Federal Trade Commission
26
On the other hand, several stakeholders argue that these concerns are overstated.
141
Some emphasize that,
to the extent the various steps in data mining lead to disparate impact, these issues are not new—they are
inherent in any statistical analysis.
142
Other writers note that, rather than disadvantaging minorities in the
hiring process, big data can help to create “a labor market that’s fairer to people at every stage of their careers.”
143
For example, companies can use big data algorithms to nd employees from within underrepresented segments
of the population.
144
ey can also use big data to identify biases so that they can choose candidates based on
merit rather than using mechanisms that depend on the reviewers’ biases.
145
Furthermore, as other stakeholders
have noted, big data can help “reduce the rate of ‘false positive’ cases that potentially make disparate treatment a
problem”
146
and can help identify whether correlations exist between prices and variables such as race, gender or
ethnicity.
147
ese stakeholders do not argue that we should ignore discrimination where it occurs; rather, they
argue that we should recognize the potential benets of big data to reduce discriminatory harm.
Common Sense Media Comment #00016, supra note 8; N.Y.U. Info. L. Inst. Comment #00015, supra note 8; World Privacy
Forum Comment #00014, supra note 19; Ctr. for Dig. Democracy & U.S. PIRG Educ. Fund Comment #00003, supra note
8. See also Barocas & Selbst, supra note 137; Crawford, supra note 39.
141 See, e.g., Big Data Tr. 75 (Gene Gsell). See generally Comment #00081 from Berin Szoka & Tom Struble, TechFreedom,
& Georey Manne & Ben Sperry, Int’l Ctr. for L. & Econ., to Fed. Trade Comm’n (Nov. 3, 2014), https://www.ftc.
gov/system/les/documents/public_comments/2014/11/00081-92956.pdf; Comment #00074 from Howard Fienberg,
Mktg. Research Assoc., to Fed. Trade Comm’n (Oct. 31, 2014), https://www.ftc.gov/system/les/documents/public_
comments/2014/10/00074-92927.pdf; Comment #00070 from Bijan Madhani, Computer & Commc’ns Indus. Assoc., to
Fed. Trade Comm’n (Oct. 31, 2014), https://www.ftc.gov/system/les/documents/public_comments/2014/10/00070-92912.
pdf; NetChoice Comment #00066, supra note 23; Ctr. for Data Innovation Comment #00055, supra note 8; Ctr. for Data
Innovation Comment #00026, supra note 8; Tech. Pol’y Inst. Comment #00010, supra note 8; V M-S
K C, B D: A R T W T H W L, W, A T (2013).
142 See, e.g., Dan Gray, Ethics, Privacy and Discrimination in the Age of Big Data, D (Dec. 3, 2014), http://
dataconomy.com/ethics-privacy-and-discrimination-in-the-age-of-big-data/. But see Je Leek, Why Big Data Is in Trouble:
ey Forgot About Applied Statistics, SS (May 7, 2014), http://simplystatistics.org/2014/05/07/why-big-data-is-
in-trouble-they-forgot-about-applied-statistics/ (noting that big data users have not given sucient attention to issues that
statisticians have been thinking about for a long time: sampling populations, multiple testing, bias, and overtting).
143 See, e.g., Don Peck, ey’re Watching You at Work, A (Dec. 2013), http://www.theatlantic.com/magazine/
archive/2013/12/theyre-watching-you-at-work/354681/.
144 See, e.g., Big Data Tr. 126 (Mark MacCarthy), 251 (Christopher Wolf). See also Software & Info. Indus. Assoc. Comment
#00067, supra note 2, at 7; Future of Privacy Forum Comment #00027, supra note 23, attached report entitled, B D: A
T F D E G, at 1–2.
145 See, e.g., Anne Loehr, Big Data for HR: Can Predictive Analytics Help Decrease Discrimination in the Workplace?, H
P (Mar. 23, 2015), http://www.hungtonpost.com/anne-loehr/big-data-for-hr-can-predi_b_6905754.html.
146 W H F. R, supra note 56, at 16.
147 Id. at 17. Economists have documented ways that data can help identify discrimination against protected groups in a wide
variety of settings. For example, a randomized experiment changed the names on resumes sent to employers from white-
sounding names to African-American sounding names; resumes with white-sounding names were 50 percent more likely to be
called back for an interview. Marianne Bertrand & Sendhil Mullainathan, Are Emily and Greg More Employable an Lakisha
and Jamal? A Field Experiment on Labor Market Discrimination, 94 A. E. R. 991, 991–1013 (2004). Research from
the early days of the Internet found that African-Americans and Latinos paid about 2 percent more for used cars purchased
oine, but paid similar prices for those purchased online; the proered reason was that individuals were anonymous online.
Fiona Scott Morton et al., Consumer Information and Discrimination: Does the Internet Aect the Pricing of New Cars to Women
and Minorities?, 1 Q M. E. 65, 65–92 (2003). See also Devin Pope & Justin Sydnor, Implementing
Anti-Discrimination Policies in Statistical Proling Models, 3 A. E. J.: E. P’ 206, 206–231 (2011), http://faculty.
chicagobooth.edu/devin.pope/research/pdf/Website_Antidiscrimination%20Models.pdf.