Validity

Option	Description
External	High confidence that findings generalize to other populations or settings.
Moderate confidence	Some ability to generalize, but caution is needed.
Internal	Strong bias control within the study but limited generalizability.
Varying	Validity depends heavily on context, data quality, or model specification.

Definition

The degree to which the method yields unbiased and reliable results (internal validity) and the extent to which those results can be generalized beyond the study sample (external validity).

Explanation

Understanding validity helps you weigh trade‑offs: a method with strong internal validity minimizes bias in effect estimation, while external validity determines whether findings can be transported to other settings, populations, or contexts. Some methods excel in one dimension but may suffer in the other.

Tools/rationale for helping assessment

• Run internal bias checks (see placebo or falsification tests).
• Test external validity by applying the method to held-out or independent data.
• Conduct simulation studies under known biases when possible.

Example

In a similar RS-based deforestation attribution, placebo tests on areas without logging yield no effect (high internal validity), but applying to a different country produces different magnitudes. A reasonable target validity for my case study is set to Moderate confidence.