phenomenon · claim-type · cross-domain-transfer · mechanism · claim-relation · replication-status · limitation — grounded in the full text| phenomenon | a young field in which one word carries several formal meanings that cannot all be satisfied at once, so that two competing analyses of the same tool can both be arithmetically correct— “Their most discussed finding was that COMPAS does not satisfy equal false positive rates by race: Among defendants who did not get rearrested, black defendants were twice as likely to be misclassified as high risk. Based largely on this and other similar findings, they described the tool as biased against blacks. Northpointe (now Equivant)… | |
| claim-type | critique— “Much of the subsequent conversation consisted of either trying to harmonize these definitions of fairness or asserting that one or the other is correct. As it turns out, there can be no harmony among definitions in a world where inequality and imperfect prediction are the reality.” | |
| method — evidence | — “The goal of this article is not to advance axiomatic definitions of fairness but to summarize the definitions and results in this area that have been formalized to date.” | |
| object of study — evidence | — “The rapid growth of this new field has led to wildly inconsistent motivations, terminology, and notation, presenting a serious challenge for cataloging and comparing definitions. This article attempts to bring much-needed order.” | |
| research question — evidence | — “The goal of this article is not to advance axiomatic definitions of fairness but to summarize the definitions and results in this area that have been formalized to date.” | |
| contribution — evidence | — “Although each flavor of fairness definition presented in this section formalizes an intuitive notion of fairness, these definitions are not mathematically or morally compatible in general. In this section, we review several impossibility results about definitions of fairness” | |
| mechanism | THE INCOMPATIBILITY IS NOT A PROPERTY OF ALGORITHMS. `the trade-off between different notions of fairness would apply even if a human were the one making the prediction`, and `Rejecting model-driven or automated decision-making is not a way to avoid these problems`. It is a property of predicting at all, under inequality— “Notably, scholars have emphasized that the trade-off between different notions of fairness would apply even if a human were the one making the prediction (Kleinberg et al. 2016). Rejecting model-driven or automated decision-making is not a way to avoid these problems.” | mechanisms[0] |
| mechanism | and the conditions under which the definitions COULD agree are named and are both unreachable: equal base rates or perfect prediction. The paper calls them `trivial, degenerate, or even utopian (representing two very different utopias)`— “Equal base rates and perfect prediction can be called trivial, degenerate, or even utopian (representing two very different utopias).” | mechanisms[1] |
| contradicts | both positions in the COMPAS dispute, insofar as either claimed to be THE fairness answer - 'there can be no harmony among definitions in a world where inequality and imperfect prediction are the reality'— “Much of the subsequent conversation consisted of either trying to harmonize these definitions of fairness or asserting that one or the other is correct. As it turns out, there can be no harmony among definitions in a world where inequality and imperfect prediction are the reality.” | claim_relations[0] |
| extends | the impossibility results of Kleinberg et al. and Chouldechova, restated here in one notation and set against a public case— “Putting this in context and recalling that sufficiency implies equality of positive predictive values and that separation implies equality of false positive rates, this result shows that both supported definitions of fairness in the COMPAS debate can only be met when (a) the rate of recidivism and the distribution of scores are the same for… | claim_relations[1] |
| supports | the psychometric literature on test fairness running back to the 1960s, named as the older form of the same question— “Though the algorithmic fairness conversation is somewhat new, it resembles older work. For example, since the 1960s, psychometricians have studied the fairness of educational tests based on their ability to predict performance” | claim_relations[2] |
| sense | fairness - NOT ONE DEFINITION BUT SEVERAL FORMAL ONES, AND THE PAPER'S RESULT IS THAT THEY CANNOT ALL HOLD: `Although each flavor of fairness definition ... formalizes an intuitive notion of fairness, these definitions are not mathematically or morally compatible in general`— “Although each flavor of fairness definition presented in this section formalizes an intuitive notion of fairness, these definitions are not mathematically or morally compatible in general. In this section, we review several impossibility results about definitions of fairness” | senses[0] |
| sense | `bias` - THE PAPER CALLS IT OVERLOADED AND GIVES THE DRIFT: `a now overloaded word that, in popular media, has come to mean that the model's performance (however defined) unjustifiably differs along social axes such as race, gender, and class`. The parenthesis is the point— “In recent years, attention has focused on how consequential predictive models may be biased—a now overloaded word that, in popular media, has come to mean that the model's performance (however defined) unjustifiably differs along social axes such as race, gender, and class.” | senses[1] |
| sense | THE COMPAS DISPUTE - two analyses of one tool, both correct. ProPublica: it `does not satisfy equal false positive rates by race`. Northpointe: it `satisfies equal positive predictive values` and calibration within groups. NEITHER SIDE MISCALCULATED. They used different definitions of the same word— “Their most discussed finding was that COMPAS does not satisfy equal false positive rates by race: Among defendants who did not get rearrested, black defendants were twice as likely to be misclassified as high risk. Based largely on this and other similar findings, they described the tool as biased against blacks. Northpointe (now Equivant)… | senses[2] |
| sense | `no harmony` - the paper's verdict on the attempts to reconcile them: `Much of the subsequent conversation consisted of either trying to harmonize these definitions ... or asserting that one or the other is correct. As it turns out, there can be no harmony among definitions in a world where inequality and imperfect prediction are the reality`— “Much of the subsequent conversation consisted of either trying to harmonize these definitions of fairness or asserting that one or the other is correct. As it turns out, there can be no harmony among definitions in a world where inequality and imperfect prediction are the reality.” | senses[3] |
| sense | the formal conditions for agreement, stated exactly: SUFFICIENCY implies equality of positive predictive values, SEPARATION implies equality of false positive rates, and both hold together only if recidivism rates and score distributions are the same across racial groups, or some group never experiences an outcome at all— “Putting this in context and recalling that sufficiency implies equality of positive predictive values and that separation implies equality of false positive rates, this result shows that both supported definitions of fairness in the COMPAS debate can only be met when (a) the rate of recidivism and the distribution of scores are the same for… | senses[4] |
| sense | algorithmic fairness - a name for a research field, and the paper notes the questions are older than it: `it resembles older work ... since the 1960s, psychometricians have studied the fairness of educational tests based on their ability to predict performance`— “Though the algorithmic fairness conversation is somewhat new, it resembles older work. For example, since the 1960s, psychometricians have studied the fairness of educational tests based on their ability to predict performance” | senses[5] |
| sense | equality of true positive rates - equality across groups in the true-positive rate— “The second pair, equality of true positive rates and equality of false negative rates, is equivalent to D⊥A|Y = 1. This definition of fairness takes the perspective of those individuals with outcome defined by one. For example, this is equivalent to the requirement that all individuals who will go on to repay a loan have the same likelihood… | senses[6] |
| sense | equality of false negative rates - equality across groups in the false-negative rate— “The second pair, equality of true positive rates and equality of false negative rates, is equivalent to D⊥A|Y = 1. This definition of fairness takes the perspective of those individuals with outcome defined by one. For example, this is equivalent to the requirement that all individuals who will go on to repay a loan have the same likelihood… | senses[7] |
| sense | equal opportunity - the paper's name for equality of true-positive and false-negative rates considered alone— “The second pair, equality of true positive rates and equality of false negative rates, is equivalent to D⊥A|Y = 1. This definition of fairness takes the perspective of those individuals with outcome defined by one. For example, this is equivalent to the requirement that all individuals who will go on to repay a loan have the same likelihood… | senses[8] |
| sense | error rate balance - Chouldechova's name for the combined outcome-conditioned fairness notions— “The second pair, equality of true positive rates and equality of false negative rates, is equivalent to D⊥A|Y = 1. This definition of fairness takes the perspective of those individuals with outcome defined by one. For example, this is equivalent to the requirement that all individuals who will go on to repay a loan have the same likelihood… | senses[9] |
| sense | equalized odds - Hardt and colleagues' name for the combined outcome-conditioned fairness notions— “The second pair, equality of true positive rates and equality of false negative rates, is equivalent to D⊥A|Y = 1. This definition of fairness takes the perspective of those individuals with outcome defined by one. For example, this is equivalent to the requirement that all individuals who will go on to repay a loan have the same likelihood… | senses[10] |
| sense | equality of negative predictive value - equality across groups in negative predictive value— “The other pair of definitions that appear on this margin are equality of positive predictive value and equality of false discovery rate. These are defined by the conditional independence relationship Y⊥A|D = 1. This definition of fairness is also sometimes called predictive parity (Chouldechova 2017), and it is assessed by an outcome test… | senses[11] |
| sense | equality of false omission rate - equality across groups in false omission rate— “The other pair of definitions that appear on this margin are equality of positive predictive value and equality of false discovery rate. These are defined by the conditional independence relationship Y⊥A|D = 1. This definition of fairness is also sometimes called predictive parity (Chouldechova 2017), and it is assessed by an outcome test… | senses[12] |
| sense | equality of positive predictive value - equality across groups in positive predictive value— “The other pair of definitions that appear on this margin are equality of positive predictive value and equality of false discovery rate. These are defined by the conditional independence relationship Y⊥A|D = 1. This definition of fairness is also sometimes called predictive parity (Chouldechova 2017), and it is assessed by an outcome test… | senses[13] |
| sense | equality of false discovery rate - equality across groups in false discovery rate— “The other pair of definitions that appear on this margin are equality of positive predictive value and equality of false discovery rate. These are defined by the conditional independence relationship Y⊥A|D = 1. This definition of fairness is also sometimes called predictive parity (Chouldechova 2017), and it is assessed by an outcome test… | senses[14] |
| sense | predictive parity - another name for equality of positive predictive value and false discovery rate— “The other pair of definitions that appear on this margin are equality of positive predictive value and equality of false discovery rate. These are defined by the conditional independence relationship Y⊥A|D = 1. This definition of fairness is also sometimes called predictive parity (Chouldechova 2017), and it is assessed by an outcome test… | senses[15] |
| sense | avoiding disparate mistreatment - Zafar and colleagues' collective name for the four predictive-measure pairs plus equal accuracy— “Zafar et al. (2017) call all four pairs of definitions (as well as equal accuracy) avoiding disparate mistreatment. Berk et al. (2017) also consider a definition based on several of the above elements of the confusion matrix. They define treatment equality in terms of the ratio of false negatives to false positives” | senses[16] |
| sense | treatment equality - a fairness definition based on the ratio of false negatives to false positives | senses[17] |
| sense | demographic parity - equal decision rates across groups irrespective of outcome Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[18] |
| sense | statistical parity - an alternative name for demographic parity Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[19] |
| sense | group fairness - an alternative name for demographic parity Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[20] |
| sense | Conditional demographic parity - parity within strata, expressed as conditional independence of decision and sensitive group Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[21] |
| sense | fairness through unawareness - conditional demographic parity when conditioning on insensitive variables Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[22] |
| sense | anticlassification - conditional demographic parity when conditioning on insensitive variables Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[23] |
| sense | treatment parity - conditional demographic parity when conditioning on insensitive variables Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[24] |
| sense | metric fairness - requiring closeness under a similarity metric to imply closeness in decisions Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[25] |
| sense | individual fairness - an alternative name for metric fairness Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[26] |
| sense | m-Lipschitz property - an alternative mathematical name for metric fairness Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[27] |
| sense | perfect metric fairness - Rothblum and Yona's name for metric fairness Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[28] |
| sense | individual counterfactual fairness - the strongest listed counterfactual criterion, requiring decision equality for every individual Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[29] |
| sense | conditional counterfactual fairness - the intermediate counterfactual criterion, requiring conditional mean-decision equality Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[30] |
| sense | counterfactual parity - the weakest listed counterfactual criterion, requiring population mean-decision equality Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | senses[31] |
| control | whether the incompatibility is an artefact of using algorithms at all — held Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. — rules out the algorithmic explanation - the trade-off 'would apply even if a human were the one making the prediction', so refusing automation does not avoid it Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[0] |
| control | under what conditions the two COMPAS criteria COULD both hold, worked out formally — held Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. — rules out any hope of reconciliation in practice - it requires equal recidivism rates across racial groups, or a group that never experiences an outcome, or perfect prediction Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[1] |
| control | whether formal fairness work is worth doing at all, against its own critics — partial Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. — rules out nothing on either side, and the paper says so: the criticisms are 'rightly pointed out', and formalism 'should not be dismissed as necessarily preserving the status quo'. Both are asserted; neither is tested Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[2] |
| weight | both positions in the COMPAS dispute, insofar as either claimed to be THE fairness answer - 'there can be no harmony among definitions in a world where inequality and imperfect prediction are the reality' — 0.9 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | claim_relations[0] |
| weight | the impossibility results of Kleinberg et al. and Chouldechova, restated here in one notation and set against a public case — 0.8 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | claim_relations[1] |
| weight | the psychometric literature on test fairness running back to the 1960s, named as the older form of the same question — 0.6 Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | claim_relations[2] |
| edge | equal false positive rates —distinct-from→ equal positive predictive values Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[0] |
| edge | quantitative notions of fairness —contrasted-with→ structural conditions Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[1] |
| audit | equality of true positive rates —renamed-across-fields→ equal opportunity — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but the span says this condition 'has been called equal opportunity' by one author; a name, and renamed-across-fields is loose. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[2] |
| note | equality of true positive rates —renamed-across-fields→ equal opportunity — The paper reports Hardt et al.'s unhedged name for this condition. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[2] |
| note | separation —renamed-across-fields→ error rate balance — The paper gives source-specific names for the same combined outcome-conditioned fairness notions. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[3] |
| note | separation —renamed-across-fields→ equalized odds — The paper gives source-specific names for the same combined outcome-conditioned fairness notions. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[4] |
| audit | Conditional demographic parity —same-as→ separation — Audit 2026-09-28: DOUBTFUL - read against its span and kept, but conditional demographic parity equals separation only when Data = Y; same-as is conditional. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[5] |
| holds when | Conditional demographic parity —same-as→ separation — Data = Y Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[5] |
| note | Conditional demographic parity —same-as→ separation — The equivalence is asserted without epistemic hedge but explicitly scoped to conditioning on outcome Y. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[5] |
| audit | Conditional demographic parity —same-as→ fairness through unawareness — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but endpoint by anaphora: 'it' = conditional demographic parity (previous sentence); the equivalence holds only when Data = X. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[6] |
| holds when | Conditional demographic parity —same-as→ fairness through unawareness — Data = X (the insensitive variables) Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[6] |
| note | Conditional demographic parity —same-as→ fairness through unawareness — The equivalence is asserted without epistemic hedge under the paper's explicit Data = X condition. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[6] |
| audit | Conditional demographic parity —same-as→ anticlassification — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but endpoint by anaphora: 'it' = conditional demographic parity (previous sentence); the equivalence holds only when Data = X. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[7] |
| holds when | Conditional demographic parity —same-as→ anticlassification — Data = X (the insensitive variables) Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[7] |
| note | Conditional demographic parity —same-as→ anticlassification — The equivalence is asserted without epistemic hedge under the paper's explicit Data = X condition. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[7] |
| audit | Conditional demographic parity —same-as→ treatment parity — Audit 2026-09-28: DOUBTFUL - read against its full span and kept, but endpoint by anaphora: 'it' = conditional demographic parity (previous sentence); the equivalence holds only when Data = X. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[8] |
| holds when | Conditional demographic parity —same-as→ treatment parity — Data = X (the insensitive variables) Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[8] |
| note | Conditional demographic parity —same-as→ treatment parity — The equivalence is asserted without epistemic hedge under the paper's explicit Data = X condition. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[8] |
| note | metric fairness —same-as→ individual fairness — The paper unhedgedly says metric fairness is also known as individual fairness. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[9] |
| note | metric fairness —same-as→ m-Lipschitz property — The paper unhedgedly gives the m-Lipschitz property as another name for metric fairness. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[10] |
| note | metric fairness —same-as→ perfect metric fairness — The paper unhedgedly gives perfect metric fairness as another name for metric fairness. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concept_edges[11] |
| outcome | control — held Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[0] |
| outcome | control — held Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[1] |
| outcome | control — partial Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | controls[2] |
| no type for this | Calibration within groups → sufficiency — The paper says calibration within groups implies sufficiency, and describes a converse after applying a function to a sufficient score. Logical implication between criteria is not one of the eight edge types. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | unhoused_relations[0] |
| no type for this | metric fairness → unawareness — When the metric uses only insensitive variables, the paper says metric fairness implies unawareness. Logical implication is not one of the eight edge types. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | unhoused_relations[1] |
| no type for this | individual counterfactual fairness → conditional counterfactual fairness — The paper orders individual counterfactual fairness above conditional counterfactual fairness in strength. A stronger-than relation is not one of the eight edge types. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | unhoused_relations[2] |
| no type for this | conditional counterfactual fairness → counterfactual parity — The paper orders conditional counterfactual fairness above counterfactual parity in strength. A stronger-than relation is not one of the eight edge types. Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | unhoused_relations[3] |
| limitation | the paper declines the job most readers want done: `The goal of this article is not to advance axiomatic definitions of fairness but to summarize the definitions and results ... formalized to date`. It catalogues a disagreement rather than settling it, and says so in the introduction Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[0] |
| limitation | and it names the gap its own subject cannot close: `gaps between mathematically convenient formalism and the larger social goals that many of these concepts were introduced to address`. THAT IS THE OPERATIONALISATION GAP, NAMED AS SUCH, by a paper whose subject is operationalisations Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[1] |
| limitation | the object being predicted is already an abstraction of the thing anyone cares about: policy goals are `sometimes clumsily, abstracted and formulated to fit into a prediction task`, and the choices that do it sit outside the model Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[2] |
| limitation | and the decision itself is flattened before any of this begins: a judge's options are many, `in this literature, this decision is often reduced to A BINARY DECISION - RELEASE OR DETAIN`. Every fairness metric in the catalogue is defined over that binary Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[3] |
| limitation | the authors record the standing criticism of their own field without dismissing it: quantitative notions of fairness `can funnel our thinking into narrow silos where we aim to make adjustments to a decision-making process, rather than to address the structural conditions that sustain inequality in society` Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | limitations[4] |
finding · quantity · variable · definition · key-term — grounded in the full text| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[0] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[1] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[2] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[3] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[4] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[5] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[6] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[7] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[8] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[9] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[10] |
| finding | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.findings[11] |
| definition | bias Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.definitions[0] |
| concept | algorithmic fairness — not defined in this paper's own words; the field name, given to work whose questions psychometrics has asked since the 1960s Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[0] |
| concept | equal false positive rates — not defined in this paper's own words; ProPublica's fairness criterion - among those who did not get rearrested, equal misclassification by race Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[1] |
| concept | equal positive predictive values — not defined in this paper's own words; Northpointe's criterion - among those called higher risk, an equal proportion rearrested, regardless of race Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[2] |
| concept | calibration within groups — not defined in this paper's own words; the third criterion COMPAS satisfies, and one of the three in Kleinberg's impossibility result Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[3] |
| concept | sufficiency — not defined in this paper's own words; the condition that implies equality of positive predictive values Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[4] |
| concept | separation — not defined in this paper's own words; the condition that implies equality of false positive rates Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[5] |
| concept | bias — not defined in this paper's own words; declared OVERLOADED - in popular media it has come to mean that model performance, however defined, unjustifiably differs along social axes Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[6] |
| concept | quantitative notions of fairness — not defined in this paper's own words; the formal definitions collectively, and the thing criticised for funnelling attention into adjustments rather than structures Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[7] |
| concept | structural conditions — not defined in this paper's own words; what the formal definitions are accused of directing attention away from Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[8] |
| concept | equality of true positive rates — not defined in this paper's own words; equality across groups in the true-positive rate Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[9] |
| concept | equality of false negative rates — not defined in this paper's own words; equality across groups in the false-negative rate Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[10] |
| concept | equal opportunity — not defined in this paper's own words; the paper's name for equality of true-positive and false-negative rates considered alone Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[11] |
| concept | error rate balance — not defined in this paper's own words; Chouldechova's name for the combined outcome-conditioned fairness notions Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[12] |
| concept | equalized odds — not defined in this paper's own words; Hardt and colleagues' name for the combined outcome-conditioned fairness notions Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[13] |
| concept | equality of negative predictive value — not defined in this paper's own words; equality across groups in negative predictive value Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[14] |
| concept | equality of false omission rate — not defined in this paper's own words; equality across groups in false omission rate Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[15] |
| concept | equality of positive predictive value — not defined in this paper's own words; equality across groups in positive predictive value Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[16] |
| concept | equality of false discovery rate — not defined in this paper's own words; equality across groups in false discovery rate Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[17] |
| concept | predictive parity — not defined in this paper's own words; another name for equality of positive predictive value and false discovery rate Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[18] |
| concept | avoiding disparate mistreatment — not defined in this paper's own words; Zafar and colleagues' collective name for the four predictive-measure pairs plus equal accuracy Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[19] |
| concept | treatment equality — not defined in this paper's own words; a fairness definition based on the ratio of false negatives to false positives Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[20] |
| concept | demographic parity — not defined in this paper's own words; equal decision rates across groups irrespective of outcome Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[21] |
| concept | statistical parity — not defined in this paper's own words; an alternative name for demographic parity Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[22] |
| concept | group fairness — not defined in this paper's own words; an alternative name for demographic parity Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[23] |
| concept | Conditional demographic parity — not defined in this paper's own words; parity within strata, expressed as conditional independence of decision and sensitive group Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[24] |
| concept | fairness through unawareness — not defined in this paper's own words; conditional demographic parity when conditioning on insensitive variables Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[25] |
| concept | anticlassification — not defined in this paper's own words; conditional demographic parity when conditioning on insensitive variables Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[26] |
| concept | treatment parity — not defined in this paper's own words; conditional demographic parity when conditioning on insensitive variables Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[27] |
| concept | metric fairness — not defined in this paper's own words; requiring closeness under a similarity metric to imply closeness in decisions Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[28] |
| concept | individual fairness — not defined in this paper's own words; an alternative name for metric fairness Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[29] |
| concept | m-Lipschitz property — not defined in this paper's own words; an alternative mathematical name for metric fairness Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[30] |
| concept | perfect metric fairness — not defined in this paper's own words; Rothblum and Yona's name for metric fairness Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[31] |
| concept | individual counterfactual fairness — not defined in this paper's own words; the strongest listed counterfactual criterion, requiring decision equality for every individual Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[32] |
| concept | conditional counterfactual fairness — not defined in this paper's own words; the intermediate counterfactual criterion, requiring conditional mean-decision equality Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[33] |
| concept | counterfactual parity — not defined in this paper's own words; the weakest listed counterfactual criterion, requiring population mean-decision equality Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | concepts[34] |
| false positive rates | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.statistics[0] |
| positive predictive values | Quotation withheld — this paper is not ours to reproduce, and the page has already quoted as much of it as the rights layer allows. The tag, its field and its confidence stand; the sentence is at the publisher. | content_tags.statistics[1] |
When we have changed our mind. Some entries in the table begin with a word in capitals and the reason for it. EVIDENCE CORRECTED — the finding still stands, but we had quoted the wrong sentence from the paper, and the quotation has been replaced with the right one. RETRACTED — the entry itself was wrong; it stays on the page, marked, with the reason, rather than being deleted. WITHDRAWN AND REPLACED — the entry was wrong, and a corrected version of it appears in the entries below. Each mark carries its own reason, in brackets, because the reasons differ from one entry to the next. An entry we have WITHDRAWN is removed from this page and kept in the record, with its reason. A line reading Audit and a date is a later re-reading of the entry against its whole sentence: DOUBTFUL means it was kept, and the reason says what is loose about it. Quotation marks inside the tagger's sentence. Where a tag is the tagger's own sentence, a phrase in quotation marks inside it is borrowed from the paper — the tagger's wording around it is not. The paper's own sentence is the evidence quote for that row; on this page it is withheld with the article, and the article is at its publisher. Nothing inside quotation marks anywhere on this page is ours: if it is quoted, it is the paper's. Whose words a concept carries. A concept entry records a term the paper itself uses, so the term, the sense given for it, and its evidence are the paper's own words, quoted where the paper fixes them. This is the opposite of a sense entry in the tag table, which is the tagger's sentence about how the paper uses a term. Where a concept entry could not be grounded in the paper's words, the entry says so on its own row.
Tag layer © the Meta-Tagging Project, CC BY 4.0. The paper text is © its authors/publisher
and is linked, not reproduced. Excerpts shown are short quotations for identification and verification.
· back to index
· concepts · authors
· עברית
© 2026 Shir Sivroni · ORCID 0009-0000-2597-2824