Added Reference Links

explainers/ai-objectivity-myth.html

@@ -112,7 +112,7 @@
         "url": "https://www.thefaircode.xyz/index.html#explainers"
       },
       "datePublished": "2026-06-14",
-      "dateModified": "2026-06-30"
+      "dateModified": "2026-08-17"
     },
     {
       "@type": "FAQPage",
@@ -285,7 +285,7 @@ <h2 id="related-concepts">Related Concepts</h2>
 <h2 id="related-projects-in-this-repo">Related Projects in This Repo</h2>
 <ul><li><a href="../index.html#project-compas"><code>COMPAS/</code></a> - the highest-profile example of a &quot;neutral risk score&quot; actively used in court decisions, with an 86.77% fairness gap before mitigation.</li><li><a href="../index.html#project-benefits"><code>Benefits Denial/</code></a> - a welfare eligibility model where &quot;objective&quot; income and household features encode sex, race, and national origin simultaneously.</li><li><a href="../index.html#project-readmission"><code>Healthcare Readmission/</code></a> - clinical &quot;risk scores&quot; that look like medicine but encode insurance access and discharge-destination disparities.</li></ul>
 <h2 id="further-reading">Further Reading</h2>
-<ul><li><a href="https://www.penguinrandomhouse.com/books/241363/weapons-of-math-destruction-by-cathy-oneil/" target="_blank" rel="noreferrer noopener">O&#x27;Neil, C. (2016): Weapons of Math Destruction</a> - the foundational case for why &quot;it&#x27;s just an algorithm&quot; is not a defense, with examples spanning credit, employment, and criminal justice.</li><li><a href="https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing" target="_blank" rel="noreferrer noopener">Angwin, J. et al. (2016): Machine Bias, ProPublica</a> - the original investigation showing COMPAS scores were presented as objective risk assessments while producing racially disparate outcomes.</li><li>Barocas, S., Hardt, M., &amp; Narayanan, A.: <em>Fairness and Machine Learning</em> (fairmlbook.org) - the standard reference for why statistical models trained on historical data inherit the properties of that history, including its inequities.</li></ul>
+<ul><li><a href="https://www.penguinrandomhouse.com/books/241363/weapons-of-math-destruction-by-cathy-oneil/" target="_blank" rel="noreferrer noopener">O&#x27;Neil, C. (2016): Weapons of Math Destruction</a> - the foundational case for why &quot;it&#x27;s just an algorithm&quot; is not a defense, with examples spanning credit, employment, and criminal justice.</li><li><a href="https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing" target="_blank" rel="noreferrer noopener">Angwin, J. et al. (2016): Machine Bias, ProPublica</a> - the original investigation showing COMPAS scores were presented as objective risk assessments while producing racially disparate outcomes.</li><li><a href="https://fairmlbook.org" target="_blank" rel="noreferrer noopener">Barocas, S., Hardt, M., &amp; Narayanan, A.: <em>Fairness and Machine Learning</em> (fairmlbook.org)</a> - the standard reference for why statistical models trained on historical data inherit the properties of that history, including its inequities.</li></ul>
 <hr>
 <p><em>Part of <a href="https://instagram.com/thefaircodeproject" target="_blank" rel="noreferrer noopener">The Fair Code Project</a> - exposing and fixing algorithmic bias with real data and open code.</em></p></article>
   </main>

explainers/ai-objectivity-myth.md

@@ -141,7 +141,7 @@ Even after a model is shown not to be objective, the institutional response is o
 
 * [O'Neil, C. (2016): Weapons of Math Destruction](https://www.penguinrandomhouse.com/books/241363/weapons-of-math-destruction-by-cathy-oneil/) - the foundational case for why "it's just an algorithm" is not a defense, with examples spanning credit, employment, and criminal justice.
 * [Angwin, J. et al. (2016): Machine Bias, ProPublica](https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing) - the original investigation showing COMPAS scores were presented as objective risk assessments while producing racially disparate outcomes.
-* Barocas, S., Hardt, M., & Narayanan, A.: *Fairness and Machine Learning* (fairmlbook.org) - the standard reference for why statistical models trained on historical data inherit the properties of that history, including its inequities.
+* [Barocas, S., Hardt, M., & Narayanan, A.: *Fairness and Machine Learning* (fairmlbook.org)](https://fairmlbook.org) - the standard reference for why statistical models trained on historical data inherit the properties of that history, including its inequities.
 
 ---
 

explainers/base-rate-fallacy.html

@@ -112,7 +112,7 @@
         "url": "https://www.thefaircode.xyz/index.html#explainers"
       },
       "datePublished": "2026-08-14",
-      "dateModified": "2026-08-14"
+      "dateModified": "2026-08-17"
     },
     {
       "@type": "FAQPage",
@@ -343,7 +343,7 @@ <h2 id="related-concepts">Related Concepts</h2>
 <h2 id="related-projects-in-this-repo">Related Projects in This Repo</h2>
 <ul><li><a href="../index.html#project-compas"><code>COMPAS/</code></a> - recidivism risk scoring audit demonstrating the real-world clash between predictive parity and equalized odds driven by racial base rate differences.</li><li><a href="../index.html#project-readmission"><code>Healthcare Readmission/</code></a> - clinical readmission model where base rate differences in hospital access corrupt risk predictions across insurance types.</li></ul>
 <h2 id="further-reading">Further Reading</h2>
-<ul><li>Bar-Hillel, M. (1980): The Base-Rate Fallacy in Probability Judgments, <em>Acta Psychologica</em>, 44(3), 211-233 - the foundational cognitive psychology paper establishing how humans ignore prior probabilities.</li><li><a href="https://arxiv.org/abs/1610.07524" target="_blank" rel="noreferrer noopener">Chouldechova, A. (2017): Fair Prediction with Disparate Impact</a> - the formal proof establishing the mathematical impossibility of satisfying predictive parity and equalized odds under unequal base rates.</li><li><a href="https://arxiv.org/abs/1609.05807" target="_blank" rel="noreferrer noopener">Kleinberg, J., Mullainathan, S., Raghavan, M. (2017): Inherent Trade-Offs in the Fair Determination of Risk Scores</a> - independent proof of the impossibility theorem for calibrated continuous scores.</li><li><a href="https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing" target="_blank" rel="noreferrer noopener">Angwin, J. et al. (2016): Machine Bias</a> - ProPublica&#x27;s seminal investigation into COMPAS error-rate disparities.</li></ul>
+<ul><li><a href="https://www.researchgate.net/publication/223684493_The_base-rate_fallacy_in_probability_judgments" target="_blank" rel="noreferrer noopener">Bar-Hillel, M. (1980): The Base-Rate Fallacy in Probability Judgments, <em>Acta Psychologica</em>, 44(3), 211-233</a> - the foundational cognitive psychology paper establishing how humans ignore prior probabilities.</li><li><a href="https://arxiv.org/abs/1610.07524" target="_blank" rel="noreferrer noopener">Chouldechova, A. (2017): Fair Prediction with Disparate Impact</a> - the formal proof establishing the mathematical impossibility of satisfying predictive parity and equalized odds under unequal base rates.</li><li><a href="https://arxiv.org/abs/1609.05807" target="_blank" rel="noreferrer noopener">Kleinberg, J., Mullainathan, S., Raghavan, M. (2017): Inherent Trade-Offs in the Fair Determination of Risk Scores</a> - independent proof of the impossibility theorem for calibrated continuous scores.</li><li><a href="https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing" target="_blank" rel="noreferrer noopener">Angwin, J. et al. (2016): Machine Bias</a> - ProPublica&#x27;s seminal investigation into COMPAS error-rate disparities.</li></ul>
 <hr>
 <p><em>Part of <a href="https://instagram.com/thefaircodeproject" target="_blank" rel="noreferrer noopener">The Fair Code Project</a> - exposing and fixing algorithmic bias with real data and open code.</em></p></article>
   </main>

explainers/base-rate-fallacy.md

@@ -225,7 +225,7 @@ Attempting to force equal false positive rates by adjusting decision thresholds
 
 ## Further Reading
 
-* Bar-Hillel, M. (1980): The Base-Rate Fallacy in Probability Judgments, *Acta Psychologica*, 44(3), 211-233 - the foundational cognitive psychology paper establishing how humans ignore prior probabilities.
+* [Bar-Hillel, M. (1980): The Base-Rate Fallacy in Probability Judgments, *Acta Psychologica*, 44(3), 211-233](https://www.researchgate.net/publication/223684493_The_base-rate_fallacy_in_probability_judgments) - the foundational cognitive psychology paper establishing how humans ignore prior probabilities.
 * [Chouldechova, A. (2017): Fair Prediction with Disparate Impact](https://arxiv.org/abs/1610.07524) - the formal proof establishing the mathematical impossibility of satisfying predictive parity and equalized odds under unequal base rates.
 * [Kleinberg, J., Mullainathan, S., Raghavan, M. (2017): Inherent Trade-Offs in the Fair Determination of Risk Scores](https://arxiv.org/abs/1609.05807) - independent proof of the impossibility theorem for calibrated continuous scores.
 * [Angwin, J. et al. (2016): Machine Bias](https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing) - ProPublica's seminal investigation into COMPAS error-rate disparities.

explainers/false-positives-vs-false-negatives.html

@@ -112,7 +112,7 @@
         "url": "https://www.thefaircode.xyz/index.html#explainers"
       },
       "datePublished": "2026-07-10",
-      "dateModified": "2026-07-10"
+      "dateModified": "2026-08-17"
     },
     {
       "@type": "FAQPage",
@@ -293,7 +293,7 @@ <h2 id="related-concepts">Related Concepts</h2>
 <h2 id="related-projects-in-this-repo">Related Projects in This Repo</h2>
 <ul><li><a href="../index.html#project-readmission"><code>Healthcare Readmission/</code></a> - the primary audit behind the concrete example above: readmission risk prediction with race and age as protected attributes.</li><li><a href="../index.html#project-insurance"><code>Insurance Denial/</code></a> - a second health-adjacent audit where the cost of a false negative (an undetected high-risk claim) and a false positive (a wrongly denied low-risk claim) are similarly asymmetric.</li></ul>
 <h2 id="further-reading">Further Reading</h2>
-<ul><li><a href="https://www.science.org/doi/10.1126/science.aax2342" target="_blank" rel="noreferrer noopener">Obermeyer, Z., Powers, B., Vogeli, C., Mullainathan, S. (2019): Dissecting racial bias in an algorithm used to manage the health of populations</a> - the study behind the concrete example above; shows how a false-negative-heavy error pattern hid behind an apparently race-neutral risk score.</li><li><a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC6594166/" target="_blank" rel="noreferrer noopener">Rajkomar, A., Hardt, M., Howell, M.D., Corrado, G., Chin, M.H. (2018): Ensuring Fairness in Machine Learning to Advance Health Equity</a> - lays out how model design, data, and clinician interaction each shape which patients absorb a clinical model&#x27;s errors.</li><li>Chouldechova, A. (2017): Fair Prediction with Disparate Impact - the proof that equalized false positive and false negative rates and equal predictive value cannot all hold at once when groups have different base rates, the statistical root of the threshold trade-off described above.</li></ul>
+<ul><li><a href="https://www.science.org/doi/10.1126/science.aax2342" target="_blank" rel="noreferrer noopener">Obermeyer, Z., Powers, B., Vogeli, C., Mullainathan, S. (2019): Dissecting racial bias in an algorithm used to manage the health of populations</a> - the study behind the concrete example above; shows how a false-negative-heavy error pattern hid behind an apparently race-neutral risk score.</li><li><a href="https://pmc.ncbi.nlm.nih.gov/articles/PMC6594166/" target="_blank" rel="noreferrer noopener">Rajkomar, A., Hardt, M., Howell, M.D., Corrado, G., Chin, M.H. (2018): Ensuring Fairness in Machine Learning to Advance Health Equity</a> - lays out how model design, data, and clinician interaction each shape which patients absorb a clinical model&#x27;s errors.</li><li><a href="https://arxiv.org/pdf/1703.0056" target="_blank" rel="noreferrer noopener">Chouldechova, A. (2017): Fair Prediction with Disparate Impact</a> - the proof that equalized false positive and false negative rates and equal predictive value cannot all hold at once when groups have different base rates, the statistical root of the threshold trade-off described above.</li></ul>
 <p><em>Part of <a href="https://instagram.com/thefaircodeproject" target="_blank" rel="noreferrer noopener">The Fair Code Project</a> - exposing and fixing algorithmic bias with real data and open code.</em></p></article>
   </main>
 

explainers/false-positives-vs-false-negatives.md

@@ -151,6 +151,6 @@ Even a model with perfectly equalized false positive and false negative rates on
 
 * [Obermeyer, Z., Powers, B., Vogeli, C., Mullainathan, S. (2019): Dissecting racial bias in an algorithm used to manage the health of populations](https://www.science.org/doi/10.1126/science.aax2342) - the study behind the concrete example above; shows how a false-negative-heavy error pattern hid behind an apparently race-neutral risk score.
 * [Rajkomar, A., Hardt, M., Howell, M.D., Corrado, G., Chin, M.H. (2018): Ensuring Fairness in Machine Learning to Advance Health Equity](https://pmc.ncbi.nlm.nih.gov/articles/PMC6594166/) - lays out how model design, data, and clinician interaction each shape which patients absorb a clinical model's errors.
-* Chouldechova, A. (2017): Fair Prediction with Disparate Impact - the proof that equalized false positive and false negative rates and equal predictive value cannot all hold at once when groups have different base rates, the statistical root of the threshold trade-off described above.
+* [Chouldechova, A. (2017): Fair Prediction with Disparate Impact](https://arxiv.org/pdf/1703.0056) - the proof that equalized false positive and false negative rates and equal predictive value cannot all hold at once when groups have different base rates, the statistical root of the threshold trade-off described above.
 
 *Part of [The Fair Code Project](https://instagram.com/thefaircodeproject) - exposing and fixing algorithmic bias with real data and open code.*

explainers/precision-recall-curve.html

@@ -300,7 +300,7 @@ <h2 id="related-concepts">Related Concepts</h2>
 <h2 id="related-projects-in-this-repo">Related Projects in This Repo</h2>
 <ul><li><a href="../index.html#project-readmission"><code>Healthcare Readmission/</code></a> - the audit behind the concrete example above, where an 11.2% base rate turns a &quot;reasonable&quot; AUC into a near-worthless F1.</li><li><a href="../index.html#project-benefits"><code>Benefits Denial/</code></a> - a second audit with a skewed target (24% positive rate) where the same ROC-vs-PR gap is worth checking.</li></ul>
 <h2 id="further-reading">Further Reading</h2>
-<ul><li><a href="https://dl.acm.org/doi/10.1145/1143844.1143874" target="_blank" rel="noreferrer noopener">Davis, J., Goadrich, M. (2006): The Relationship Between Precision-Recall and ROC Curves</a> - the paper that formally connects the two curve families and proves a curve dominating in ROC space also dominates in PR space, while showing why their <em>visual</em> impression of &quot;how good&quot; a model looks can differ sharply under imbalance.</li><li><a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0118432" target="_blank" rel="noreferrer noopener">Saito, T., Rehmsmeier, M. (2015): The Precision-Recall Plot Is More Informative than the ROC Plot When Evaluating Binary Classifiers on Imbalanced Datasets</a> - a direct empirical demonstration of the exact mechanism this explainer describes: ROC curves that look consistently strong across imbalance levels while the matching PR curves reveal deteriorating precision.</li><li>Barocas, S., Hardt, M., Narayanan, A. (2019): <em>Fairness and Machine Learning</em> - the classification chapter&#x27;s discussion of base-rate differences across groups applies directly here: a shared threshold cannot equalize precision across groups with different base rates without giving something else up.</li></ul>
+<ul><li><a href="https://dl.acm.org/doi/10.1145/1143844.1143874" target="_blank" rel="noreferrer noopener">Davis, J., Goadrich, M. (2006): The Relationship Between Precision-Recall and ROC Curves</a> - the paper that formally connects the two curve families and proves a curve dominating in ROC space also dominates in PR space, while showing why their <em>visual</em> impression of &quot;how good&quot; a model looks can differ sharply under imbalance.</li><li><a href="https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0118432" target="_blank" rel="noreferrer noopener">Saito, T., Rehmsmeier, M. (2015): The Precision-Recall Plot Is More Informative than the ROC Plot When Evaluating Binary Classifiers on Imbalanced Datasets</a> - a direct empirical demonstration of the exact mechanism this explainer describes: ROC curves that look consistently strong across imbalance levels while the matching PR curves reveal deteriorating precision.</li><li><a href="https://fairmlbook.org/classification.html" target="_blank" rel="noreferrer noopener">Barocas, S., Hardt, M., Narayanan, A. (2019): <em>Fairness and Machine Learning</em></a> - the classification chapter&#x27;s discussion of base-rate differences across groups applies directly here: a shared threshold cannot equalize precision across groups with different base rates without giving something else up.</li></ul>
 <p><em>Part of <a href="https://instagram.com/thefaircodeproject" target="_blank" rel="noreferrer noopener">The Fair Code Project</a> - exposing and fixing algorithmic bias with real data and open code.</em></p></article>
   </main>
 

explainers/precision-recall-curve.md

@@ -165,6 +165,6 @@ Two groups can have identical AP while one group's curve reaches that area throu
 
 * [Davis, J., Goadrich, M. (2006): The Relationship Between Precision-Recall and ROC Curves](https://dl.acm.org/doi/10.1145/1143844.1143874) - the paper that formally connects the two curve families and proves a curve dominating in ROC space also dominates in PR space, while showing why their *visual* impression of "how good" a model looks can differ sharply under imbalance.
 * [Saito, T., Rehmsmeier, M. (2015): The Precision-Recall Plot Is More Informative than the ROC Plot When Evaluating Binary Classifiers on Imbalanced Datasets](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0118432) - a direct empirical demonstration of the exact mechanism this explainer describes: ROC curves that look consistently strong across imbalance levels while the matching PR curves reveal deteriorating precision.
-* Barocas, S., Hardt, M., Narayanan, A. (2019): *Fairness and Machine Learning* - the classification chapter's discussion of base-rate differences across groups applies directly here: a shared threshold cannot equalize precision across groups with different base rates without giving something else up.
+* [Barocas, S., Hardt, M., Narayanan, A. (2019): *Fairness and Machine Learning*](https://fairmlbook.org/classification.html) - the classification chapter's discussion of base-rate differences across groups applies directly here: a shared threshold cannot equalize precision across groups with different base rates without giving something else up.
 
 *Part of [The Fair Code Project](https://instagram.com/thefaircodeproject) - exposing and fixing algorithmic bias with real data and open code.*