Sample Selection Bias as a Specification Error¶
Heckman, J. J. (1979). Sample Selection Bias as a Specification Error. Econometrica, 47(1), 153-161.
Cited by¶
5 citations across 5 artifacts.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Confounding
This sourceFrames selection bias as an omitted-variable specification error and introduces the two-step (Heckman) correction; relevant to the selection-vs-confounding distinction. (Bibliography only — link added.)
- Selection Bias
- The concept spans both experimental-design/statistics origins (Berkson's 1946 hospital-bias recognition, Neyman's sampling-selection discussion) and econometrics origins (Heckman's 1979 formal treatment and 2000 Nobel-winning sample-selection model), making it a dual-origin principle essential to causal inference, observational research, and meta-analysis
This sourceHeckman sample selection bias econometrics labor market wage estimation.
- The concept spans both experimental-design/statistics origins (Berkson's 1946 hospital-bias recognition, Neyman's sampling-selection discussion) and econometrics origins (Heckman's 1979 formal treatment and 2000 Nobel-winning sample-selection model), making it a dual-origin principle essential to causal inference, observational research, and meta-analysis
- Vantage-Induced Omission
- Selection bias names the case where a samplable selection mechanism distorts data; vantage-induced omission generalises to institutional, instrumental, and positional selection, including cases where no samplable selection function exists at all — which stories a newsroom can reach is not a probabilistic sample.
This sourceSelection bias as a samplable selection mechanism distorting inference, the formal-statistical special case the prime generalizes beyond.
- Selection bias names the case where a samplable selection mechanism distorts data; vantage-induced omission generalises to institutional, instrumental, and positional selection, including cases where no samplable selection function exists at all — which stories a newsroom can reach is not a probabilistic sample.
Mechanisms¶
- Holdout Ground-Truth Audit
- If the holdout is not truly shielded — if audited cases are, deliberately or not, easier ones, or ones the proxy already flagged — the audit inherits the proxy's blind spots and flatters it; this is selection bias, and it is the quiet killer of audit programs.
This sourceIdentifies nonrandom sample selection as a source of biased inference.
- If the holdout is not truly shielded — if audited cases are, deliberately or not, easier ones, or ones the proxy already flagged — the audit inherits the proxy's blind spots and flatters it; this is selection bias, and it is the quiet killer of audit programs.
- Revealed Preference Choice Log
- Its deepest failure mode is survivorship bias: the log only sees choices that were made
This sourceShows that behavioral outcomes can be observed only for units whose self-selection decision places them in the recorded sample.
- Its deepest failure mode is survivorship bias: the log only sees choices that were made
Verification¶
This reference passed the adversarial substantiation pipeline: it was checked to exist and to support the claim it is attached to. See how references were verified.
Registry ID ref:a5c0a93f31bc · see in the full table