Introduction
Methods
No. | Matches search terms | PsychInfo | PhilPapers | SocIndex | CINAHL | PubMed | Web of science |
---|---|---|---|---|---|---|---|
1 | “Big data” OR “digital data” OR “data mining” OR “data linkage” | 2385 | 179 | 507 | 944 | 13214 | 23740 |
2 | Discriminat* OR *equality OR vulnerab* OR *justice OR ethic* OR exclusion | 69,435 | 46,349 | 46,624 | 38,096 | 245,604 | 414,661 |
3 | 1 AND 2 | 156 | 67 | 88 | 55 | 769 | 1177 |
Results
Author, Year, Country | Design | Participants | Discipline | Field of application | Definition of discrimination | Reference to legislation/regulatory text |
---|---|---|---|---|---|---|
Ajana (2015) [1], UK | Theoretical | Social Sciences | Migration | Unequal treatment | ||
Ajunwa et al. (2016) [2], USA | Theoretical | Bioethics | Employment | Not given—self explanatory | ||
Bakken and Reame (2016) [6], USA | Theoretical | Bioethics | Healthcare research | Not applicable—digital divide | ||
Barocas and Selbst (2016) [8], USA | Theoretical | Law | Employment | Disparate treatment/disparate impact | ||
Berendt and Preibusch (2014) [10], Belgium-UK | Other | Computer Science | Various | Juridical—legally protected classes | ||
Berendt and Preibusch (2017) [11], Belgium-UK | Other | Computer Science | Various | Illegitimate discrimination on grounds of four protected attributes | ||
Boyd and Crawford (2012) [12], Australia-USA | Theoretical | Social Sciences | Digital divide in research | Not applicable—digital divide | ||
Brannon (2017) [13], USA | Theoretical | Social Sciences | Social disparity | Not given—inequality | ||
Brayne (2017) [14], USA | Qualitative | A sample of Employees of LAPD (Officers and Civilians) | Social Sciences | Policing/criminology | Not given—inequality | |
Calders and Verwer (2010) [17], Netherlands | Other | Computer Science | Various | Not given—self explanatory | ||
Casanas i Comabella and Wanat (2015) [18], UK | Theoretical | Bioethics | Digital divide in research | Not applicable—digital divide | ||
Cato et al. [19], USA | Theoretical | Bioethics | Healthcare | Not given—injustice | Belmont Report; 1976 | |
Chouldechova (2017) [20], USA | Other | A sample of Caucasian/African American US Defendants | Computer Science | US criminal justice system | Disparate impact | |
Citron and Pasquale (2014) [21], USA | Theoretical | Law | Credit scoring | Not given—reference to protected classes | ||
Cohen et al. (2017) [22], USA | Theoretical | Bioethics | Healthcare | Not given—inequality | ||
d’Alessandro et al. (2017) [25], USA | Theoretical | Computer Science | Various | Disparate treatment/disparate impact | ||
de Vries (2010) [27], Belgium | Theoretical | Philosophy | Various | Unwarranted discrimination | ||
Francis and Francis (2017) [30], USA | Theoretical | Law | Healthcare and healthcare research | Not given—stigmatization and harm | ||
Hajian and Domingo-Ferrer (2013) [32], Spain | Other | Computer Science | Various | Not given—self explanatory | ||
Hajian et al. (2014) [33], Spain | Other | Computer Science | Various | Unfair or unequal treatment | Australian Legislation 2008; European Union Legislation 2009 | |
Hajian et al. (2015) [34], Italy-Spain | Other | Computer Science | Various | Unfair or unequal treatment | Australian Legislation 2014; European Union Legislation 2014 | |
Hildebrandt and Koops (2010) [35], USA | Theoretical | Law | Ambient intelligence | Unlawful/unfair discrimination | ||
Hirsch (2015) [36], USA | Theoretical | Law | Various | Not given—elusive concept | ||
Hoffman (2010) [37], USA | Theoretical | Social Sciences | Employment | Unlawful discrimination on basis of disability | Americans with Disabilities Act (ADA), 1990; Genetic Information Nondiscrimination Act (GINA), 2003; Health Insurance Portability and Accountability Act (HIPAA), 1996 | |
Hoffman (2017) [38], USA | Theoretical | Social Sciences | Employment | Unlawful discrimination on basis of disability | Americans with Disabilities Act (ADA), 1990; Genetic Information Nondiscrimination Act (GINA), 2003; Health Insurance Portability and Accountability Act (HIPAA), 1996 | |
Holtzhausen (2016) [39], USA | Theoretical | Social Sciences | Various | Not given—self explanatory | ||
Kamiran and Calders (2012) [42], Netherlands-UK | Other | Computer Science | Various | Unfair and unequal treatment | Australian Sex Discrimination Act, 1984; US Equal Pay Act, 1963; US Equal Credit Opportunity Act, 1974; European Council Directive, 2004 | |
Kamiran et al. (2013) [43], Netherlands-Saudi Arabia-UK | Other | Computer Science | Various | Unfair and unequal treatment | Australian Sex Discrimination Act, 1984; US Equal Pay Act, 1963 | |
Kennedy and Moss (2015) [44], UK | Theoretical | Social Sciences | Society and culture | Not given—self explanatory | ||
Kroll et al. (2017) [45], USA | Theoretical | Law | Various | Not given—opposite of fair treatment | ||
Kuempel (2016) [46], USA | Theoretical | Law | Various | Not given—self explanatory | ||
Le Meur et al. (2015) [47], France | Quantitative | A sample of pregnant women | Bioethics | Healthcare | Not given | |
Leese (2014) [48], Germany | Theoretical | Ethics | Aviation/migration | Principle of equality and non discrimination | [60]; European Convention on Human Rights, 1953; Treaty on the Functioning of the European Union, 1958 | |
Lerman (2013) [49], USA | Theoretical | Law | Digital divide in social participation | Social marginalization/exclusion | ||
Lupton (2015) [51], Australia | Theoretical | Social Sciences | Society | Not given—stigmatization | ||
MacDonnell (2015) [53], Ireland | Theoretical | Social Sciences | Insurance | Not given | ||
Mantelero (2016) [54], China-Italy | Theoretical | Social Sciences | Various | Unjust or prejudicial treatment | ||
Mao et al. (2015) [55], USA | Quantitative | A sample of citizens from Cote D’Ivoire | Social Sciences | Economic development | Not given—related to social and economic disparity | |
Newell and Marabelli (2015) [58], UK-USA | Theoretical | Social Sciences | Various | Not given—Harm towards vulnerable individuals | ||
Nielsen et al. (2017) [58], Brasil-USA | Quantitative | A sample of Twitter users in Brazil | Social sciences | Public health | Not given—self explanatory | |
Pak et al. (2017) [60], Belgium | Quantitative | Citizens of Brussels using “Fix My Street” App | Social Science | Urban and social involvement | Not given—social exclusion/disparity | |
Peppet (2014) [62], USA | Theoretical | Law | Various | Illegal or unwanted discrimination | ||
Ploug and Holm (2017) [64], Denmark | Theoretical | Bioethics | Society | Differential treatment and stigmatization | ||
Pope and Sydnor (2011) [66], USA | Other | Full sample of UI claimants from the State of New Jersey between 1995 and 1997 | Computer Science | Employment | Not given—self explanatory | |
Romei et al. (2013) [70], Italy | Quantitative | Italian female researchers | Computer Science | Academia | Unjustified distinction of individuals based on their membership | European Union Legislation, 2010 |
Ruggieri et al. (2010) [71], Italy | Other | Computer Science | Various | Juridical | Australian Legislation, 2010; European Union Legislation, 2010; United Nations Legislation, 2010; U.K. Legislation, 2010; U.S. Federal Legislation, 2010 | |
Sharon (2016) [74], Netherlands | Theoretical | Bioethics | Healthcare and Healthcare Research | Not given—self explanatory | ||
Schermer (2011) [73], Netherlands | Theoretical | Social Sciences | Not Defined | Not given—self explanatory/Stigmatization | ||
Susewind [76], Germany | Quantitative | Selected Asian countries | Social Sciences | Various | Not given—self explanatory | |
Taylor (2016) [78], Netherlands | Qualitative | West Africa Population (Cote d’Azur) | Social Sciences | Surveillance | Not given—self explanatory | |
Taylor (2017) [79], Netherlands | Theoretical | Social Sciences | Various | Disparity/inequality/exclusion | ||
Timmis et al. (2016) [80], UK | Theoretical | Social Sciences | Education | Not given—social exclusion/disparity | ||
Turow et al. (2015) [81], USA | Theoretical | Social Sciences | Marketing | Social discrimination | ||
Vaz et al. (2017) [83], Canada | Quantitative | Social Sciences | Urban development | Social inequalities | ||
Veale (2017) [84], UK | Theoretical | Social Sciences | Various | Not given—opposite of fairness and equality | ||
Voigt (2017) [85], Canada | Theoretical | Social Sciences | Healthcare | Inequality | ||
Zarate et al. (2016) [91], USA | Qualitative | Participants of the PGP (Personal Genome Project) | Bioethics | Various | Not given—self explanatory | |
Zarsky (2014) [93], Israel | Theoretical | Law | Various | Illusive concept—unfair or Unequal Treatment of the individual | ||
Zarsky (2016) [92], Israel | Theoretical | Law | Credit scoring | Unfairness and inequality | ||
Zliobaite and Custers (2016) [95], Finland-Netherlands | Other | Computer Science | Various | Juridical | Race Equality Directive (2000/43/EC), Employment Equality Directive (2007/78/EC), Gender Recast Directive (2006/54/EC), Gender Goods and Services Directive (2006/113/EC) | |
Zliobaite (2017) [94], Finland-Netherlands | Other | Computer Science | Various | Adversary treatment of people based on belonging to some group | Race Equality Directive (2000/43/EC), Employment Equality Directive (2007/78/EC), Gender Recast Directive (2006/54/EC), Gender Goods and Services Directive (2006/113/EC) |
Discrimination and data mining
Forms, targets and consequences of discrimination
Discriminatory outcomes | Paper references |
---|---|
1. Forms of discrimination | |
1.1. Accidental/involuntary discrimination | |
1.2. Direct voluntary discrimination | |
2. Victims/targets of discrimination | |
2.1. Vulnerable groups/populations | |
2.2. Larger groups | |
3. Discriminatory consequences | |
3.1. Social marginalization and stigma | |
3.2. Exacerbation of existing inequalities | |
3.3. New forms of discrimination | |
3.3.1. Economic discrimination | |
3.3.2. Health prediction discrimination |
Causes of discrimination
Causes of discrimination | Related articles |
---|---|
1. Algorithmic causes | |
1.1. Definition of the target variable | |
1.2. Data issues Training data (Historically biased data sets) | |
1.3. Data issues Training data (manual assignment of class labels) | |
1.4. Data issues Data collection (Overrepresentation and underrepresentation) | |
1.5. Proxies | |
1.6. Feedback loop | |
1.7. Overfitting | |
1.8. Feature selection | Barocas and Selbst 2016 [8] |
1.9. Cost function Error by omission | d'Alessandro et al. 2017 [25] |
1.10 Masking Proxies | |
2. Digital divide | |
2.1. Skills | |
2.2. Resources | |
2.3. Geographical location | |
2.4. Age | Casanas i Comabella and Wanat 2015 [18] |
2.5. Income | |
2.6 Gender | Boyd and Crawford 2012 [12] |
2.7. Education | Boyd and Crawford 2012 [12] |
2.8 Race | |
3. Data linkage |
Algorithmic causes of discrimination
Digital divide
Data linkage and aggregation
Suggested solutions
Suggested solutions | Paper references |
---|---|
1. Computer science and technical solutions | |
1.1. Pre-processing | |
1.2. In-processing | |
1.3. Post-processing | Hajian et al. 2015 [34] |
1.4.Mixed methods | d'Alessandro et al. 2017 [25] |
1.5. Implementation of transparency | |
1.6. Privacy preserving strategies | |
1.7. Exploratory fairness analysis | Veale and Binns 2017 [84] |
2. Legal solutions | |
3. Human based solutions | |
3.1. Human in the loop | |
3.2. Third parties | |
3.3. Multidisciplinary involvement | |
3.4. Education | |
3.5. Implementing EHR flexibility | Hoffman 2010 [37] |
Practical computer science and technological solutions
Legal solutions
Human-centered solutions
Obstacles to fair data mining
Obstacles to fair data analytics | Paper references |
---|---|
1. Black box | Hildebrandt and Koops 2010 [35], Ruggieri et al. 2010 [71], Schermer 2011 [73], Berendt and Preibusch 2014 [10], Citron and Pasquale 2014 [21], Cohen et al. 2014 [22], Leese 2014 [48], Zarsky 2014 [93], Kennedy and Moss 2015 [44], Newell and Marabelli 2015 [58], Turow, McGuigan et al. 2015 [81], Mantelero 2016 [54], Zarsky 2016 [92], Brannon 2017 [13], Brayne 2017 [14], d'Alessandro et al. 2017 [25], Kroll et al. 2017 [45], Taylor 2017 [79] |
2. Human bias | Boyd and Crawford 2012 [12], Kamiran and Calders 2012 [42], Citron and Pasquale 2014 [21], Zarsky 2014 [93], Ajana 2015 [1], Ajunwa et al. 2016 [2], Barocas and Selbst 2016 [8], Berendt and Preibusch 2017 [11], Brayne 2017 [14], d'Alessandro et al. 2017 [25], Veale and Binns 2017 [84], Voigt 2017 [85] |
3. Conceptual challenges | de Vries 2010 [27], Hoffman 2010 [37], Lerman 2013 [49], Leese 2014 [48], Zarsky 2014 [93], Ajana 2015 [1], Hirsch 2015 [36], MacDonnell 2015 [53], Barocas and Selbst 2016 [8], Kuempel 2016 [46], Mantelero 2016 [54], Francis and Francis 2017 [30], Hoffman 2017 [38], Kroll et al. 2017 [45], Taylor 2017 [79] |
4. Inadequate legislation |
Beneficial adoption of Big Data technologies
Beneficial adoption of Big Data | Paper references |
---|---|
1. Promotion of objectivity in classification | |
2. Uncover and assess discriminatory practices | |
3. Integration of data for promotion of equality and social integration | |
3.1. Healthcare | |
3.2. Economic growth and urban development | |
3.3. Migration | |
4. Beneficial use of social media |