Sterilization Racism and Pan-Ethnic Disparities of the Past Decade:The Continued Encroachment on Reproductive Rights
In the late 1960s and through the 1970s, reports of coercive, involuntary, and otherwise nonconsenting sterilizations of American Indian, African American, Mexican, and Puerto Rican origin women began surfacing in the United States.1 These revelations came at a time of intense civil rights activity and political consciousness among non-"white"2 groups in the United States. American Indian and African American women and girls were especially impacted by sterilization abuse.
In a well-known case, recounted by Jane Lawrence, Dr. Connie Pinkerton-Uri saw a twenty-six-year-old patient in early November 1972 who visited her clinic and requested a "womb transplant."3 It turned out that the woman was given a full hysterectomy (for alcoholism) at age twenty after being told by an Indian Health Services (IHS) doctor that the procedure was reversible. Other scholars have noted cases of American Indian women receiving hysterectomies as young as age eleven.4 These cases are similar to the experiences of African American women and girls, such as the Relf sisters, ages twelve and fourteen, who were the unwilling and unknowing recipients of tubal sterilization as well as guinea pigs for intrauterine devices and what were then experimental Depo-Provera shots (along with their older sister, Katie) in the early 1970s.5 African American civil rights leader Fannie Lou Hamer was compelled to get involved in the modern civil rights movement, in part, after receiving a "Mississippi Appendectomy" [End Page 17] (an unbeknownst hysterectomy) upon visiting a doctor to have a benign fibroid uterine tumor removed.6
Many recent studies have taken stock of the eugenics history of the United States.7 Furthermore, there have been several historical studies of forced sterilization of American Indian women that covered cases from the 1970s.8 While such scholarship has provided rich insight into specific cases from the 1970s, very little is known about contemporary pan-ethnic differences in sterilization.9 One recent study, in the medical journal Obstetrics and Gynecology and based on the 2002 National Survey of Family Growth, found that African American women were more likely than European American women to have undergone tubal sterilization.
I extend this work here by focusing on contemporary pan-ethnic disparities in tubal sterilization using a different dataset (the CDC's 2004 Behavioral Risk Factor Surveillance System survey), developing a concept of sterilization racism and comparing African Americans, American Indians, and European American women. This study is motivated by the recent outpouring of scholarship on sterilization abuse directed at women of color and the paucity of contemporary large-sample statistical studies of pan-ethnic and other ethnic disparities in women's sterilization. As with any quantitative study of pan-ethnic disparities, unequal results are not definitive proof that racism is driving the differences. Instead, the statistical analysis in this paper will adjust for differences to rule out alternative explanations for the disparities and will provide a controlled statistical portrait of sterilization among women of color and European American women.10 I now turn to a theoretical discussion of the causes of disparities by conceptualizing sterilization racism as an important factor shaping reproductive healthcare provision in the United States.
Conceptualizing Sterilization Racism
The origins of racism and racial oppression in the Americas are the result of European conquest and colonization. The early colonization process involved various forms of population control. European colonizers established a system of capital accumulation in the Americas, in part, by controlling the population sizes of American Indians and African Americans. In the case of American Indians, various genocidal and "removal" policies aimed at women and children, by European colonizers and their descendants, were implemented to free territory on which to build plantations.11 In the case of African Americans, people were sold and kidnapped, and women were forced to reproduce in order to provide a laboring population that fueled the slave-based mode of production.12 This history is important because it suggests that controlling the reproduction of women of color has a long history that shares continuity with contemporary sterilization practices. [End Page 18]
In this study, I argue that contemporary sterilization racism occurs because the United States is what Joe Feagin calls a "total racist society."13 Following sociologists Noel A. Cazenave and Darlene Alvarez Maddern, racism is defined as ". . . a highly organized system of race-based group privilege that operates at every level of society and is held together by a sophisticated ideology of color/race supremacy."14 Part of this racist system includes various oppressive ideological constructs that define, categorize, and place different pan-ethnic groups in a hierarchy.15 The racist hierarchy of "whites" on the top and people of color on the bottom maps on to an ordering of reproductive rights. In this ordering, European American women are least likely to have external authorities (e.g., the state, reproductive healthcare providers) constraining their reproductive abilities, while women of color are most likely to have such institutions influencing their reproductive lives. This system is held together and supported by racist controlling images. Racist controlling images are racist and sexist stereotypes and caricatures of people of color that sustain racial, class, and gender oppression.16 African American and American Indian women are targeted by distinct racist controlling images.
Racist Controlling Images of American Indian Women
American Indian women have been represented in mainstream America media as either a Squaw or an Indian Princess, according to some scholars.17 The Squaw image depicts American Indian women as dirty, subservient, abused, alcoholic, and ugly, and as women who love to torture "white" men, while the Indian Princess image depicts American Indian women as exotic, beautiful "princesses" who leave their society to elope with suave European American men.18 The Squaw image suggests a dirty and sinful body in need of cleansing and, consequently, sterilization.19 The Squaw is the "darker twin" of the Indian Princess.20 The Squaw has the same vices as "Indian" men, such as drunkenness, thievery, and stupidity.21 As a controlling image of American Indian women's reproductive behaviors, Squaws may live with "Indian" men and ". . . work for their lazy bucks and bear large numbers of fat 'papooses.'"22 Unlike the Indian Princess, the Squaw may be overweight and have a darker skin tone, a problem with alcohol, and fewer "European features" than the Indian Princess. Another characteristic of the Squaw is that she is not capable of the same human emotions as "white" women and consequently neglects her children.23 Thus, the Squaw controlling image may be a pervasive stereotype that reproductive healthcare providers rely on to suggest sterilization for American Indian women. Instead of seeing an individual woman, reproductive healthcare providers may see a "Squaw" and label her as someone whose reproductive abilities threaten the continued colonization of American Indian people.24 Related yet distinct racist controlling images of African Americans are also prevalent in U.S. media and culture. [End Page 19]
Racist Controlling Images of African American Women
"White"-controlled and -owned media and cultural producers promoted Jezebel images of African American women during slavery. The Jezebel is the stereotype of a sexually aggressive "black" woman promoted by media during slavery to explain the high birth rates of bonded women and numerous allegations of sexual assault claimed by enslaved women.25 These images functioned to justify the sexual violence inflicted by European American male slave owners on women and children of African descent. European American male slave owners used enslaved women to "breed" future generations of people held in bondage, who were accumulated as wealth holdings. Thomas Jefferson recognized this when he stated, "I consider a woman who brings a child every two years as more profitable than the best man on the farm. . . . What she produces is an addition to the capital, while his labors disappear in mere consumption."26
Between the 1960s and through the 2000s, various racist controlling images emerged to justify controlling African American women's reproduction. The Welfare Queen image emerged in the 1960s as African American women gained access to public assistance. This racist image suggests that African American women have babies simply to enrich themselves with "welfare money."27 This is quite similar to the stereotype that American Indian women collect public assistance money to buy alcohol.28 Other racist images of "black" women as "Matriarchs," "Baby Mamas," and "Welfare Queens" were created in the post-civil rights movement era by social scientists29—possibly as part of a "white" racial backlash against the perceived overzealousness of the modern civil rights movement.30 Just as in the case of American Indian women, reproductive healthcare providers may not see an individual African American woman patient but instead rely on racist controlling images of Jezebels, Hoochie Mamas, and/or Welfare Queens to guide the type of reproductive healthcare they give to women of African descent. In light of this discussion, I offer the following concept of sterilization racism.
Sterilization Racism is defined as the organization of racist controlling images, policies, and practices of delivering reproductive healthcare that operate to constrain, minimize, or completely eliminate the reproductive activities of women of color.31 In offering this definition, I argue that racist controlling images become solidified as justificatory ideologies for continued systemic sterilization abuse.
Figure 1 depicts how sterilization racism may operate. In the context of seeking reproductive healthcare services, a woman interacts with a reproductive healthcare provider. The reproductive healthcare provider views the patient through the cognitive lens of stereotypes and racist controlling images. If the woman appears "black," then she may be defined and discriminated against accordingly as a Jezebel, Hoochie Mama, and/or Welfare Queen. Since these controlling images suggest a sexually "loose" woman who cannot control her sexual urges [End Page 20] and who may already have too many children, then sterilization may be suggested because it permanently prevents conception.
If the woman appears "Indian," then she may be viewed as a Squaw who has "painless childbirths"32 and is "too dumb to use birth control," who consequently should be encouraged to get her tubes tied. However, if the woman appears "white," then she is likely to be positively stereotyped as a competent user of birth control; in the event that she has had no or few children, and given that she may change her mind, she may not be encouraged to undergo sterilization (and she may be dissuaded from undergoing sterilization if she requests the procedure).33 My theoretical model leads to the following hypothesis: American Indian and African American women are more likely to have been sterilized, net of confounding factors, than European American women. I now turn to a discussion of the quantitative methods used to test this hypothesis.
Data and Methods
The data used in this study is the 2004 Behavioral Risk Factor Surveillance System (BRFSS) survey conducted by the Centers for Disease Control and Prevention (CDC) and administered by individual states. The BRFSS is designed with the aim of measuring behavioral risk factors among the population age eighteen and older living in households. Coverage includes the fifty states plus other U.S. territories (D.C., Guam, Puerto [End Page 21] Rico, and the Virgin Islands). The sampling design is a disproportionate stratified sample (DSS) of households with telephones. Each state, in most cases, is a single stratum and, within each state, telephone numbers are divided into three groups of high-, medium-, and low-density strata sampled separately to obtain the probability sample of all households with telephones. Data for each state is collected either directly by the state health department or through a subcontractor. In the 2004 BRFSS, the survey was conducted by state health departments, university survey research centers, or commercial firms. Interviews are conducted using computer-assisted telephone interviewing. State personnel and contractors follow CDC protocols and must be certified to work on the BRFSS. In the 2004 wave, a core module of questions on family planning methods was included in the survey.
The dataset contained 303,822 completed interviews. This analysis is restricted to women age eighteen through forty-four who did not have a hysterectomy, were not currently pregnant, were using a form of contraception to prevent pregnancy, and lived within the fifty states or District of Columbia. The sample is further restricted to African American, American Indian, and European American respondents who did not have missing values on the variables of interest, yielding a sample size of 32,941 women. The outcome variable in the study is whether a woman was currently using tubal sterilization (at the time of interview) versus another method of contraception (pill, condoms, foams/jelly/creams, diaphragm, IUD, injectables, implants, withdrawal, rhythm, vasectomy, or other methods).
The main independent variables of interest are categorical indicators of pan-ethnicity: African American ("Non-Hispanic Black"), American Indian ("Non-Hispanic American Indian or Alaskan Native"), and European American ("Non-Hispanic White"). Potential confounders are entered into the logistic regression equation and include age, number of children under age eighteen living in the respondent's household, marital status, household income, whether or not the respondent had insurance at the time of the interview, the state of residence, and educational attainment. Past research suggests that age and parity (estimated as the number of children under age eighteen living in the household) should be the best predictors of tubal sterilization. All statistical analyses were programmed in SUDAAN version 9.0.3 to make the appropriate adjustment for the complex survey sampling design of the BRFSS. This study was approved by the Institutional Review Board of the University of Connecticut.
Analysis of Results
Figure 2 presents a bar chart of pan-ethnic disparities in tubal sterilization among women age eighteen through forty-four. American Indian women have the highest rate of tubal sterilization, at 33.9 percent. [End Page 22]
Following closely behind American Indian women are African American women, with a tubal sterilization rate of 30.1 percent. European American women have the lowest rate of tubal sterilization, at 18.7 percent. Using European American women as a basis for comparison, the differences in percentages are all statistically significant (American Indian women, z = 3.71, p < .001; African American women, z = 9.32, p < .001). African American and American Indian women are not statistically significantly different from each other in the percentage sterilized (z = 0.88, p = .367).
The differences raise many questions and some skeptics might aver that these differences in raw sterilization rates are nothing to get excited about because they are not adjusted for other factors (e.g., age and parity) that may influence the prevalence of tubal sterilization. Table 1 presents the proportion of women within each pan-ethnic group with each given characteristic.
For example, a critic might argue that American Indian and African American women have higher sterilization rates because American Indian and African American women have higher fertility rates and, when you control for fertility, the disparities should disappear. Similarly, one could make the same argument about age, education, income, marital status, and insurance status and assert that, once these variables are controlled, all disparities should vanish.
In Table 2a logistic regression model is fit to the 2004 BRFSS data where the dependent variable is whether or not the respondent is relying on tubal ligation for contraception (1 = yes, 0 = no). As expected, women with more children under age eighteen living in their household are more likely to use tubal sterilization compared to women with no children under age eighteen living in their household. There is not a statistically significant difference between women with no children [End Page 23] under age eighteen and women with a single child under age eighteen living in their household at p < .001. Similarly, older women are more likely to have undergone tubal sterilization.
Class may also play a role because women in households with higher incomes are less likely to have been sterilized compared to [End Page 24] women on lower incomes. Thus, a woman from a household with an annual income of $75,000 or higher has odds of tubal sterilization about 74 percent lower than a woman from a household with an annual income of $10,000 or less. Similarly, women with higher levels of educational completion have a lower likelihood of tubal sterilization. A college graduate, net of all other factors, is much less likely to undergo tubal sterilization than a high school dropout (84 percent lower odds). Women with no health insurance are also more likely to
[End Page 25]
have undergone tubal sterilization (76 percent higher odds than for a woman with health insurance). Women who have ever been married (currently married, divorced, widowed, or separated) are more likely than never married women to have been sterilized. The model also includes fifty categorical indicators (Wyoming as reference) for all states and the District of Columbia. The inclusion of these variables adjusts for any state-to-state variation in the likelihood of tubal sterilization not explained by the other independent variables.
As previously stated, a critic might argue that the aforementioned factors should explain away pan-ethnic disparities in sterilization. This argument is rejected by results of the fitted logistic regression model. Both African American and American Indian women have a greater likelihood of tubal sterilization than European American women (p < .001). Since p-values are less reliable in large samples, the BIC is also reported for individual coefficients.34 For American Indian women, the evidence is positive as the BIC = 3.0 and, for African American women, BIC = 32.4, indicating strong evidence against [End Page 26] the null hypothesis of no difference between these women of color and European American women. Net of all other variables included in the model, African American women have odds of tubal sterilization increased by 1.75 above European American women. American Indian women have odds of tubal sterilization increased by 2.23 above European American women. The odds of tubal sterilization, net of other variables controlled in the model, are 123 percent greater for American Indian women and 75 percent greater for African American women compared to European American women.
We can better understand the disparities by computing predicted probabilities from the logistic regression results and then comparing them between pan-ethnic groups. The first column in Table 2 presents the logit estimates. These are the expected change in the log odds of sterilization, given a unit difference in the respective independent variable. Since all independent variables are measured as categorical indicators (i.e., dummy variables), then each logit estimate is the difference in the log odds of sterilization comparing a given group with the reference group, holding all else constant (e.g., American Indian women vs. European American women, a difference in log odds of 0.80).
Since changes in the log odds and odds ratios do not have an intuitive interpretation, we should instead look at differences in the predicted probabilities. To compute the predicted probability for a woman with a given set of characteristics, we simply add up the appropriate logits plus intercept and then find the antilog of the predicted logit divided by 1.0 plus itself. The most simple case is a woman represented by all reference groups. This is a European American woman, age eighteen to twenty-four, who has no children under age eighteen in the household, has never been married, has an annual household income less than $10,000, did not complete high school, and has no health insurance. Her predicted log odds are -2.66 and the predicted probability of such a woman using tubal sterilization for contraception is [(exp2.66)/(1 + exp2.66)] = 0.07. A similarly situated African American woman has a probability of using tubal sterilization for contraception of 0.11 = [(exp2.66 + 0.56)/ (1 + exp2.66 + 0.56)]; a similarly situated American Indian woman has a predicted probability of being sterilized of 0.13 = [(exp2.66 + 0.80)/(1 + exp2.66 + 0.80)]. While a woman with these characteristics has a low risk of sterilization, the pan-ethnic disparities are notable with the probability nearly twice as high for American Indian women and significantly higher for African American women compared to European American women.
Age and number of children under age eighteen living in the household increase the likelihood of sterilization. A useful comparison is to look at younger and older women, holding the number [End Page 27] of children under age eighteen living in the household constant at two. Consider a woman who is age twenty-five to twenty-nine; has been divorced, widowed, or separated; has two children under age eighteen in the household; has an income of $25,000-34,999; and has a high school degree and health insurance. A European American woman with these characteristics has a predicted 0.33 probability of sterilization compared to a probability of 0.47 for an African American woman and 0.52 for an American Indian woman. Likelihood of sterilization increases with age, such that the same hypothetical woman, at age thirty-five to thirty-nine, if European American, has a probability of sterilization of 0.69 compared to 0.80 for her African American counterpart and 0.83 for her American Indian counterpart.
The results of the statistical analysis based on the data from the 2004 BRFSS indicate that both African American and American Indian women are at increased risk of using tubal sterilization for contraception compared to similarly situated European American women. The disparity remains after controlling for age, number of children under age eighteen living in the household, marital status, income, education, insurance status, and state of residence. The disparity is noteworthy because of the history of coerced, non-consenting, and otherwise deceptive sterilizations that American Indian and African American women experienced.
The statistical data is merely suggestive, as it reveals patterns but does not elucidate the underlying causal processes that generate the pan-ethnic disparities. Some women may be choosing sterilization as a means of contraception, while other women may be pressured into sterilization by healthcare providers. Andrea Smith has argued that the "pro-choice vs. pro-life" paradigm is a false dichotomy because the experiences of American Indian women, and other women of color who have experienced racist sterilization policies aimed at curtailing the population sizes of various non-"white" groups, are important for understanding contemporary patterns. Physicians may be all too eager to sterilize women of color, as in the case of one physician, who, when asked why he sterilized a "white" woman, stated, "It was done like in Alabama. You sterilize the black girls, except this is elective. . . . It was something social, you elect to sterilize this person." Bahati Kuumba argues that women of African descent have been targeted for the colonization of their wombs as a form of "reproductive imperialism."35 Such forms of reproductive imperialism and colonialism may be in operation for American Indian women as well. [End Page 28]
The statistical analysis presented in this article and the patterns discussed herein are provocative and require further study. Given historical charges of genocide and documented cases of abuse, the high contemporary proportion of sterilized African American and American Indian women cannot be dismissed lightly.
Thomas W. Volscho is assistant professor of sociology at the City University of New York-College of Staten Island. In his dissertation, he developed a theory of sterilization racism to explain high rates of sterilization and long-acting hormonal contraceptive use among women of color. He is currently working on a book entitled The Revenge of the Capitalist Class, which explores the connection between rising inequality and increasing state repression.
I want to thank Noel Cazenave, Evelyn Simien, Xae Alicia Reyes, and Simon Cheng for challenging me and strengthening this work.
1. Jennifer Nelson, Women of Color and the Reproductive Rights Movement (New York: New York UniversityPress, 2003); and Dorothy Roberts, Killing the Black Body: Race, Reproduction, and the Meaning of Liberty (New York: Vintage Books, 1997).
2. Theodore Allen, The Invention of the White Race (London: Verso, 1994).
3. Jane Lawrence, "The Indian Health Service and the Sterilizationof Native American Women," American Indian Quarterly 24, no. 3 (2000): 400-19; D. Marie Ralstin-Lewis, "The Continuing Struggle against Genocide: Indigenous Women's Reproductive Rights," Wicazo Sa Review: A Journal of Native American Studies 20, no. 1 (2005): 71-95; and Andrea Smith, Conquest: Sexual Violence and American Indian Genocide (Boston: South End Press, 2005).
4. Myla Vincenti Carpio, "The Lost Generation: American Indian Women and Sterilization Abuse," Social Justice 31, no. 4 (2004):40-53.
5. Nancy Ordover, American Eugenics: Race, Queer Anatomy, and the Science of Nationalism (Minneapolis: University of Minnesota Press, 2003).
6. Harriet Washington, Medical Apartheid: The Dark History of Medical Experimentation on Black Americans from Colonial Times to the Present (New York: Doubleday, 2006).
7. See, for example, Edwin Black, War against the Weak: Eugenics and America's Campaign to Create a Master Race (New York: Four Walls Eight Windows, 2003); Harry Brunius, Better for All the World: The Secret History of Forced Sterilization and America's Quest for Racial Purity (New York: Knopf, 2006); and Ordover, American Eugenics.
8. Sally J. Torpy, "Native American Women and Coerced Steriliza-tion: On the Trail of Tears in the 1970s," American Indian Culture and Research Journal 24, no. 2 (2000): 1-22; Lawrence, "The Indian Health Service and the Sterilization of Native American Women"; Carpio, "The Lost Generation"; and Ralstin-Lewis, "The Continuing Struggle against Genocide."
9. There is no biological evidence for the "race" concept. SociologistNoel A. Cazenave has argued that the very concept of race is confusing, injurious, and erroneous. In this article, I do not use racist color-coded terms, such as "red," "white," or "black," to refer to survey respondents. Instead, I use the imperfect alternative of ethnic or pan-ethnic identifiers: American Indian, European American, and African American. See Noel A. Cazenave, "Conceptualizing 'Race' and Beyond," Association of Black Sociologists Newsletter (2004):4-6; Audrey Smedley, Race in North America: Origins and Evolution [End Page 29] of a Worldview (Boulder, Colo.: Westview Press, 2007); and Kimberly Tallbear, "DNA, Blood, and Racializing the Tribe," Wicazo Sa Review: A Journal of Native American Studies 18, no. 1 (2003): 81-107.
10. Some readers may wonder why I did not include Latinas in this study. I did this because detailedethnic identifiers are not available in the BRFSS dataset that I use in this study. The experiences of Puerto Rican, Mexican, and other Latinas may be obscured by relying on "Hispanic"as a crude pan-ethnic or substitute for a national ethnic identifier. For recent work on Latinas and sterilization abuse, see Leo R. Chavez, The Latino Threat: Constructing Immigrants, Citizens, and the Nation (Berkeley: University of California Press, 2008); Elena Gutiérrez, Fertile Matters: The Politics of Mexican Origin Women's Reproduction (Austin: University of Texas Press, 2008); and Iris López, Matters of Choice: Puerto Rican Women's Struggle for Reproductive Freedom (New Brunswick, N.J.: Rutgers University Press, 2008).
11. David E. Stannard, American Holocaust: The Conquest of the New World (New York: Oxford University Press, 1993); and Smith, Conquest.
12. Carter A. Wilson, Racism: From Slavery to Advanced Capitalism (Thousand Oaks, Calif.: Sage, 1996).
13. Joe R. Feagin, Racist America: Roots, Current Realities, and Future Reparations (New York: Routledge, 2000), 26.
14. Noel A. Cazenave and Darlene Alvarez Maddern, "Defending the White Race: White Male Faculty Opposition to a 'White Racism' Course," Race and Society 2, no. 1 (1999): 25-50.
15. Here I am extending Patricia Hill Collins's concept of "controlling images" and calling them racist controlling images. See Patricia Hill Collins, Black Feminist Thought (New York: Routledge, 2000).
16. Noel A. Cazenave and Kenneth J. Neubeck, Welfare Racism: Playing the Race Card against America's Poor (New York: Routledge, 2001), 32.
17. Devon A. Mihesuah, American Indians: Stereotypes and Realities (Gardena, Calif.: Clarity Press, 1996); and S. Elizabeth Bird, "Gendered Construction of the American Indian in Popular Media," Journal of Communication 49, no. 3 (1999): 61-83.
18. Mihesuah, American Indians, 61.
19. See Smith, Conquest, 92.
20. Rayna Green, "The Pocahontas Perplex: The Image of Indian Women in American Culture," The Massachusetts Review 16, no. 4 (1975): 698-714.
21. Bird, "Gendered Construction of the American Indian in Popular Media."
22. Green, "The Pocahontas Perplex," 711.
23. Bird, "Gendered Construction of the American Indian in Popular Media"; see also Angie K. Beeman, "Emotional Segregation: A Content Analysis of Institutional Racism in US Films, 1980-2001," Ethnic and Racial Studies 30, no. 5 (2007):687-712.
24. Smith, Conquest, 79.
25. K. Sue Jewell, From Mammy to Miss America and Beyond: Cultural Images and the Shaping of U.S. Social Policy (New York: Routledge, 1993); and Roberts, Killing the Black Body. [End Page 30]
26. Quoted in Mary Beth Norton, Liberty's Daughters: The Revolutionary Experience of American Women, 1750-1800 (Ithaca, N.Y.: Cornell University Press, 1996), 96.
27. Neubeck and Cazenave, Welfare Racism.
28. Beatrice Medicine, Drinking and Sobriety among the Lakota Sioux (Lanham, Md.: Altamira Press, 2006).
29. Collins, Black Feminist Thought; Neubeck and Cazenave, Welfare Racism; and Evelyn M. Simien, Black Feminist Voice in Politics (Albany: State University of New York Press, 2006).
30. Noel. A. Cazenave, Impossible Democracy: The Unlikely Success of the War on Poverty Community Action Programs (Albany: State University of New York Press, 2007).
31. This definition is inspired by Neubeck and Cazenave's definition of welfare racism. See Neubeck and Cazenave, Welfare Racism, 36.
32. Smith, Conquest.
33. Low-income European American women may be stereotyped as "irresponsible reproducers" compared to higher-income European American women. Thus, sterilization outcomes of poor European Americans may be shaped by class oppression. The assumption in my model is that class differences are held constant (i.e., controlled for).
34. Fred C. Pampel, Logistic Regression: A Primer (Thousand Oaks, Calif.: Sage, 2000).
35. Bahati Kuumba, "Population Policy in the Era of Globalization: A Case of Reproductive Imperialism," Agenda 48, no. 1 (2001): 22-30. [End Page 31]