Statistical Discrimination Research Articles

To Segregate, or to Discriminate – That Is the Question: Experiment on Identity and Social Preferences

How do various sources of social identity affect segregation and discrimination decisions? In our laboratory experiment, social identity originates either from similar preferences, income, ability, randomly or from shared socioeconomic status. For the latter, we exploit Colombia's unique (public information) stratification system which assigns households to socioeconomic strata based on its residential block amenities. Subjects decide with whom to interact in a Dictator and Trust Game. We find high socioeconomic status senders segregate against out-group receivers in the Dictator Game, while low socioeconomic ones do so in the Trust Game. This segregation pattern is partly explained by payoff-maximizing behavior. In the Trust Game, we gather evidence for statistical discrimination. In the Dictator Game, evidence points to a taste for redistribution when identity originates from socioeconomic status or income level. No matter the source of identity, our subjects expect being segregated but not discriminated against.

Age Discrimination in Hiring Decisions: A Factorial Survey among Managers in Nine European Countries

AbstractThis article analyses old-age discrimination in managers’ hypothetical hiring decisions. We expect that older job candidates are less likely to be hired than equally qualified younger candidates. Statistical discrimination theory argues that when recruiters have more information about the candidate’s skills, age is less important for hiring decisions. Given inconclusive results of previous studies, we elaborate on the theory by focusing on the content rather than the amount of information. We argue that information is primarily influential if it debunks, rather than confirms, ageist stereotypes. To test this argument, a factorial survey was conducted among 482 managers in nine European countries. The findings show that older candidates indeed receive lower hireability scores, and this finding is robust across countries and sectors. However, we do not find that stereotype-rejecting information moderates age discrimination: it does not matter whether recruiters have information that debunks or confirms ageist stereotypes; age is equally important in both situations. Our findings suggest that for hiring decisions, the valuation of applicants’ skills and their age are largely independent.

A Characterisation of ‘Phelpsian’ Statistical Discrimination

Abstract We establish that a type of statistical discrimination—that based on informativeness of signals about workers’ skills and the ability appropriately to match workers to tasks—is possible if and only if it is impossible uniquely to identify the signal structure observed by an employer from a realised empirical distribution of skills. The impossibility of statistical discrimination is shown to be equivalent to the existence of a fair, skill-dependent, remuneration for workers. Finally, we connect the statistical discrimination literature to Bayesian persuasion, establishing that if discrimination is absent, then the optimal signalling problem results in a linear pay-off function (as well as a kind of converse).

Do Women Receive Worse Financial Advice?

Trained undercover men and women posed as potential clients and visited all 65 local financial advisory firms in Hong Kong. We find that women are more likely than men to receive advice to buy only individual or only local securities. This effect is significant for financial planners, but not for securities firms. Women who signal that they are highly confident, highly risk tolerant, or have a domestic outlook, are especially likely to receive suboptimal advice. Our theoretical model interprets these patterns as an interaction between statistical discrimination and advisors’ incentives. Taste-based discrimination is unlikely to explain the results.

Proof-of-Concept and Feasibility Study to Evaluate the Effect of β-Glucan on Protective Qi Deficiency in Adults.

To preliminarily explore the potential effect of β-glucan on Chinese medicine (CM) concept protective qi deficiency (PQD), and the methodology for future definitive studies. To have a standardized assessment of PQD, a list of 13 potentially PQD-relevant parameters were firstly created, each with defined quantitative or categorial scales. Using the data from 37 participants with (21 cases) or without (16 cases) PQD, multivariate logistic modeling was conducted to create a preliminary diagnostic PQD risk score. Subsequently, 21 participants diagnosed with PQD were treated with β-glucan in a dose of 200 mg/day for 8 weeks. Data were collected for trial acceptability measures (rate of recruitment, withdrawal, and compliance), and the participants were assessed for PQD status at baseline and every 2 weeks thereafter. The preliminary logistic model consisted of 3 parameters (low voice and apathy, aversion to wind and cold, and Cun pulse). The resulting risk score demonstrated a degree of PQD-predicting accuracy that, as evaluated by statistical (discrimination and classification) methods, was higher than those obtained from any of the individual candidate parameters. The 21 PQD participants treated with β-glucan demonstrated good receptibility and a time-dependent improvement in PQD status as evidenced by the decrease of PQD participant to 9.5% at the end of study. This study demonstrated the effect of proof-of-concept of β-glucan on improving PQD and the proof-of-concept of a multivariate-model-derived diagnostic PQD risk score. It also indicated feasibility for future definitive studies. Studies like this embody an innovative approach that uses therapies derived from the mainstream biomedicine to enrich therapeutics guided by CM principle. (Trial registration No. NCT03829228).

Algorithmic Discrimination and Input Accountability under the Civil Rights Acts

The disproportionate burden of COVID-19 among communities of color, together with a necessary renewed attention to racial inequalities, have lent new urgency to concerns that algorithmic decision-making can lead to unintentional discrimination against members of historically marginalized groups. These concerns are being expressed through Congressional subpoenas, regulatory investigations, and an increasing number of algorithmic accountability bills pending in both state legislatures and Congress. To date, however, prominent efforts to define algorithmic accountability have tended to focus on output-oriented policies that may facilitate illegitimate discrimination or involve fairness corrections unlikely to be legally valid. Worse still, other approaches focus merely on a model’s predictive accuracy—an approach at odds with long-standing U.S. antidiscrimination law. We provide a workable definition of algorithmic accountability that is rooted in the caselaw addressing statistical discrimination in the context of Title VII of the Civil Rights Act of 1964. Using instruction from the burden-shifting framework, codified to implement Title VII, we formulate a simple statistical test to apply to the design and review of the inputs used in any algorithmic decision-making processes. Application of the test, which we label the input accountability test, constitutes a legally viable, deployable tool that can prevent an algorithmic model from systematically penalizing members of protected groups who are otherwise qualified in a legitimate target characteristic of interest.

Comprehensive Chemical Characterization of the Pistacia vera Fruits through Original NMR Quantification Methods

(1) Background: Pistacia vera is a dry fruit spread all over the world, commonly known as pistachio, and is very appreciated as a snack, as a precious ingredient for confectionery industries and also as a blessed touch for gourmet dishes; (2) Methods: The simultaneous NMR profiles concerning the hydrophilic and hydrophobic phases, enabled by a specific throughput over grinded pistachio seeds, led to the comprehensive chemical characterization of these nuts coming from different territories; (3) Results: The customized extraction method is described in detail as well as the employed nuclear magnetic resonance (NMR) experiments. The chemical quantification takes advantage from the innovative multi-assignment recovered analysis (MARA-NMR) technique able to provide experimentally robust panels for both hydrophobic and hydrophilic chemicals. Beyond the assessed repeatability and the interesting statistical discrimination, this paper adds information about the quantitative and qualitative composition of pistachio as inferred by NMR data. This updated background can pave the way toward the development of wider and wider chemical panels improving the knowledge about the pistachio’s composition. This approach can be easily extended to other matrices.

Statistical Discrimination, Employer Learning, and Employment Gap by Race and Education

Tests of statistical discrimination require evaluation records provided by employers or variables that employers do not observe directly but are observed by researchers. As such variables are difficult to obtain, this paper develops a strategy that uses variables available in usual data sets. This paper derives testable implications for statistical discrimination by exploiting the heterogeneity in employer learning processes. Evidence from analysis using the March Current Population Survey for 1971-2016 is consistent with the theoretical predictions. The empirical findings are not explained by alternative hypotheses, such as human capital theory, taste-based discrimination, or search and matching models.

Opto-Electronic Nose Coupled to a Silicon Micro Pre-Concentrator Device for Selective Sensing of Flavored Waters

Headspace analysis of highly humid samples remains a challenge for artificial olfaction. Based on surface plasmon resonance imaging and bio-based sensors, the NeOse Pro olfactive analyzer yields multivariate data and enhances the statistical discrimination capacity of odor patterns. However, the presence of a high background signal, such as water vapor from aqueous samples, may deteriorate its discriminant ability. Recently, miniaturized pre-concentrators packed with hydrophobic adsorbent have been developed to improve the detection limit of gas analysis methods and to enhance their selectivity by reducing the water’s background signal. This work presents, for the first time, the coupling of a miniaturized silicon micro pre-concentration unit (µPC) to a bio-based opto-electronic nose (NeOse Pro). The results showed that the coupling of a silicon µPC with the NeOse Pro led to an improvement in the detection limit of n-nonane by at least a factor of 125. Additionally, principal component analysis (PCA) of eight different flavored waters showed an enhanced discrimination ability of the coupled set-up in highly humid conditions.

The Gender Gap in Bank Credit Access

We use a sample of over 80,000 Spanish companies started by a sole entrepreneur between 2004 and 2014, and distinguish between male and female entrepreneurs demand for credit, credit approval ratio, and credit performance. We find that female entrepreneurs who start a business are less likely to ask for a loan. Of the female entrepreneurs requesting a credit, the probability of obtaining one in the founding year is significantly lower than their male peers in the same industry. This lower credit access disappears over the subsequent years, once the company has a track record of profits and losses. We also observe that women-led companies that receive a loan in the founding year are less likely to default as compared to men-led companies. This superior performance disappears for subsequent years, coinciding with the disappearance of the lower credit access. Taking all these results together, we rule out both taste-based discrimination and statistical discrimination in the credit industry, and point to the possible presence of double standards which might be a consequence of implicit (unconscious) discrimination.

Disabling Effects of Enabling Social Policies on Gender Equality in Organizational HR Development

Human capital and its use on the labour market are associated with desirable outcomes for individuals, organisations and societies. A variety of social policies, such as paid parental leave and formal childcare, are designed to enhance parents’ labour force participation. These policies vary considerably in extent and nature between states, with different influence on men’s and women’s decision regarding their labour market supply and on organisations’ expectations towards labour availability, potentially affecting employers’ willingness to invest in their employees. Thus, social policies may indirectly influence individuals’ human capital and career development. Using a sample of over 13,500 individuals from 19 countries, we investigate the moderation effect of gender on the relationship between social policies and human resource development practices. Both policies have disabling effects for women, but not for men. We provide explanations for the counterintuitive result regarding the provision of childcare with reference to the statistical discrimination effect.

1H NMR-based metabonomics for infertility diagnosis in men with varicocele.

"Omics" techniques have been used to understand and to identify biomarkers of male infertility. We report on the first metabonomics models created to diagnose varicocele and infertility among men with varicocele. We recruited 35 infertile men with varicocele (VI group), 21 fertile men with varicocele (VF group) and 24 fertile men without varicocele (C group). All men underwent standard semen analysis, scrotal duplex ultrasonography, and sexual hormone level measurement. Hydrogen-1 nuclear magnetic resonance (1H NMR) spectra of seminal plasma were used to create metabonomics models to discriminate between men with and without varicocele, and between fertile and infertile men with varicocele. Using the statistical formalisms partial least square discriminants analysis and genetic algorithm-based linear discriminant analysis (GA-LDA), we created two models that discriminated the three groups from each other with accuracy of 92.17%. We also created metabonomics models using orthogonal partial least square discriminants analysis and GA-LDA that discriminated VF group from VI group, with an accuracy of 94.64% and 100% respectively. We identified 19 metabolites that were important in group segregation: caprate, 2-hydroxy-3-methylvalerate, leucine, valine, 3-hydroxybutyrate, lactate, alanine, 4-aminobutyrate, isoleucine, citrate, methanol, glucose, glycosides, glycerol-3-phosphocoline, n-acetyltyrosine, glutamine, tyrosine, arginine, and uridine. 1HNMR-based metabonomics of seminal plasma can be used to create metabonomics models to discriminate between men with varicocele from those without varicocele, and between fertile men with varicocele from those infertile with varicocele. Furthermore, the most important metabolites for group segregation are involved in the oxidative stress caused by varicocele.

Over-education among immigrants: the role of demographics, time, and firm characteristics

ABSTRACT This paper puts the nexus between immigration and over-education to an updated test, capitalizing on the authors’ access to rich Belgian matched employer-employee data for the period 1999–2010. These data enable the authors to i) measure over-education with higher precision, ii) examine the heterogeneous effects of workers’ birth countries, and iii) test the role of key moderators. Using ordered probit estimates, the authors highlight that workers born in developing countries (especially in Asia and North Africa) are much more likely to be over-educated than their opposite numbers born in developed countries. However, the results also show that this gap in over-education is entirely driven by tertiary-educated workers. Gender-based differences in immigrants’ penalties, in contrast, are found to be quite modest overall. In line with statistical discrimination, results further show that tenure has a moderating effect on the likelihood for immigrants born in developing countries to be over-educated and that citizenship acquisition is also associated with improved job matches for those immigrants. Finally, regarding the role of firm characteristics, the authors find that the likelihood for immigrants from developing countries to be over-educated is significantly smaller in bigger firms and when working conditions are collectively renegotiated at the firm level.

Gender bias and statistical discrimination against female instructors in student evaluations of teaching

This paper uses administrative data from a public university in Taiwan to examine gender bias in teaching evaluations. We test for statistical discrimination against female instructors using the employer learning model in which the instructor value-added to the grades of the current course and follow-on course is used to measure teaching effectiveness. The results show that statistical discrimination is a significant source of gender bias in teaching evaluations, especially among male students and in STEM departments where female faculty is underrepresented. Gender bias in teaching evaluations was reduced by nearly 50% after ten years of teaching. The results also suggest that the gender gap in teaching evaluations changes over time as male and female students evaluate male and female instructors differentially. Statistical discrimination is closely related to the underrepresentation of women in academia. For female students, the gender gap in evaluation scores narrows when the share of female faculty in the department rises. By contrast, male students are less sensitive to the percentage of female faculty in the department.

Gender Discrimination in the Hiring of Skilled Professionals in Two Male-Dominated Occupational Fields: A\xa0Factorial Survey Experiment with Real-World Vacancies and Recruiters in Four European Countries

The present article investigates gender discrimination in recruitment for two male-dominated occupations (mechanics and IT professionals). We empirically test two different explanatory approaches to gender discrimination in hiring; namely, statistical discrimination and taste-based discrimination. Previous studies suggest that, besides job applicants’ characteristics, organisational features play a role in hiring decisions. Our article contributes to the literature on gender discrimination in the labour market by investigating its opportunity structures located at the recruiter, job and company level, and how gender discrimination varies across occupations and countries.The analysed data come from a factorial survey experiment conducted in four countries (Bulgaria, Greece, Norway and Switzerland). Real job advertisements were sampled, and the recruiters in charge of hiring for these positions (n = 1,920) rated up to ten hypothetical CVs (vignettes). We find gender discrimination in Bulgaria and Greece and to a lesser degree in Switzerland, but not in Norway. The degree of gender discrimination appears to be greater in mechanics than in IT. Multivariate analyses that test a number of opportunity structures for discrimination suggest that mechanisms of statistical discrimination rather than those of taste-based discrimination might be at work.

Does employer learning with statistical discrimination exist in China? Evidence from Chinese Micro Survey Data

Because employers cannot perfectly recognize workers’ actual ability when recruiting, they have rational expectations for worker productivity only through considering easily observable characteristics such as education. However, with the increase in labour market experience, employers gradually learn actual productivity through historical working performance. The above phenomenon is called employer learning with statistical discrimination (EL-SD). This paper confirms that EL-SD significantly exists in the Chinese labour market based on Chinese Household Income Program (CHIP) data from 2013 and Chinese Family Panel Studies (CFPS) data from 2012 to 2014. Furthermore, EL-SD is hardly corroborated for advantaged workers such as males, college graduates, highly skilled workers, high wage earners, and workers in large businesses or the urban labour market.

Discrimination against people with disabilities in accessing microfinance

The purpose of this paper is to explore the state of existing research on microfinance lending discrimination against people with disabilities. It argues that existing literature suggests people with disabilities face notable discrimination in accessing microfinance (Labie et al., 2015). The attitudes of employees within microfinance institutions (MFIs) are one of the principal sources of such discrimination, which has important implications for the lives of people with disabilities (Cramm & Finkenflugel, 2008; Mersland et al., 2009; Labie et al., 2015). Moreover, studies conducted by Beisland & Mersland (2012) and Nuwagaba et al. (2012) found that people with disabilities tend not to apply for microfinance due to the anticipation of such rejection. People with disabilities face both ‘taste-based discrimination’ (prejudice) and ‘statistical discrimination’ (when a decision-maker uses individuals’ observable characteristics as a substitute for unobservable ones). It is reasonable to assume that reducing discrimination in the microfinance market would greatly benefit existing and prospective entrepreneurs with disabilities. However, this requires targeted interventions aimed at inducing systematic institutional reforms, changing the mindsets of employees and people with disabilities, and integrating appropriate accommodations within microfinance operations. In order to implement these changes, I conclude that microfinance regulators must work to identify areas of discrimination that are not correctly covered by existing lending practices.

Platon: identification and characterization of bacterial plasmid contigs in short-read draft assemblies exploiting protein sequence-based replicon distribution scores.

Plasmids are extrachromosomal genetic elements that replicate independently of the chromosome and play a vital role in the environmental adaptation of bacteria. Due to potential mobilization or conjugation capabilities, plasmids are important genetic vehicles for antimicrobial resistance genes and virulence factors with huge and increasing clinical implications. They are therefore subject to large genomic studies within the scientific community worldwide. As a result of rapidly improving next-generation sequencing methods, the quantity of sequenced bacterial genomes is constantly increasing, in turn raising the need for specialized tools to (i) extract plasmid sequences from draft assemblies, (ii) derive their origin and distribution, and (iii) further investigate their genetic repertoire. Recently, several bioinformatic methods and tools have emerged to tackle this issue; however, a combination of high sensitivity and specificity in plasmid sequence identification is rarely achieved in a taxon-independent manner. In addition, many software tools are not appropriate for large high-throughput analyses or cannot be included in existing software pipelines due to their technical design or software implementation. In this study, we investigated differences in the replicon distributions of protein-coding genes on a large scale as a new approach to distinguish plasmid-borne from chromosome-borne contigs. We defined and computed statistical discrimination thresholds for a new metric: the replicon distribution score (RDS), which achieved an accuracy of 96.6 %. The final performance was further improved by the combination of the RDS metric with heuristics exploiting several plasmid-specific higher-level contig characterizations. We implemented this workflow in a new high-throughput taxon-independent bioinformatics software tool called Platon for the recruitment and characterization of plasmid-borne contigs from short-read draft assemblies. Compared to PlasFlow, Platon achieved a higher accuracy (97.5 %) and more balanced predictions (F1=82.6 %) tested on a broad range of bacterial taxa and better or equal performance against the targeted tools PlasmidFinder and PlaScope on sequenced Escherichia coli isolates. Platon is available at: http://platon.computational.bio/.

A natural experiment on discrimination in elections

We exploit a natural experiment to study discrimination in elections. In Illinois Republican presidential primaries, voters vote for delegates bound to presidential candidates, but delegates' names convey information about their race and gender. We identify discrimination from variation in vote totals among delegates bound to the same presidential candidate and who face the same voters. Examining delegate vote totals from 2000 to 2016, we estimate nonwhite delegates receive 9% fewer votes. We find essentially no gender discrimination. Negligible incentives for statistical discrimination, costs to preferred presidential candidates, and heterogeneity are consistent with an interpretation of this behavior as taste-based.

Discrimination and Deterrence with Enforcer Liability

Abstract Taste-based discrimination (i.e. discrimination due to racist preferences) receives more attention than statistical discrimination in the enforcement literature, because the latter allows enforcers to increase their “success rates.” I show here that when enforcers’ incentives can be altered via liabilities and rewards, all types of discrimination reduce deterrence. Moreover, adverse effects of statistical discrimination on deterrence are more persistent than similar effects due to taste-based discrimination. I identify crime minimizing liabilities and rewards when enforcers engage in racial discrimination and consider the robustness of the analysis in alternative settings.