLS is BLUE: Best Linear Unbiased Estimator.
Var(θ̂ efficient) ≤ Var(θ̃)
A visual comparison for efficiency will be added later.