A comprehensive analysis of sex and race inequities in unemployment insurance benefits.

This research makes a unique contribution to the growing body of literature on the welfare system by examining the relationship between sex, race, and social insurance benefits in a rural state. Using data from the West Virginia Unemployment Compensation Program, this research investigates sex and race differences in (1) monetary disqualifications for unemployment insurance (UI) benefits and (2) separation issue and nonseparation issue disqualifications of UI benefits. The analyses indicate that unemployed women, people of color, younger, and low income workers are the most likely to fail the monetary qualifications for UI benefits and to lose qualified weeks of UI benefits.

Keywords: gender, inequities, racial inequities, unemployment, unemployment insurance, benefits, distribution of benefits, social insurance benefits

Introduction

An individual's occupation determines his or her social status, income, potential for advancement, type of benefits, the potential for unemployment, and even the resources available to them if they become unemployed. The type of jobs available to workers in the United States has changed drastically over the last 25 years. High-wage manufacturing jobs have been steadily replaced by part-time, contingency, and/or service employment (Schram 1995; Wilson 1997). Jobs in the service industry are known for their minimum wage dependence, lack of unionization, limited job benefits, and limited job security. Consequently, as the service sector continues to grow, so does the potential economic vulnerability and hardship of our labor force.

The Unemployment Insurance (UI) program is one part of the welfare system that is supposed to address the problems associated with such labor market vulnerability. The question many social scientists have raised is the overall effectiveness and fairness of this program, particularly for women and people of color (Amott 1990; Bassi and McMurrer 1997; Bingham 1995; Blaustein 1993; Gordon 1994; Latimer 1999; Maranville 1992; Mink 1990; Nelson 1990; Pearce 1990; Pearce 1986; Pearce and McAdoo 1981).

This research project makes a unique contribution to the growing body of literature on the welfare system by examining the relationship between sex, race, and social insurance benefits for unemployed West Virginia workers. The primary purpose of this research is to investigate state level trends in receipt of unemployment insurance (UI) for workers who are unemployed and have applied for UI in West Virginia in 1997. Using the feminist scholarship on the welfare state as the overall theoretical framework, this research investigates sex and race differences in (1) unemployment insurance qualification (i.e., whether or not workers monetarily qualify for unemployment insurance benefits once they become unemployed), (2) separation issue disqualifications of UI benefits (i.e., losing eligible weeks of benefits because they voluntarily quit their job, were fired for misconduct, etc.), and (3) nonseparation issue disqualifications of UI benefits (i.e., losing eligible weeks of benefits because they failed to register for work did not accept suitable employment, etc.).

There are several reasons why extensive state-level data are required for this research. First, the major monetary qualification for UI benefits (i.e., the minimum earnings requirement) is state specific (Nicholson 1997). Second, variations in UI receipt may be partly a function of state government policy on unemployment insurance and local variation in administration. Thus, national level data obscures the discretionary power of individual states. In addition, national samples typically do not contain enough cases from rural states to analyze the issues. Even with large data sets that oversample disadvantaged workers (i.e., the National Longitudinal Surveys of Youth-NLSY), many critical questions are left unanswered because the data sets are not designed to examine patterns in UI receipt.

Research on a rural state's social insurance program is particularly significant given (1) the majority of the research on poverty and the welfare system has focused on poor families in urban areas (Rural Sociological Society Task Force 1993), (2) the poverty rates in rural areas are consistently higher than those in urban areas and these rates are not declining (Zimmerman et al 1999), (3) researchers have already documented rural deficits in public assistance benefits (Amott 1990; Pearce 1989; Rank 1994; Tickamyer et al. 1993), (4) economic inequality has grown over the past 25 years (particularly the socioeconomic gap between rural and urban areas) while the actual percentage of workers qualifying for unemployment insurance benefits has declined, and (5) one of the best ways to test the UI program (i.e., a program designed to provide income security to displaced workers) is to examine its effectiveness in helping economically vulnerable workers in a rural state with consistently the highest unemployment and poverty rates in the nation.

This research on West Virginia's social insurance program adds to our understanding of stratification by examining sex and race inequities in state policies for disadvantaged workers. This research is particularly relevant in a postindustrial society where global competition and continued institutionalized occupational and income inequality compound the historic problems with distribution programs designed for an industrialized society.

The Unemployment Compensation Program

Feminist scholarship on the welfare system documents a strong connection between discrimination, domination, and control of the state (Gordon 1994). More specifically, feminist researchers argue that by consistently providing minimum income assistance to some groups (i.e., social assistance) or by excluding certain groups from higher paying income security programs (i.e., social insurance), the welfare system performs the following functions: (1) it reinforces sexual and racial economic subordination, (2) it maintains white male domination over women and people of color, and (3) it mediates power relations "between politically dominant and politically repressed groups" (Quadagno 1990:26). Thus, welfare programs, as an embodiment of state capacity and decision-making individuals and processes (Weir, Orloff, and Skocpol 1988), "have a great deal to do with maintaining social and economic inequities" (Piven and Cloward 1971:xvii), especially those that stem from sex, race, and place (Gordon 1990; Mink 1990; Nelson 1990; Pearce 1990; Quadagno 1990).

According to this scholarship, the Unemployment Insurance (UI) program was specifically designed during the depression of the 1930s for male heads of households with longstanding commitments to the labor force who became unemployed due to structural changes in the labor market (Pearce 1990). This federal/state partnership was designed based on the labor force patterns of regularly employed white male urban workers. Consequently, it is structurally biased against any worker whose labor force participation patterns differ from workers with longstanding attachments to the labor market (Bassi and McMurrer 1997; Bingham 1995; Blaustein 1993; Gordon 1994; Latimer 1999; Maranville 1992; Pearce 1986).

For example, not all jobs are covered by the unemployment insurance program. States have enormous discretion in setting the specific tax provisions and benefits for the UI system. In general, the status of the employer (i.e., paying or not paying UI taxes) determines the UI coverage of each employee (Bassi and McMurrer 1997). According to Bingham (1995) "approximately 12% of the work force works in employment not covered by unemployment compensation" (p. 944). Nationwide, jobs that are not covered by unemployment insurance are disproportionately filled by women, people of color, and rural workers. They include self-employed workers, domestic laborers who are paid less than $1,000 in one quarter, workers employed in religious organizations, and in some states, agricultural workers who work on small farms (Bassi and McMurrer 1997).

In West Virginia, employers have to cover domestic workers if they pay $1,000 or more in a calendar quarter. Religious, charitable, educational, or other nonprofit organizations are liable for UI taxes if they have four or more individuals for some portion of the day, for at least 20 weeks a year. Individuals working in a rehabilitative program are not covered for UI benefits. Not all agricultural jobs are covered by UI benefits in West Virginia. Employers who have agricultural employment of ten or more individuals for some portion (e.g., it can be for one hour or eight hours) of at least one day for 20 different weeks in a calendar year must cover each of their employees (West Virginia Bureau of Employment Programs 1997).

West Virginia, like thirteen other states, has "adopted special provisions intended to restrict the eligibility of seasonal workers" (Nicholson 1997:99). All employers who have (1) employed one or more individuals during some portion of a day for 20 weeks out of a calendar year or (2) pay $1500 in total wages in a calendar quarter, must cover their seasonal employees with UI benefits in West Virginia. For those who are self-employed (i.e., the sole proprietor) or a partner, they are not liable to cover themselves, their spouses, their children, or their parents with UI benefits. If the organization is a corporation, everyone must be covered (West Virginia Bureau of Employment Programs 1997).

Coverage does not automatically guarantee receipt of UI benefits. In order to qualify for unemployment insurance, workers must meet a minimum earnings and work time requirement. Individuals in West Virginia are monetarily denied UI benefits if (1) they do not make enough wages in the base period to qualify, (2) they make at least $2,200 in the base period but it was all in one quarter, or (3) they do not re-qualify for benefits. Part-time, seasonal, service, and agricultural workers have lower average earnings and more difficulty meeting the minimum income requirement for UI.

A monetary eligibility for UI benefits does not guarantee full payment of those benefits. Unemployed West Virginia workers who monetarily qualify for UI benefits can receive up to 26 weeks of UI payments if certain weeks do not get disqualified. An unemployed worker must have become unemployed "on good terms" to avoid weeks of disqualification due to a separation issue. In other words, the terms of their unemployment can cause them to lose weeks of benefits. Employers provide the reasons for unemployment and if disputed by the employee, an UI benefits officer confers with both to determine the terms of unemployment. Examples of separation issue disqualifications include voluntary quits or being discharged due to misconduct. Workers who voluntarily quit their jobs for "good cause" can collect UI in West Virginia if the employer is at fault for the quit. The worker must prove that the employer did something wrong such as sexually harassing the employee, changing their hiring agreement (i.e., drastically cutting their hours or their pay), or trying to make them relocate. Workers who voluntarily quit to follow their spouses or because they cannot find satisfactory child care to enable them to work, would be denied benefits in West Virginia.

Unemployed workers can also have weeks of their UI benefits disqualified because they failed to follow criteria outlined by the UI office to remain eligible for UI benefits (i.e., a non-separation issue). Frequent causes for a non-separation disqualification are that the claimant was not able and available for work, the claimant failed to meet the reporting requirements (i.e., did not report to the UI office every two weeks or provide a valid reason for not reporting), the claimant was only partially unemployed, the claimant was not registered for work with job services within six weeks, the claimant was receiving annuity, pension pay, and the claimant failed to accept suitable work (i.e., a job within 10% of their previous wages). Individuals who have pieced together two part-time jobs to survive economically would be denied benefits upon losing one job because they are seen as only partially unemployed. Transportation barriers are not seen as legitimate reasons for failing to meet the non-separation qualifications.

In summary, there are structural inequities built into the social insurance program in terms of coverage and qualifications for benefits. Individuals who are over-concentrated in part- time jobs, seasonal jobs, agricultural jobs, family owned businesses and/or low paying service sector jobs are unlikely to qualify for the better paying, less stigmatized unemployment insurance when they become unemployed. These disadvantaged workers must rely on the highly limited, stigmatized social assistance programs such as AFDC/TANF, Food Stamps, Medicaid, and public housing. Labor market disadvantage clearly translates into welfare vulnerability.

The Connection Between Unemployment Insurance, Sex, Race, and Place

The occupations that are not covered by UI benefits or that have the most difficulty meeting the eligibility requirements for UI benefits are disproportionately filled by women and people of colon A recent report by the Congressional Research Service found that individuals who are "young, did not head families, and were not the primary source of income within their families" and had "lower-than-average incomes both before and after their unemployment spell" (Bassi and McMurrer 1997:73) were the least likely of all unemployed individuals to be receiving unemployment insurance benefits.

Because women, whether they have partners or not, are held responsible for domestic labor and childcare (Folbre 1984;Hurst 2001; McLanahan 1985), they face the choice of working fewer hours and having less income or enduring greater stress (Tickamyer and Tickamyer 1988). Divorce further restricts women's labor force participation as women become solely responsible for their children's physical, economic, and emotional welfare (Pearce 1990). These responsibilities can be even more intense for African American women because overall black communities have greater poverty than white communities and consequently, fewer resources to offer these women (Hurst 2001; Pearce 1989).

Women also work fewer hours than men because there is a growing preference for part-time employment, particularly in female dominated occupations and industries. In the late eighties and early 1990s, twenty percent of all workers were employed part-time and over twenty-five percent of all female workers worked part-time. African American women are even more likely to be part-time workers than white women (Reskin and Padavic 2003;Pearce 1989).

Part-time workers have lower earnings (due to fewer working hours), higher levels of occupational segregation (Reskin and Padavic 2003), and more difficulty meeting the minimum income or the availability qualification requirement for UI. Pearce argues that some working mothers may actually choose to work part-time, but others are forced into part-time work because day care or after-school care costs outweigh the benefits of full-time employment. Thus, women's childrearing responsibilities, the lack of publicly subsidized child care, and "the increasing demand for relatively cheap services come together to reinforce the pattern of part-time, low wage employment for many women workers" (Pearce 1986:147).

Women's labor force participation patterns also differ from men by the sectors in which they continue to be highly concentrated. The sex segregated occupational structure relegates and isolates women, particularly women of color, primarily in the periphery, secondary, and service sectors of the labor market (Hurst 2001; Reskin and Padavic 2003; Stafford and Fossett 1989). Although women's labor force participation and placement have changed dramatically, a large proportion of women workers occupy traditionally "female" jobs (Pearce 1986). In fact Reskin and Padavic (2003) found that "of the 57 million women in the labor force in 1990, one third worked in just 10 of the 503 detailed occupations" (p. 422). They also document that 60 percent of black women and 53 percent of white women would have to shift occupations to jobs predominately occupied by white males in order for the labor market to be completely integrated by sex and race (Reskin and Padavic 2003:422).

Recent changes in the structure of the economy (such as the decrease in high wage industry jobs have increased competition for more limited employment opportunities, and therefore, led to an even greater confinement of women in lower paid sectors of the occupational structure. Technological innovations and deskilling have eliminated many of the traditional jobs held by women (i.e., clerical work) (Tickamyer and Tickamyer 1988).

Women's child care responsibilities, their lower average hours worked per week, their greater part-time employment, their segregation and isolation in the secondary sector of the labor market, as well as continued gender and racial discrimination (Reskin and Padavic 2003) result in average hourly earnings that are significantly lower than hourly wages earned by men. The average woman still earns about 76 percent of the total average income earned (per week) by males (Hurst 2001:68). Single mothers earn, on average, between 30% to 40% of the income earned by two-parent families headed by men. Even when single mothers with children (under six years old) work full-time at paid labor, more than one-third of these individuals are poor (Pearce and McAdoo 1981). According to Hurst (2001) "Female-headed families have continued to possess poverty rates that are over six times those of married-couples" (p. 27). Race compounds the effects of gender and household structure. African American women are even more disadvantaged with an income that is approximately 86 percent of the total income earned by white single women (Hurst 2001:106). These lower average incomes make it difficult for women to meet the minimum earnings qualifications for unemployment insurance benefits (Pearce 1986).

Rural areas have a disproportionate concentration of the occupations that are not covered by UI benefits and/or that have the most difficulty meeting the eligibility requirements for UI benefits. Rural labor markets lack diversity and are typically dominated by farm based economies, the service industries, and nondurable manufacturing sectors (i.e., periphery and secondary sector jobs) (Tickamyer and Duncan 1990).

Rural communities have been particularly hard hit by deindustrialization and economic restructuring. According to Zimmerman et al. (1999) "Between 1969 and 1992, rural manufacturing employment dipped from 20.4% to 16.9% of total employment" (p. 8). Almost one-quarter of rural workers were employed in the service sector in 1996. Most of the service sector growth in rural areas has been in the low wage sector of the service industry (Gorham, 1992). The higher paying jobs have education and training qualifications that few rural workers can meet (Tickamyer and Duncan 1990). Both the working poor and involuntary part time workers in rural areas had a significantly more difficult time finding adequate jobs in the 1990s than their counterparts in the 1970s (Zimmerman et al. 1999).

The low wage employment that dominates rural areas is compounded by the fact that neither the sex segregation of the labor market (Goudy et al. 1986) nor racial inequalities in income (Cho and Ogunwole 1989) have significantly decreased in rural areas. In addition, rural people have fewer child care and transportation resources and longer commutes to potential jobs than their urban counterparts (Zimmerman et al. 1999).

The statistics on West Virginia's labor force reflect these sex, race, and rural trends. West Virginia has one of the highest unemployment rates in the nation (higher for African Americans than whites), a higher than average rate of occupational and industrial sex segregation, an overconcentration of workers in the service industry (i.e., 50 percent of women and 20 percent of men found in the service industries), a larger percent of women working part-time and only part of the year than in the nation, the lowest college graduation rate in the nation and one of the highest income gaps between full-time male and female workers (i.e., women earn on average 58 percent of their male counterparts) (Hannah 1995: 17-29).

Research Design

Data

Data for this research come from the West Virginia Unemployment Compensation Program, a subunit of the West Virginia Bureau of Employment Programs. The data include all claims made for unemployment insurance in January, May, September, and December of 1997. The cost of data extraction restricted the study to four months of analysis. Data for the two months in 1997 with the most UI claims made and the two months in 1997 with the least claims made were extracted to examine the relationship between high/low demands on the system and UI receipt.

Approximately 37,000 unemployed workers in West Virginia applied for UI benefits during these four months. According to O'Leary (2000) only about 35 percent of all unemployed individuals receive UI benefits (p. 2). The total number of unemployed workers in West Virginia for the four months examined in this research was 220,900 people. Thus, only 16.8 percent of the unemployed in West Virginia actually applied for their benefits. Other research shows that the rural poor and unemployed are much less likely to apply for public assistance benefits than their urban counterparts (Rank and Hirschl 1993). The reasons given are that those economically disadvantaged in rural areas have less knowledge about their eligibility for such programs, and/or there is a greater stigma in using these benefits in these areas (Rank and Hirschl 1993). I would argue that these factors also explain lower applications for UI benefits in rural areas. I would also add that the over-concentration of minimum wage jobs in rural areas make the UI benefits so low that they are simply not worth the time and energy required to receive them.

The UI data set contains individual level information such as sex, race/ethnicity, age, education, county of residence, occupation the respondent asked the program to job search for, whether or not the applicant meets the monetary requirement for UI, which part of the monetary requirement did the individual fail, the weekly UI benefit amount for qualified workers, the non-monetary reasons for disqualification, and length of time disqualified due to a non-monetary violation. Unfortunately, the occupation variable could not be used because about fifty percent of the cases were missing.

The UI data were merged with West Virginia county level data from the census (i.e., Census of Housing and Population, County-City Data Book, County-Statistics File 4, and the Regional Economic Information System) so that measures of the local labor market could be included in the models. The West Virginia county level data include the following information on the individual's county of residence: is it a rural or urban county, what is the population density, does the county have a diverse industry structure, is it in a metropolitan area or near an interstate, and does the county have large concentrations of federal, manufacturing, mining, or farming employment.

Measures

Dependent Variables

The first dependent variable measures monetary eligibility for UI benefits. If the worker (1) failed to make $2,200 in the base period, (2) made at least $2,200 but it was not in the first four of the last five completed quarters, or (3) did not earn enough money to re-qualify for benefits, they have a MONETARY DISQUALIFICATION. About 3% of unemployed West Virginia workers in this study failed to meet the monetary requirements for UI benefits.

The second dependent variable represents whether or not the worker is disqualified due to a separation issue. About 5.4% of unemployed West Virginia workers receiving UI lose at least one week of their UI benefits due to a separation issue. If an unemployed worker loses weeks of eligibility for any of the four major separation issues (i.e., voluntarily quitting their previous job, voluntarily quitting due to retirement, being discharged from their previous job due to misconduct, and being discharged for gross or aggravated misconduct), they have a SEPARATION ISSUE DISQUALIFICATION.

The third dependent variable measures whether or not the worker is disqualified due to a nonseparation issue. About 3% of unemployed West Virginia workers receiving UI lose at least one week of their UI benefits due to a nonseparation issue. If an unemployed worker loses weeks of eligibility for any of the six major nonseparation disqualifications (i.e., the claimant was not able and available for work, the claimant failed to meet the reporting requirements, the claimant was only partially unemployed, the claimant was not registered for work, the claimant was receiving annuity, pension pay, and the claimant failed to accept suitable work), they have a NONSEPARATION ISSUE DISQUALIFICATION.

Independent Variables

A number of individual level and county level measures which are linked to economic vulnerability are used as independent variables. The sex of the respondent is indicated in the FEMALE variable. WHITE is the race/ethnicity of the respondent where black, Asian, Native American, or Latino individuals are coded as zero and white is coded as one. AGE is the actual age of the respondent. EDUCATION is the number of years of education completed by the respondent. Originally, there were about 40% of the responses missing on education. To avoid losing this variable, I created 13 age ranges and calculated the average education level based on these five year age categories (i.e., under 25 years old, between 25 and 29 years old, between 30 and 34 years old, between 35 and 39 years old, between 40 and 44 years old, between 75 and 79 years old, and greater than 79 years old). I then replaced the missing cases with the educational mean for the appropriate age category. BASE PERIOD WAGES is the amount of wages used to determine the claimant's weekly benefit amount.

Researchers have found that the probability of a welfare recipient exiting the welfare system is positively associated with the county's unemployment rate and the welfare caseload (Brazzell, Lefbert, and Opitz 1989; O'Neill et al 1984; Rank 1994). These results appear counterintuitive. Rank explains (1994:165): "The key to understanding this relationship is the effect that increased pressure has on finite resources". As unemployment and poverty rates increase, more and more individuals turn to public assistance and put "greater pressure on the existing system and its resources, both financial and otherwise (e.g., number of staff, physical facilities, available time, and so on)" (Rank 1994:165). It is possible that increased demands for UI services can cause a "tightening up" of eligibility requirements for these benefits. To investigate this possibility, the HIGH APPLICANTS variable was constructed. HIGH APPLICANTS represents the month in which the originally UI claim was made. If a claim was made in the two months with the highest number of UI applicants (i.e., January or December) the response is coded as 1. A zero represents claims made in either of the two months with the lowest number of UI applicants (i.e., May or September).

The local labor market measures come from several county-level data sources: Census of Housing and Population, County-City Data Book, County-Statistics File 4, and the Regional Economic Information System. The measures of the ecological structure are METROPOLITAN AREA or URBAN AREA. Both of these similar variables measure the competitiveness of the local labor market area. Workers who live in an urban area or a metropolitan area should have more opportunities than workers outside these areas. METROPOLITAN AREA is whether or not the claimant's county falls into a metropolitan area. The URBAN AREA variable indicates that the county in which the respondent lives is urban or rural. Using a category scheme devised by the West Virginia Bureau of Employment Programs, urban counties are defined as those counties located in a metropolitan statistical area or any counties with cities of 10,000 individuals or more. This classification designates 17 out of 55 counties in West Virginia as urban (i.e., Kanawha, Putnam, Cabell, Wayne, Wood, Ohio, Marshall, Brooke, Hancock, Mineral, Berkeley, Jefferson, Monongalia, Marion, Harrison, Raleigh, and Mercer). The remaining 38 counties are designated as rural.

In 1995, 61 percent of the state's total population resided in the urban counties (WV Bureau of Employment Programs 1996). This rural/urban measure is also an indirect measure of the unemployment rate. The unemployment rate for the rural counties in 1995 was double the national average at 10.8 percent. In contrast, all of the urban counties experienced single-digit unemployment rates and captured over 75% of the new jobs developed between 1987 and 1995. Correspondingly, the per capita income in urban counties ($18,851) is well above the average per capita income earned in rural counties ($14,393) (West Virginia Bureau of Employment Programs 1996).

Two indicators of the local business climate are the log of the total 1989 earnings in a county from the mining sector (LOG MINING) and the agriculture/forestry sector (LOG FARMING). There is also a SUSTENANCE DIVERSITY variable. SUSTENANCE DIVERSITY is a measure of the dispersion of private non-farm employees in the county across seven industry sectors. The industry sectors include manufacturing, transportation, services, finance, insurance, and real estate, construction, wholesale trade and retail trade. High positive values on this variable indicate that the local economy is complex and has a diverse industry structure (Mencken 1997). This variable was standardized by converting it into z-scores.

Data Analysis

The statistical technique used to analyze the data is logistic regression. Given that one goal of this research is to investigate variations in the impacts of labor market measures across the state, the labor market measures are entered first for each model and are then followed by the individual variables in the second stage.

Results

About 76 percent of all unemployment insurance applicants in West Virginia are male. Almost 3 percent of the applicants are people of colon The UI applicants are on average 38 years old, have completed 12 years of education, and received $18,425 in wages during the base period. The majority of the applicants live in a rural county (52.9 percent) and applied for benefits during either January or December of 1997 (71.9 percent). About 97 percent of all applicants meet the monetary requirement for unemployment insurance benefits, 5.5 percent have weeks of benefits disqualified due to a separation issue, and 3.0 percent lose benefits due to a nonseparation issue.

The results from the bivariate crosstabulations and t-tests are found in Table 1. There are significant sex and race differences in monetary eligibility, separation issue disqualifications, nonseparation issue disqualifications, base period wages, and weekly UI benefit amounts. A significantly larger percent of women and people of color are monetarily ineligible for UI (i.e., not having enough wages in their base period or did not re-qualify for benefits), have weeks of UI receipt disqualified due to a separation issue, and lose weeks of UI receipt due to a non-separation issue. In addition, women and people of color have significantly lower average base period wages and weekly UI benefit amounts.

Table 2 contains the results from the MONETARY DISQUALIFICATION logistic regression. METROPOLITAN AREA is a significant predictor of meeting the monetary qualifications for UI benefits in Step 1. Unemployed workers who live in a county that falls within a metropolitan area are 1.23 times more likely to meet the monetary requirements for unemployment insurance than unemployed workers who live in a county outside of a metropolitan area. METROPOLITAN AREA remains significant and LOG FARMING becomes significant when the individual level variables are added in Step 2. Unemployed workers who live in a county with higher earnings in the agriculture/forestry sector are significantly less likely to monetarily qualify for UI benefits than unemployed workers in counties with low agriculture/forestry earnings. In fact, a one percent increase in the total county earnings in the agriculture/forestry sector results in a 6 percent decrease in the odds that an unemployed worker will meet the monetary qualification for UI benefits.

Sex, race, and age are also significant predictors of meeting the monetary requirements for UI benefits. As expected, unemployed men, whites, and older workers are more likely than unemployed women, people of color, and younger workers to monetarily qualify for UI benefits. Unemployed female workers are 1.54 times less likely than their male counterparts and unemployed people of color are 1.54 times less likely than their white counterparts to meet the monetary requirements for unemployment insurance benefits. Further analyses revealed that younger workers and females are more likely to fail the monetary requirement because they did not make enough money in the base period while older workers and males are more likely to fail the monetary requirement because all of their income came within one quarter.

A one year increase in age results in a 3% increase in the odds that an unemployed worker will monetarily qualify for UI benefits. It is interesting to note that all these significant relationships disappear when the model controls for the worker's base period wages. This finding implies that it is the labor market income inequities experienced by workers who are female, people of color, and younger that puts them at such a disadvantage in terms of qualifying for UI benefits. Overall, these results correspond with Bassi and McMurrer's (1997) finding that "The majority of the unemployed who do not meet their state monetary eligibility requirements are either new entrants to the labor force, reentrants to the labor force, or individuals with sporadic labor force participation" (p.76).

Table 3 contains the results from the SEPARATION ISSUE DISQUALIFICATION logistic regression. Both URBAN AREA and LOG FARMING are significant predictors of being disqualified due to a separation issue in Step 1 of the analysis. Unemployed workers who live in an urban county are 1.46 times more likely than unemployed rural workers to lose weeks of UI benefits due to a separation issue (i.e., they are more likely to have voluntarily quit their previous job or to be discharged from their previous job due to misconduct). Thus, unemployed rural workers are either less likely to voluntarily quit their jobs or they are less likely to have been discharged due to misconduct than their urban counterparts. It is possible that workers in urban counties have more potential job opportunities and thus are less tolerant of unsatisfactory working circumstances.

In terms of farming concentration, unemployed workers in counties with low farming concentration are more likely to have weeks disqualified due to a separation issue than their counterparts in high farming concentration counties. A one percent increase in the total county earnings in the agriculture/forestry sector results in a 7 percent decrease in the odds that an unemployed worker will have UI benefits disqualified due to a separation issue. URBAN AREA and LOG FARMING remain significant even when the individual level variables are added in Step 2.

All of the individual level variables except education are significant predictors of a separation issue disqualification. As in the monetary qualification model unemployed women, people of color, and younger workers are more likely than unemployed men, whites, and older workers to have weeks of their UI benefits disqualified due to a separation issue (i.e., they are more likely to have voluntarily quit or to be discharged due to misconduct). Unemployed women are 1.87 times more likely than unemployed men to be disqualified due to a separation issue. Unemployed people of color are 2.11 times more likely than their white counterparts to be disqualified from UI benefits due to a separation issue. A one year increase in age results in a 3 percent decrease in the odds that an unemployed worker will experience a separation issue disqualification of their UI benefits. In addition, unemployed low income workers are more likely to experience a separation issue disqualification than unemployed high income workers. A one dollar increase in base period wages results in a 1% decrease that an unemployed worker will experience a separation issue disqualification of their UI benefits. Further analyses revealed that younger workers, males, and people of color are more likely to be disqualified because they were discharged due to misconduct while whites, older workers, and females are more likely to be disqualified because they voluntarily quit their last job.

One surprising finding was the significant relationship between month of UI application and a separation issue disqualification. Unemployed workers who applied for UI benefits in the months with the least applicants (i.e., May or September) are over two times more likely than January/December (i.e., the months with the most applicants) applicants to have their UI benefits reduced due to a separation issue. It is possible that May and September have the least applicants because the unemployment level is lower during these months. When the unemployment level is lower, there are more job opportunities for workers. In times of high employment West Virginia workers maybe less tolerant of unsatisfactory working circumstances and thus are more likely to quit these jobs or voice their dissatisfaction and be discharged because of misconduct. It is also possible that UI officers have more time to scrutinize each application when the application rate is lower.

Table 4 contains the results from the NONSEPARATION ISSUE DISQUALIFICATION logistic regression. LOG FARMING is a significant predictor of being disqualified due to a nonseparation issue in Step I of the analysis. Unemployed workers in counties with low farming concentration are more likely to have weeks disqualified due to a nonseparation issue than their counterparts in high farming concentration counties. A one percent increase in the total county earnings in the agriculture/forestry sector results in a 14 percent decrease in the odds that an unemployed worker will have UI benefits disqualified due to a nonseparation issue. LOG FARMING remains significant even when the individual level variables are added in Step 2.

All of the individual level variables are significant predictors of a nonseparation issue disqualification. As in the previous model, unemployed women, people of color, younger workers, and low income workers are more likely than unemployed men, whites, older workers, and high income workers to have weeks of their UI benefits disqualified due to a nonseparation issue (i.e., because they failed to follow criteria outlined by the UI office to remain eligible for UI benefits). Unemployed women are 1.83 times more likely than unemployed men to be disqualified due to a nonseparation issue. Unemployed people of color are 1.70 times more likely than their white counterparts to be disqualified from UI benefits due to a nonseparation issue.

A one year increase in age results in a 1% decrease in the odds that an unemployed worker will experience a nonseparation issue disqualification of their UI benefits. A one dollar increase in base period wages results in a 1% decrease that an unemployed worker will experience a nonseparation issue disqualification of their UI benefits. Further analyses revealed that older workers and females were more likely to have a nonseparation disqualification because they were not able and ready to work. Younger, male workers were more likely to have a nonseparation disqualification because they did not meet the reporting requirements or they did not register to work.

As in the separation issue model, unemployed workers who applied for UI benefits in the months with the least applicants (i.e., May or September) are 1.66 times more likely than those who applied in the months with the most applicants to have their UI benefits reduced due to a nonseparation issue. In addition, unemployed high education workers are more likely to experience a nonseparation issue disqualification than unemployed workers with low education. A one year increase in educational attainment results in a 7 percent increase in the odds that an unemployed worker will experience a nonseparation issue disqualification of their UI benefits.

Discussion/Conclusion

This research adds to existing poverty literature by documenting sex, race, place, age, and income disparities in unemployment insurance disqualifications in a poor rural state. Several findings from this research support feminist assertions that the UI program is structurally biased against certain workers. For example, unemployed West Virginia workers who live in a county outside of a metropolitan area are significantly less likely to meet the monetary qualification for UI insurance than workers within a metropolitan area. This finding is not surprising given that the types of employment that dominate rural and nonmetro areas (i.e., service industries, nondurable manufacturing sectors, agriculture, and resource extraction) are low waged and/or low security employment.

Thus, it appears that many rural workers are engaged in various forms of unpaid, seasonal, or part-time labor which have contributed to the survival of their households but are not covered under the UI program or do not qualify them for benefits. Their lower average earnings make it more difficult for them to meet the minimum earnings requirement (over a two quarter period) for unemployment insurance. The good news for rural workers in West Virginia is that once they monetarily qualify for UI benefits, they are significantly less likely than urban workers to lose benefit payments due to a separation issue or they are no more likely than urban workers to be disqualified due to a nonseparation issue. A similar pattern is found for unemployed workers in counties with high concentrations of farming (i.e., significantly less likely to monetarily qualify, or lose weeks of UI benefits due to a separation issue or nonseparation issue). It is possible that rural workers voluntarily quit, get discharged due to misconduct, or refuse to follow the eligibility criteria (i.e., they are "better behaved") for UI less often than urban workers because of their more limited employment opportunities.

In addition, unemployed women, people of color, younger workers, and low income workers in West Virginia in 1997 consistently (1) have the most difficult time meeting the monetary requirement for UI benefits, (2) are the most likely to be disqualified from UI benefits because of a separation issue, and (3) are the most likely to have a nonseparation issue disqualification of UI benefits. These findings are particularly troublesome in a postindustrial society where global competition, continued institutionalized occupational and income inequality, and a restructuring of the public assistance program compound the historic problems with distribution programs designed for an industrialized urban society (Schram 1995).

Economic restructuring and welfare reform have made the historical inequities built into the unemployment insurance program more problematic than ever before. The UI system has remained virtually unchanged since its inception in the 1930s (O'Leary 2000). We now live in a postindustrial society where workers attain jobs "that increasingly must be subsidized with supplemental wages, tax credits, or the extension of public assistance benefits if they are to serve as a means of subsistence" (Schram 1995:173). Historically, part-time labor was associated with a weak attachment to the labor market. Today, "the rapid growth in flexible working arrangements has made such an assumption increasingly untenable" (Nicholson 1997:115). The current unemployment insurance program has not adjusted to the transformations in the nation's economy and work force.

The sex, race, age, and income disparities in unemployment insurance disqualifications also indicate big problems for the large number of current and former welfare recipients pushed into the labor market by the welfare reform law of 1996 (PWRORA, the Personal Responsibility and Work Reconciliation Act of 1996-PL 104-193). Unemployed workers who are denied UI benefits have historically turned to the stigmatized, restrictive social assistance benefits (i.e., AFDC, Food Stamps, and Medicaid) to supplement their insufficient incomes. These programs were "downsized" and restructured in 1996 making this already overtaxed inefficient system an even more limited back-up system to the UI program.

Two specific changes to this system, a lifetime cap on benefits and stiff work requirements for recipients, have the potential to increase the labor force participation of both former and current welfare recipients and thus create stronger connection between the social assistance and social insurance programs (i.e., TANF and UI) (Lare 1999). Unfortunately, researchers have found that only about 13 percent of unemployed people with a history of welfare receipt qualified for UI benefits (Gornick 1999:53). Former welfare recipients have trouble meeting the monetary requirement for UI benefits because many simply do not earn enough money to qualify and "... conditioning UI eligibility on earnings-rather than hours, days, or weeks of employment-during the base period discriminates against lower-wage workers" (Gornick 1999:53). Hobbie, Wittenburg, and Fishman (1999) state that "Former TANF UI beneficiaries will have special needs, such as child care and transportation assistance, that historically have not been addressed by the UI system" (p. 16) but could create more voluntary quits. Another barrier is that "Many recent welfare leavers, especially those with young children, are seeking part-time work, which disqualifies them from UI in most circumstances" (Gornick 1999:49). Hobbie et al. (1999) also document that the "Able to work" qualification for UI benefits can be particularly difficult for adults with disabilities or who are living with a family member with a physical or mental disability. The final barrier to UI benefits is that given their historical exclusion from this program "Many TANF recipients may not understand how the UI system works or how to apply for UI benefits" (Hobbie et al. 1999:16).

The structural limitations of the UI program combined with economic structuring and the rapidly shrinking public assistance safety net place low income individuals in a rural state like West Virginia in a uniquely postmodern economic crunch. Until the unemployment insurance program is transformed to accommodate the part-time, contingent, and self-employed workers who increasingly dominate our postindustrial economy, work force, and welfare system, the unemployment insurance system will continue to re-enforce and legitimate sex, race, and place inequities in employment and income security.

Table 1
Results from Bivariate Crosstabulations and T-Tests by Sex and Race

                                             MALES          FEMALES

MONETARY INELIGIBILITY                          2.91%         4.17% ***
Not Enough Wages in Base Period                 2.24%         3.12% ***
Wages All in One Quarter                        0.29%         0.16% *
Did Not Re-qualify                              0.38%         0.89% ***
SEPARATION ISSUE DISQUALIFICATIONS              4.27%         9.19% ***
NON-SEPARATION ISSUE DISQUALIFICATIONS          2.36%         4.98% ***
BASE PERIOD WAGES                         $20,488.55    $12,224.50 ***
                                         ($13,405.49)   ($8,664.090)
WEEKLY UI BENEFIT AMOUNT                     $188.38       $124.00 ***
                                             ($91.66)      ($74.16)

                                                               PEOPLE
                                               WHITES         OF COLOR

MONETARY INELIGIBILITY                          3.10%         4.57% **
Not Enough Wages in Base Period                 2.35%         3.12%
Wages All in One Quarter                        0.25%         0.42%
Did Not Re-qualify                              0.50%         1.04% *
SEPARATION ISSUE DISQUALIFICATIONS              5.33%        13.30% ***
NON-SEPARATION ISSUE DISQUALIFICATIONS          2.92%         5.84% ***
BASE PERIOD WAGES                         $18,644.77    $14,928.26 ***
                                         ($12,937.20)  ($12,198.49)
WEEKLY UI BENEFIT AMOUNT                     $174.08       $143.99 ***
                                             ($91.88)      ($90.38)

*** Indicates significant sex or race differences at p = .001

** Indicates significant sex or race differences at p = .01

* Indicates significant sex or race differences at p = .05

Table 2
Logistic Regression of MONETARY DISQUALIFICATION on Labor
Market and Individual Variables (N = 37011)

                                     Step 1

                       b (1)        S.E. (2)   O/R (3)

Metropolitan          -.2047 **      .0759       .815
  Area               (-.0522)                  (1.23)
Log Farming            .0458         .0242      --
                      (.0330)
Log Mining             .0043         .0179      --
                      (.0045)
Female
White
Age
Education
High Applicants
Intercept             3.13
[chi square]         16.03 ***
p[R.sup.2]

                                     Step 2

                         b            S.E.       O/R

Metropolitan        -.1879 **        .0767       .829
  Area             (-.0479)                    (1.21)
Log Farming          .0560 *         .0243      1.06
                    (.0405)
Log Mining           .0120           .0181      --
                    (.0124)
Female               .4315 ***       .0656      1.54
                    (.1011)
White               -.4342 **        .1604       .648
                   (-.0374)                    (1.54)
Age                 -.0335 ***       .0029       .967
                   (-.2068)                    (1.03)
Education           -.0314           .0201      --
                   (-.0258)
High Applicants      .0529           .0677      --
                    (.0131)
Intercept           1.20
[chi square]      206.80 ***
p[R.sup.2]           .214

(1) The unstandardized coefficients are presented first
with the standardized coefficients listed below in parentheses.

(2) These are the standard errors.

(3) The numbers in this column are the odds/ratios.

Table 3
Logistic Regression of SEPARATION ISSUE DISQUALIFICATION
on Labor Market and Individual Variables (N = 37011)

                                     Step 1

                       b (4)        S.E. (5)   O/R (6)

Urban                .3756 ***       .0513      1.46
  Area              (.1034)
Sustenance           .0117           .0068      --
  Diversity         (.0307)
Log Farming         -.705 ***        .0179       .932
                   (-.0509)                    (1.07)
Female
White
Age
Education
High Applicants
Based Period
  Wages
Intercept          -3.43
[chi square]      121.94 ***
p[R.sup.2]

                                     Step 2

                         S            S.E.       O/R

Urban                 .2784 ***      .0522      1.32
  Area               (.0766)
Sustenance            .0093          .0067      --
  Diversity          (.0244)
Log Farming          -.0693 ***      .0181       .933
                    (-.0501)                   (1.07)
Female                .6238 ***      .0505      1.87
                                                (.1462)
White                -.7459 ***      .1038       .474
                    (-.0642)                   (2.11)
Age                  -.0313 ***      .0024       .969
                    (-.1932)                   (1.03)
Education            -.0092          .0157      --
                    (-.0076)                     .439
High Applicants      -.8226 ***      .0468     (2.28)
                    (-.2038)                    1.00
Based Period         -.0000 ***      .0000
  Wages             (-.1230)
Intercept            -.795
[chi square]      1105.64 ***
p[R.sup.2]            .285

(4) The unstandardized coefficients are presented first
with the standardized coefficients listed below in parentheses.

(5) These are the standard errors.

(6) The numbers in this column are the odds/ratios.

Table 4
Logistic Regression of NONSEPARATION ISSUE DISQUALIFICATION
on Labor Market and Individual Variables (N = 37011)

                                     Step 1

                       b (7)        S.E. (8)   O/R (9)

Urban               .0356            .0668      --
  Area             (.0098)
Sustenance          .0080            .0085      --
  Diversity        (.0211)
Log Farming        -.1326 ***        .0215       .876
                  (-.0958)                     (1.14)
Female

White

Age

Education

High Applicants

Base Period
  Wages
Intercept         -4.22
[chi square]      43.79 **
p[R.sup.2]
                                     Step 2

                         b            S.E.       O/R

Urban               -.0426           .0676      --
  Area             (-.0117)
Sustenance           .0058           .0084      --
  Diversity         (.0152)
Log Farming         -.1317 ***       .0216       .877
                   (-.0951)                    (1.14)
Female               .6039 ***       .0672      1.83
                    (.1415)
White               -.5289 ***       .1462       .589
                   (-.0455)                    (1.70)
Age                 -.0076 *         .0030       .992
                   (-.0466)                    (1.01)
Education            .0699 ***       .0210      1.07
                    (.0575)
High Applicants     -.5099 ****      .0632       .601
                   (-.1263)                    (1.66)
Base Period         -.0000 ***       .0000      1.00
  Wages            (-.0858)
Intercept          -3.86
[chi square]      303.03 ***
p[R.sup.2]           .207

(7) The unstandardized coefficients are presented first
with the standardized coefficients listed below in parentheses.

(8) These are the standard errors.

(9) The numbers in this column are the odds/ratios.

References

Amott, Teresa. 1990. "Black Women and AFDC: Making Entitlement Out of Necessity". Pp.280-298 in Women, the State, and Welfare edited by Linda Gordon. Madison, Wisconsin: The University of Wisconsin Press.

Bassi, Laurie J. and Daniel P. McMurrer. 1997. "Coverage and Recipiency". Pp. 51-87 in Unemployment insurance in the United States: Analysis of Policy Issues edited by Christopher J O'Leary and Stephen A. Wander. Michigan: W.E. Upjohn Institute for Employment and Research.

Bingham, Stephen. 1995. "Replace Welfare for Contingent Workers With Unemployment Compensation". Fordham Urban Law Journal 22:937-59.

Blaustein, Saul J. 1993. Unemployment Insurance in the United States: The First Half Century. Michigan: W. E. Upjohn Institute for Employment and Research.

Brazzell, Jan F., Irving Lefbert, and Wolfgang Opitz. 1989. "The Impact of Population Size and the Economy on Welfare Caseloads: The Special Case of Welfare Reform". Applied Demography 4:1-7.

Cho, Woong and S. Ogunwole. 1989. "Black Workers in Southern Rural Labor Markets". Pp. 189-206 in Research in Rural Sociology and Development edited by William Falk and Thomas Lyson. London: JAI Press.

Folbre, Nancy. 1984. "The Pauperization of Motherhood: Patriarchy and Public Policy in the United States". Review of Radical Political Economics 16(4):72-88.

Fletcher, Cynthia, Jan Flora, Barbara Gaddis, Mary Winter, and Jacquelyn Litt. 2000. "Small Towns and Welfare Reform". Poverty Research News 4(5):8-9.

Gennetian, Lisa, Cindy Redcross, and Cynthia Miller. 2000. "Rural-Urban Differences in the Minnesota Family Investment Program (MFIP)". Poverty Research News 4(5):17-19.

Goetz, Stephan J. and David Freshwater. 1997. "Effects of Welfare Reform on Rural Counties and Labor Markets". American Journal of Agricultural Economics 79(5):1608-1613.

Gordon, Linda. 1994. Pitied But Not Entitled: Single Mothers and The History of Welfare. Cambridge, Massachusetts: Harvard University Press.

Gorham, Lucy. 1992. "The Growing Problem of Low Earnings in Rural Areas". Pp. 21-39 in Cynthia M. Duncan (ed.) Rural Poverty in America. New York, NY: Auburn House.

Gornick, Janet C. 1999. "Income Maintenance and Employment Supports For Former Welfare Recipients: The United States in Cross-National Perspective. Pp. 49-90 in Rethinking Income Support for the Working Poor: Perspectives on Unemployment Insurance, Welfare, and Work edited by Evelyn Ganzglass and Karen Glass. Washington, D.C.: The National Governor's Association.

Goudy, Willis, Rogelio Saenz, Sandra Burke, and David Rogers. 1986. "Occupational Segregation in Selected Iowa Communities: 1960-1980". Paper presented at the Annual Meeting of the Rural Sociological Society, Salt Lake City, Utah.

Hannah, Karen L. 1993. West Virginia Women: In Perspective 1980-1995. Charleston, WV: West Virginia Women's Commission.

Harvey, Mark and Gene Summers. 2000. "The Impact of Welfare Reform in Persistently Poor Rural Regions". Poverty Research News 4(5):10-12.

Hobbie, Richard, David C. Wittenburg, and Michael E. Fishman. 1999. "Temporary Assistance for Low-Wage Workers: Evolving Relationships Among Work, Welfare, and Unemployment Insurance" Pp. 15-48 in Rethinking Income Support for the Working Poor: Perspectives on Unemployment Insurance, Welfare, and Work edited by Evelyn Ganzglass and Karen Glass. Washington, D.C.: The National Governor's Association.

Hurst, Charles E. 2001. Social Inequality: Forms, Causes, and Consequences. Boston, MA: Allyn and Bacon.

Latimer, Melissa. 1999. "A Longitudinal Analysis of Unemployment Insurance Receipt: 1987-1991". Sociological Focus 32(4):383-400.

Lare, Barry Van. 1999. "Issue and Options for States in Creating a Work-Centered Safety Net for Working Poor" Pp.1-14 in Rethinking Income Support for the Working Poor: Perspectives on Unemployment Insurance, Welfare, and Work edited by Evelyn Ganzglass and Karen Glass. Washington, D.C.: The National Governor's Association.

McLanahan, Sara. 1985. "Family Structure and the Reproduction of Poverty". American Journal of Sociology 90:873-901.

Maranville, Deborah. 1992. "Feminist Theory and Legal Practice: A Case Study on Unemployment Compensation Benefits and the Male Norm". The Hastings Law Journal 43:1081-94.

Mencken, F. Carson. 1997. "Regional Differences in Socioeconomic Well-Being in Appalachia". Sociological Focus 30(1):79-97.

Mink, Gwendolyn. 1990. "The Lady and the Tramp: Gender, Race, and the Origins of the American Welfare State". Pp. 92-122 in Women, the State, and Welfare edited by Linda Gordon. Madison, Wisconsin: The University of Wisconsin Press.

Nelson, Barbara. 1990. "The Origins of the Two-Channel Welfare State: Workmen's Compensation and Mother's Aid". Pp. 123-151 in Women, the State, and Welfare edited by Linda Gordon. Madison, Wisconsin: The University of Wisconsin Press.

Nicholson, Walter. 1997. "Initial Eligibility for Unemployment Compensation". Pp. 91-124 in Unemployment Insurance in the United States: Analysis of Policy Issues edited by Christopher J O'Leary and Stephen A. Wander. Michigan: W.E. Upjohn Institute for Employment and Research.

O'Leary, Christopher J. 2000. "U.S. Unemployment Insurance: Progress and Prospects". Employment Research 7(2):1-4. W.E. Upjohn Institute for Employment Research.

O, Neill, June A., Douglas A. Wolf, Laurie J. Bassi, and Michael T. Hannan. 1984 (June). An Analysis of Time on Welfare. Final report to the U.S. Department of Health and Human Services. Washington, D.C.: The Urban Institute.

Pearce, Diana. 1990. "Welfare is Not for Women: Why the War on Poverty Cannot Conquer the Feminization of Poverty". Pp.265-279 in Women, the State, and Welfare edited by Linda Gordon. Madison, Wisconsin: The University of Wisconsin Press.

Pearce, Diana. 1989. "Farewell to Alms: Women's Fare Under Welfare". Pp. 493506 in Women (Fourth Edition) edited by Jo Freeman. California: Mayfield Publishing Company.

Pearce, Diana. 1986. "Toil and Trouble: Women Workers and Unemployment Compensation. Pp. 141-161 in Women and Poverty edited by Barbara Gelpi, Nancy Hartsock, Clare Novak, and Myra Strober. Chicago: The University of Chicago Press.

Pearce, Diana and Harriet McAdoo. 1981. Women and Children: Alone and in Poverty. Washington, D. C.: National Advisory Council on Economic Opportunity.

Piven, Francis Fox and Richard A. Cloward. 1971. Regulating the Poor: The Functions of Public Welfare. New York, NY: Pantheon.

Quadagno, Jill. 1990. "Race, Class, and Gender in the U.S. Welfare State: Nixon's Failed Family Assistance Plan". American Sociological Review 55:11-28.

Rank, Mark Robert. 1994. Living on the Edge: The Realities of Welfare In America. New York: Columbia University Press.

Rank, Mark R. and Thomas A. Hirschl. 1993. "The Link Between Population Density and Welfare Populations". Demography 30(4):607-623.

Reskin, Barbara E and Irene Padavic. 2003. "Sex, Race, and Ethnic Inequality in United States Workplaces". Pp. 420-436 in The Meaning of Difference edited by Karen E. Rosenblum and Toni-Michelle C. Travis. New York: McGraw Hill.

Rural Sociological Society Task Force. 1993. Persistent Poverty in Rural America. Boulder, CO: Westview Press.

Schram, Sanford F. 1995. Words of Welfare: The Poverty of Social Science and the Social Science of Poverty. Minneapolis, MN: University of Minnesota Press.

Stafford, M. Therese and Mark A. Fossett. 1989. "Occupational Sex Inequality in the Nonmetropolitan South, 1960-1980". Rural Sociology 54:169-194.

Tickamyer, Ann R., Janet Bokemeier, Shelley Feldman, Rosalind Harris, John Paul Jones, and Dee Ann Wenk. 1993. "Women and Persistent Rural Poverty". Pp. 200-229 in Persistent Poverty in Rural America. Boulder, Colorado: Westview Press.

Tickamyer, Ann R. and Cynthia M. Duncan. 1990. "Poverty and Opportunity Structure in Rural America". Annual Review of Sociology 16(1):67-86.

Tickamyer, Ann R. And Cecil Tickamyer. 1988. "Gender and Poverty in Central Appalachia". Social Science Quarterly 69(4):874-891.

Weir, Margaret, Ann Shola Orloff, and Theda Skocpol. 1988. "Introduction: Understanding American Social Politics". Pp. 3-27 in The Politics of Social Policy in the United States edited by Margaret Weir, Ann Shola Orloff, and Theda Skocpol. New Jersey: Princeton University Press.

West Virginia Bureau of Employment Programs. October 1997. "Handbook for Employers". www.state.wv.us/bep/uc/HDBKEMP.HTM.

West Virginia Bureau of Employment Programs. August 1996. "The Two West Virginias". Charleston, WV: West Virginia Bureau of Employment Programs.

Wilson, Julius Wilson. 1997. When Work Disappears: The World of the New Urban Pooh New York, NY: Alfred A. Knopf.

Zimmerman, Julie, Sarah Dewees, Lynn Reinschmiedt, and Tom Hirschl. 1999. Rural America and Welfare Reform: An Overview Assessment. Rural Policy Research Institute. P99-3, University of Missouri, Columbia, MO.

MELISSA LATIMER

West Virginia University

Dept. Of Sociology and Anthropology