Weak Instrument Detection Bootstrap Test with Monte Carlo Simulations

A bootstrap-based test for detecting weak instruments in instrumental variable regression, developed by Zhaoguo Zhan. Monte Carlo simulations indicate the test has good size and power. The methodology is illustrated with an application analyzing the return to schooling.

Use Cases

• Evaluating instrument strength in econometric models based on the bootstrap test described.
• Simulating test performance under different conditions based on the Monte Carlo experiments mentioned.
• Applying the test to real-world regression problems, such as estimating returns to schooling as illustrated.
• Comparing bootstrap distributions to standard normal distributions to diagnose weak identification.

Strengths

• The methodology includes Monte Carlo simulations demonstrating test performance.
• The test is illustrated with a concrete application from Card's work.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset consists primarily of source files (PDF, BBL, TEX, EPS, BIB, PNG) rather than structured data tables.

Provenance

Source

figshare

Collection Method

The dataset likely contains files related to the development and documentation of a statistical test.

Freshness

Last updated 2026-05-25 13:55:42; freshness should be verified.