Demographic aspects of first names

Tzioumis, Konstantinos

doi:10.1038/sdata.2018.25

Cited by 70 publications

(73 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In other countries, surname lists have been utilized to identify religious groups and those of Middle Eastern descent (Mateos, 2007). Others have created first-name based lists to identify R/E using first name and surname, finding that first names might more accurately identify White individuals than surnames (Tzioumis, 2018). As reference name lists continue to be created, Mateos (2007) cautions researchers to make sure that these lists were based on a large enough population to make valid inferences and to be aware of temporal differences, regional differences, differences in average ratio of people per surname, history of name adoption, and surnames reflecting only patrilineal descent.…”

Section: Resultsmentioning

confidence: 99%

How Administrative Data Collection and Analysis Can Better Reflect Racial and Ethnic Identities

Viano

Baker

2020

Review of Research in Education

View full text Add to dashboard Cite

Measuring race and ethnicity for administrative data sets and then analyzing these data to understand racial/ethnic disparities present many logistical and theoretical challenges. In this chapter, we conduct a synthetic review of studies on how to effectively measure race/ethnicity for administrative data purposes and then utilize these measures in analyses. Recommendations based on this synthesis include combining the measure of Hispanic ethnicity with the broader racial/ethnic measure and allowing individuals to select more than one race/ethnicity. Data collection should rely on self-reports but could be supplemented using birth certificates or equivalent sources. Collecting data over time, especially for young people, will help identify multiracial and American Indian populations. For those with more complex racial/ethnic identities, including measures of country of origin, language, and recency of immigration can be helpful in addition to asking individuals which racial/ethnic identity they most identify with. Administrative data collection could also begin to incorporate phenotype measures to facilitate the calculation of disparities within race/ethnicity by skin tone. Those analyzing racial/ethnic disparities should understand how these measures are created and attempt to develop fieldwide terminology to describe racial/ethnic identities.

show abstract

Section: Resultsmentioning

confidence: 99%

How Administrative Data Collection and Analysis Can Better Reflect Racial and Ethnic Identities

Viano

Baker

2020

Review of Research in Education

View full text Add to dashboard Cite

show abstract

“…All names appear in each source, except Maile does not appear for the last source. 23 For example, there is no Census or Social Security Administration tabulation of first names by race as there is for last names (Tzioumis 2018) and there is little information that suggests that Native American or Alaska Native first names are sufficiently common. Furthermore, no Alaska Native-specific names appear in the Social Security database in Alaska for the years 1985-1987.…”

Section: First Name As An Indigenous Signal (Native Hawaiian Only)mentioning

confidence: 99%

Employment discrimination against Indigenous Peoples in the United States: Evidence from a field experiment

Button

Walker

2020

Labour Economics

View full text Add to dashboard Cite

We conducted a resume correspondence experiment to measure discrimination in hiring faced by Indigenous Peoples in the United States (Native Americans, Alaska Natives, and Native Hawaiians). We sent employers realistic 13,516 resumes for common jobs (retail sales, kitchen staff, server, janitor, and security) in 11 cities and compared callback rates. We signaled Indigenous status in one of four different ways. We almost never find any differences in callback rates, regardless of the context. These findings hold after numerous robustness checks, although our checks and discussions raise multiple concerns that are relevant to audit studies generally.

show abstract

“…Because the data points do not have names associated with them, we generate synthetic first names using the race and gender attributes. First, we use the dataset of Tzioumis (2018) to identify "white" and "nonwhite" names. For each name, if the proportion of "white" people with that name is higher than 0.5, we deem the name to be "white;" otherwise, we deem it to be "non-white."…”

Section: Datasetsmentioning

confidence: 99%

“…And, because biographies are typically written in the third person and because pronouns are gendered in English, we can extract (likely) self-identified gender. We infer race for each data point by sampling from a Bernoulli distribution with probability equal to the average of the probability that an individual with that first name is "white" (from the dataset of Tzioumis (2018), using a threshold of 0.5, as described above) and the probability that an individual with that last name is "white" (from the dataset of Comenetz (2016), also using a threshold of 0.5). 3 Finally, like , we consider two versions of the Bios dataset: one where first names and pronouns are available to the classifier and one where they are "scrubbed.…”

Section: Datasetsmentioning

confidence: 99%

What’s in a Name?

Romanov¹,

De‐Arteaga²,

Wallach³

et al. 2019

Proceedings of the 2019 Conference of the North

View full text Add to dashboard Cite

There is a growing body of work that proposes methods for mitigating bias in machine learning systems. These methods typically rely on access to protected attributes such as race, gender, or age. However, this raises two significant challenges: (1) protected attributes may not be available or it may not be legal to use them, and (2) it is often desirable to simultaneously consider multiple protected attributes, as well as their intersections. In the context of mitigating bias in occupation classification, we propose a method for discouraging correlation between the predicted probability of an individual's true occupation and a word embedding of their name. This method leverages the societal biases that are encoded in word embeddings, eliminating the need for access to protected attributes. Crucially, it only requires access to individuals' names at training time and not at deployment time. We evaluate two variations of our proposed method using a large-scale dataset of online biographies. We find that both variations simultaneously reduce race and gender biases, with almost no reduction in the classifier's overall true positive rate.

show abstract

Demographic aspects of first names

Cited by 70 publications

References 7 publications

How Administrative Data Collection and Analysis Can Better Reflect Racial and Ethnic Identities

How Administrative Data Collection and Analysis Can Better Reflect Racial and Ethnic Identities

Employment discrimination against Indigenous Peoples in the United States: Evidence from a field experiment

What’s in a Name?

Contact Info

Product

Resources

About