Article body

Introduction

The past several decades have reflected a proliferation in non-standard jobs, including part-time and over-time work, double-shifts, and temporary positions in all sectors of the labour market (Presser, 2003; Vosko, 2006). These jobs are often defined as precarious in nature as they provide low wages, few benefits, and modest security (Vosko, 2000), and may have health consequences for workers (Kim et al., 2008). Exceptions include some voluntary shift work, or alternate or reduced work schedules that provide workers the opportunity to juggle competing work and family demands (Thompson, Beauvais, and Lyness, 1999).

Traditionally, research in this area focussed on “non-standard” employment, defined as anything deviating from the standard employment relations (SER) of the early 1900’s (Economic Council of Canada, 1990; Kalleberg, 2000; Krahn, 1991). Yet recent research has argued against such crude characterizations and instead emphasizes levels of precarity across non-standard employment (Cranford, Vosko, and Zukewich, 2003; Kalleberg, Reskin, and Hudson, 2000; Vosko, 2006). To address variations in precarious employment, research in the 1990’s shifted its focus from specific work settings, such as part-time, contract, or temporary work, towards the analysis of personal characteristics and job quality, which acknowledged the breadth of precarious work situations across both full-time and part-time occupational settings (Kalleberg et al., 2000; Rodgers and Rodgers, 1989; Webber and Williams, 2008). The following paper examines a combination of part-time, or non-standard work positions and overall job quality to better understand the determinants and inequalities across precarious employment.

The most common definition of precarious employment can be found in the work of Rodgers and Rodgers (1989). These authors characterize precarious employment across several dimensions, including working wages, work intensity, the presence of a union, or statutory protection(s), job stability, and access to personal or family benefits. The authors argue that the presence of such characteristics indicates high job quality. This definition deviates from previous literature on non-standard work or temporary positions (see Krahn, 1991), since it emphasizes the “quality” of work situations. More recent work has considered precarious employment as a continuum ranging from high precariousness to low precariousness, depending on wage work and regulatory protection across statuses, including full-time and part-time permanent and temporary positions (see Cranford et al., 2003). Accordingly, less lucrative work situations without stability and benefits are considered highly precarious for employees.

Not surprisingly, women are overrepresented in precarious work situations both nationally and internationally (Cranford et al., 2003; Presser, Parashar, and Gornick, 2008), and recent research suggests these trends are increasing (Fudge and Vosko, 2001; Ruyter and Warnecke, 2008; Vosko, 2006). In the U.S., thirty-three percent of women work in part-time positions compared to 12 percent of all men (U.S. Bureau of Labor Statistics, 2009). And within these positions, women earn approximately, 20 percent less than their male counterpart, both in Canada (Statistics Canada, 2007) and the U.S. (U.S. Census Bureau, 2007). These findings are further supported by research on the gender-wage gap, highlighting income discrepancies, despite similarities in job titles and tasks (Blau and Kahn, 2000). Women are also less likely to hold permanent positions and within these positions, work far fewer hours compared to their male counterparts (Cranford et al., 2003; Presser, 2003; Ruyter and Warnecke, 2008; Webber and Williams, 2008). Finally, research highlights that women are less likely to have union protection compared to their male counterparts, which may be directly related to employment in smaller sized firms (Cranford et al., 2003; Kalleberg et al., 2000).

These gendered differences in wages, security, work hours, and union protection suggest that women experience a greater degree of precariousness in various types of employment. Phrases such as the gender of precarious employment and the feminization of temporary employment norms underscore the problematic patterns of women’s presence in precarious work settings (Cranford et al., 2003; Fuller and Vosko, 2007; Vosko, 2006). Despite advances in research documenting women’s increased representation in these employment situations, there have been few studies exploring why women are more likely to occupy precarious job positions. The present study examines this enquiry directly, thereby contributing to the growing body of literature on gender and precarious employment. More specifically, the current study analyzes the gendered nature of precarious employment and the human capital investments and family situations that affect the likelihood of working in such settings. So while previous research documents women’s segregation in highly precarious jobs, findings from the present study provide a better understanding of why women are persistently overrepresented in such jobs, characterized by fewer rewards, protection, and work hours.

This paper focuses on two competing theories for why women are over-represented in precarious work settings: 1) rational choice theory or more popularly referred to as human capital theory, and 2) gender stratification theory. While human capital theorists emphasize individual choices and human capital investments in obtaining secure and lucrative job positions (Becker, 1994), stratification theories accentuate employers’ demands for, and personal discrimination against particular sectors of the population (England, 1992; Phelps, 1972). Following human capital theory, some suggest that women invest less in work-related capital and more in family obligations, resulting in less rewarding job opportunities (Becker, 1994).

Comparatively, gender stratification theory argues that women experience personal discrimination in work-related domains. Gender stratification theory has its roots in traditional segregation theories that argue women’s continual and often unjust segregation into particular occupations (Bergmann, 1974; Blau, 1972). In effect, women do not have equal opportunity in achieving rewarding job positions, despite their human capital investments (Bielby and Baron, 1986). Both rational choice and stratification perspectives provide explanations for the growing representation of women in precarious work settings; however, neither explains these patterns completely. Since both human capital theory and gender stratification theory suggest that women are more likely to work in precarious settings, similar findings are expected in the current study.

Hypothesis 1: Women are more likely to work in precarious settings.

Using Rodgers and Rodgers (1989) criteria of “precarious employment”, the following paper tests human capital and gender stratification theories by questioning: a) the gendered patterns of precarious employment; b) the effects of human capital investments and family obligations on precarious employment; and c) the extent to which human capital investments and family obligations affect precarious employment criteria differently for men and women. Results from the current study therefore provide an important contribution to the literature on the gender of precarious employment by offering a point of departure for analyzing why women remain overrepresented in sub-standard work settings compared to their male counterparts across all occupational settings.

Background

Human Capital Theory and Precarious Employment

A large literature on men’s and women’s workplace rewards and job situations has incorporated human capital models to explain gender discrepancies (Melamed, 1996; Reskin and McBrier, 2000). Human capital theory explains that those who invest in their careers through education, experience, skills, and training will be subsequently rewarded in the workplace and are more likely to obtain high quality jobs because he or she becomes more valuable to the employer. The more valuable a worker, the more rewards they receive in the labour market either through promotions, benefits, or raises (Becker, 1994). Such rewards diminish the precariousness of job positions by providing security.

Human capital investments in skills, training, and experience, also influence an individual’s occupation and productivity within that occupation (Becker, 1994). For example, those with more education and experience are likely to have higher quality jobs and greater career success (Kalleberg et al., 2000). Measurements of productivity are often contested, so given methodological constraints, the following study excludes productivity but focuses on three other attributes of human capital investments: education, experience, and occupation.

Several studies focus on the workplace benefits accrued as a result of capital investments in education, experience, and occupation. This literature supports human capital theory, reporting that job training, post-secondary education, and academic performance all influence one’s workplace rewards (Robson and Wallace, 2001). Others report the importance of occupation and productivity for earnings, specialization, and workplace rewards (Blau, Ferber, and Winkler, 2002; Leahey, 2007). Therefore, those with more education, experience, and occupational success should occupy less precarious work settings as defined by workplace rewards and job security.

Family Determinants

In addition to the various human capital investments that contribute to employees’ workplace precariousness, family also influences employment opportunities. Investments such as marital status, number of children, time spent caring for children, and time on household tasks may have negative implications for workplace rewards and security because they undermine the amount of time and effort one can dedicate to work responsibilities. Greenhaus and Beutell (1985) explain that competing family obligations may compromise an employee’s time, effort, and commitment towards their work roles, thereby jeopardizing their employment situation. Human capital theorists support this perspective and further argue that workers who prioritize family over work obligations may be less valuable in the labour market and subjected to more precarious work settings (Mincer and Polachek, 1974).

Human capital theory also states that since women are highly committed to their current, or expected family responsibilities, they invest less time and effort in their careers, which affects their experience, training, and occupational outcomes. Men, however, tend to invest more in their career than in their family responsibilities, which leads to greater employment rewards (Becker, 1985). In addition, this theory assumes that women’s skills depreciate while they are out of the workforce caring for their children (Reskin, 1993). The decrease in women’s skills at the expense of domestic responsibilities means women may instead choose jobs that are compatible with their child rearing obligations, where their skills do not depreciate over time. These jobs are generally characterized by low pay, part-time work, and few benefits because they require less valuable skills, including domestic and childcare-related responsibilities (i.e., cleaning and cooking services; daycare providers, etc.) (Glass, 1990; Webber and Williams, 2008).

Literature in this area documents three general family determinants that impact workplace rewards and security, including marital situation, childcare, and household responsibilities (Christie-Mizell, 2006; Kalleberg et al., 2000). Although working men are more likely to be married, women commit more time to their relationship compared to their partners (Stevens, Kiger, and Riley, 2001). Similarly, women generally spend more time than men on childcare and household chores (Bianchi, Robinson, and Milkie, 2006). So human capital theorists would suggest that women work in more precarious settings because they choose to invest heavily in their family instead of in their personal capital. So logically, men and women who invest equally in their family and in their human capital should be equally likely to work in lucrative, full-time, and therefore, non-precarious job positions. In accordance with these arguments, this study hypothesizes the following:

Hypothesis 2a: Human capital attributes decrease the likelihood of working in precarious settings, and these effects are equal for men and women.

Hypothesis 2b: Family-related investments increase the likelihood of working in precarious settings, and these effects are equal for men and women.

Gender Stratification Theory

Gender stratification theory does not deny that workplace rewards are a product of investments in human resources, but instead emphasizes inequalities between men and women in obtaining and benefiting from such investments (Blau, 1972; England, 1992; Reskin, 1993). So women’s greater presence in precarious work settings results not from personal choice, but rather from employers’ discriminatory practices to segregate women in these positions. In general, men and women do not have equal opportunities to invest in their careers, given women’s traditional role in the domestic sphere (Bergmann, 1974; England, 1992). Despite advances, the majority of these women still tend to their domestic responsibilities, including child care and household responsibilities, while others have decided to opt out of motherhood entirely (Blair-Loy, 2003; Fortin and Huberman, 2002; Gerson, 2004).

Notwithstanding women’s efforts to advance their human capital, employers may still inefficiently penalize women by offering them less lucrative, less secure, or ultimately precarious positions (Blau, 1972). Wright, Baxter, and Birkelund (1995) underscore several explanations of why employers would discriminate against women in the workplace including a) organizational norms perpetuating the value of men over women in the occupation (i.e., the “old boys club” mentality, see Epstein et al., 1995); b) the belief that women are too emotional to handle certain positions; and, c) the power of statistical discrimination in hiring decisions, which is discrimination based on stereotypes about women (Reskin, and McBrier, 2000). Taken together, these explanations address employers’ reservations towards a particular type of worker based on logical expectations regarding the workers’ productivity. These forms of workplace discrimination transcend comparable worth arguments for human capital indicators. In other words, women are more likely to be employed in precarious work settings as a result of employers’ discriminatory preferences, as opposed to human capital alone. These inequalities persist both across and within occupations, suggesting men are rewarded more for their efforts compared to women, despite occupational title or tasks (Acker, 1989; Fortin and Huberman, 2002). To this end, gender stratification theory hypothesizes that human capital investments operate differently for men’s and women’s job rewards and status.

Hypothesis 3a: Human capital investments decrease the likelihood of working in precarious settings, but these effects are greater for men than they are for women.

Family Determinants

In contrast to human capital theory, gender stratification theory argues that men’s and women’s job positions are affected differently by family investments, such as marital status, time on childcare, and time on household responsibilities (see Bianchi et al., 2006; England, 1992). More specifically, women are penalized for their current or expected family commitments because of discriminatory practices in the labour market (Budig and England, 2001; Wright et al., 1995). Consequently, women become segregated into more precarious work settings because of presumed, rather than actual, family investments. A similar argument is proposed for domestic obligations (Budig and England, 2001). Since traditional gender expectations designate women as primarily responsible for domestic responsibilities, employers internalize these stereotypes and hire based upon these expectations (Kaufman, 2002; Christie-Mizelle, 2006). Accordingly, gender stratification theory would argue that compared to women, men experience fewer work-related consequences for family investments.

Hypothesis 3b: Family-related investments increase the likelihood of working in precarious settings, but these effects are greater for women than they are for men.

Methods

The current study uses American data from the National Survey of the Changing Workforce (2002), with information on individuals across the United States. Potential participants included those 18 or older, employed in the labour force (Bond, Thompson, Galinsky, and Prottas, 2003). In total, 3,504 interviews were conducted with eligible participants between October 2002 and June 2003 using a random sample stratified by region. Interviews took approximately 45 minutes to complete using a computer-assisted telephone interviewing system. Eligibility included people who were 18 years of age or older, worked in the paid labour force, resided in the contiguous 48 states, and owned a telephone. One eligible participant was randomly selected from each household to complete the interview. Each participant was offered a $25 cash honoraria for completing the interview (Bond et al., 2003).

In total, 28,000 telephone numbers were called. Of these, 14,778 were non-residential or non-working numbers and 3,609 were considered ineligible (often language-related). The remaining 3,578 numbers were eligible, and of these households, 3,504 completed an interview. By these figures, the NSCW (2002) has a completion rate of 98 percent, excluding the eligibility criteria for the remaining 6,035 cases, which is difficult to determine. From the available eligibility information about these remaining 6,035 cases, Bond et al. (2003) propose that 3,146 of the 6,035 households were eligible for interviews. According to these figures, the NSCW (2002) maintains an overall response rate of 52% (N = 3,504 / (3,578 + 3,146).

Respondents missing information on focal measures were excluded from the current analyses. Data were also restricted to wage and salary workers, given that self-employed and contractual employment involve complexities that could not be accounted for in the current study due to measurement limitations. For example, precariousness of self-employed work situations may vary depending on business type. After restricting the data, a total of 2,024 cases were used for analyses: 1,189 were women and 835 were men. Descriptives for the current sample are presented separately for men and women in Table 1.

Measures for Precarious Employment. Four separate measures were used to represent precarious employment, including hourly wage, availability of benefits, union protection, and part-time work status. The validation of these measures was based on Rodgers and Rodgers’ (1989) definition of precarious employment, among other sources that emphasize particular work schedules as precarious (see Fuller and Vosko, 2007; Kim et al., 2008). In general, high quality, and therefore non-precarious work settings are operationalized by high hourly wages, the availability of benefits, the presence of union protection, and full-time work status (35 or more hours per week; U.S. Bureau of Labor Statistics, 2009). Hourly wages represent respondents’ hourly wage from their main job. Given the positive skew of the distribution, hourly wages was logged (after adding 1 to remove 0’s). Benefits are measured by averaging the availability of workplace benefits. Nine items were used asking about personal and family health benefits, retirement plans offered by employers, and paid vacation time. Responses are coded (2) yes or (1) no. Scores were summed and averaged by the total number of items. Union protection was measured by a single item asking if respondents are a member of a union OR collective bargaining unit. Responses are coded (1) yes or (0) no. Finally, part-time work status was measured by asking respondents how many hours they work per week. By national U.S. standards those who reported 34 hours or less per week were defined “part-time” compared to “full-time” (reference, 35 or more; U.S. Bureau of Labor Statistics, 2009).

TABLE 1

Descriptive Statistics for Men and Women, Human Capital Investments, Family Determinants, and Control Variables (N = 835, Men; N = 1,189, Women)

Descriptive Statistics for Men and Women, Human Capital Investments, Family Determinants, and Control Variables (N = 835, Men; N = 1,189, Women)

Note: Means for categorical variables represent the percentage of respondents in each category

Asterisks represent significant gender differences.

* p < .10; ** p < .05; *** p < .01.

-> See the list of tables

Gender. Respondents’ gender was coded (0) for women and (1) for men.

Work-related Measures. Human capital and work-related determinants are captured by three separate measures, including education, experience, and occupation. Education is represented by four categories: “less than high school” (reference category), “high school”, “some post secondary education”, and “4-year or graduate degree.” Occupation was measured by title of respondents’ main job. These responses were recoded based on the 1990 U.S. Census codes, including “professional” (executive, administrative, and managers), “technical” (technical, sales, and administrative support), “service” (private services, such as launders, or housekeepers; public services, such as police, health care, and personal services), and “labourers” (operators, production workers, and repair workers). Professionals were used as the reference category. These categories reflect commonly used occupational classifications (Bond et al., 2003). Experience is measured by respondents’ years working in the labour force since the age of 18. In accordance with NSCW (Bond et al., 2003) categories, these figures were categorized by “less than 12 years” (reference category), “12 to 21 years”, “22 to 30 years”, and “more than 30 years.”

Family-Related Determinants. Family-related determinants include marital status, number of children, time on childcare, and time on household responsibilities. Marital status included those “married (including common-law)” compared to respondents who have “never married” or were “previously married.” Number of children included all children in the household under 18. Time on childcare was gathered by asking, “On average, on days when you’re working, about how much time do you spend taking care of or doing things with your (child/children)?” Responses were coded in hours and multiplied by seven to represent hours spent on childcare per week. Time on housework includes respondents’ answer to the following question: “On average, on days when you’re working, about how much time do you spend on home chores—things like cooking, cleaning, repairs, shopping, yard work, and keeping track of money and bills?” Responses were multiplied by seven to represent hours spent on housework per week.

Control Measures. To properly specify the following model for precarious work settings, additional factors are controlled for, including organizational size, race/ethnicity, and immigrant status. Size includes 10 categories capturing approximately how many people are employed by the respondent’s company or organization at their work location, including themselves. Categories include “under 25”, “25 to 49”, “50 to 74”, “75 to 99”, “100 to 249”, “250 to 499”, “500 to 999”, “1000 to 5999”, “6,000 to 9,999”, and “10,000 or more.” Immigrant status measures whether the respondent is an immigrant to the U.S. in reference to those in the sample who are not. Finally, respondents’ ethnicity is measured as “White” versus all other categories of race, including “Black or African American”, “Native American or Alaskan Native”, “Asian, Pacific Islander, or Indian (from India)”, and “Other, including mixed.” Originally, age was controlled in all models; however, due to high collinearity with years of work experience (r = .84), this variable was removed from the model. Descriptives for the measures are reported separately for men and women in Table 1.

Analytical Strategy. Hypothesis 1 was tested by comparing significant differences in men’s and women’s mean values for each of the precarious employment measures (Table 1) by occupation (Table 2). These differences are identified by asterisks in each table. Next, the main effects of human capital and family-related investments were estimated, controlling for gender. This procedure serves two purposes: first, the effects of human capital and family-related determinants must be examined to answer my second research question. Second, to test hypotheses 2a and 3a, it is necessary to model these effects while controlling for gender to determine whether human capital and family-related investments explain any of the observed gender differences in precarious work criteria. If in fact human capital and family-related investments explain gender differences in precarious work, results would support human capital arguments that men and women simply invest differently in these situations. Table 3 presents the main effects model results. Models for hourly wage and benefits used ordinary least squares (OLS) regression techniques to estimate the effects of human capital investments and family-related determinants. Binary logistic techniques were conducted using maximum likelihood estimation to model the effects of human capital and family-related determinants on union protection and part-time work status.

TABLE 2

Cross-Tabulation of Precarious Employment Measures by Occupational Category (N = 2,024)

Cross-Tabulation of Precarious Employment Measures by Occupational Category (N = 2,024)

Note: Means for categorical variables represent the percentage of respondents in each category

Asterisks represent significant gender differences. * p < .10; ** p < .05; *** p < .01.

-> See the list of tables

TABLE 3

Ordinary Least Squares and Logistic Regression Results for the Effects of Human Capital Investments and Family Determinants on Precarious Employment Indicators (N = 2,024)

Ordinary Least Squares and Logistic Regression Results for the Effects of Human Capital Investments and Family Determinants on Precarious Employment Indicators (N = 2,024)

† Work hours per week was excluded from model 4 because of high collinearity between part-time status and hours worked.

Note: Pseudo R2 is reported for Models 3 through 4, in accordance with logit models.

a Compared to less than high school education. b Compared to professionals. c Compared to less than 12 years experience. d Compared to married.

* p < .05; ** p < .01; *** p < .001.

-> See the list of tables

After estimating the main effects of human capital and family-related determinants, controlling for gender, differences between men and women still persist among hourly wages and part-time work status. To better understand why these differences exist, I test hypotheses 3a and 3b from a stratification perspective using a gender interaction approach (see Wright et al., 1995: 409). Recall hypotheses 3a and 3b argue that human capital and family-related determinants function differently for men and women, which would be supported by the presence of significant interactions between gender and human capital and family-related explanatory measures. A gender interaction approach tests stratification theory by estimating additional equations for the precarious work measure in question. These equations include gender interactions for each of the explanatory measures, minus the controlled conditions. Interaction models were only tested for hourly wages and part-time work since gender differences in access to benefits and union protection were already explained by the main effects. The main models for hourly wages and part-time work were each re-estimated two times: an additional model for the gender interactions with human capital investments and one for the gender interactions with family determinants. Table 4 highlights significant multiplicative terms by comparing differences in overall model fit. Significant r-square (R2) and chi-square (?2) values are presented for each set of interaction terms where applicable.

The initial results in Table 4 reveal several gender-contingent work and family-related effects. These results justify re-estimating separate models for men’s and women’s hourly wages and part-time work status. A total of four models were estimated in this last step of analyses. Table 4 presents the regression coefficients and odds ratios for human capital and family-related determinants for men’s and women’s hourly wages and part-time work status.

Results

Gendered Differences in Precarious Work Settings

Table 1 reports descriptives for men (n = 989) and women (n = 1,189) separately. This table also highlights significant gender differences for each of the four dependent measures, including hourly wage, benefits, union protection, and part-time employment. According to these tests (not shown), men earn significantly more per hour compared to women ($22.77 for men; $15.97 for women), and are more likely to have benefits (mean = 1.812 for men; mean = 1.762 for women). Men are also more likely to have union protection (21% of men; 18% of women), and compared to men (6%), women are more likely to work part time (17%). Figure 1 presents a pictoral reference of the gender differences in precarious work criteria. Table 2 suggests gendered patterns persist across almost all occupational categories On average, women receive less hourly pay and benefits in all occupations, and are less likely to have union protection and more likely to work part-time. These results contribute to my first research question—what is the gendered nature of precarious employment? Results support hypothesis 1, which predicted that women would be more likely than men to work in precarious settings or situations.

TABLE 4

Gender-Interaction Effects of Determinants on Precarious Employment Indicators (N = 2,024)

Gender-Interaction Effects of Determinants on Precarious Employment Indicators (N = 2,024)

Note: All models control for firm size, immigrant status, and ethnicity. Models 1 through 4 also control for hours worked per week.

† Significance refers to difference in R2 between the main effects and interaction models.

a Compared to less than high school education. b Compared to professionals. c Compared to less than 12 years experience. d Compared to married.

* p < .05; ** p < .01; *** p < .001.

-> See the list of tables

FIGURE 1

Gender differences in Precarious Employment Criteria

Gender differences in Precarious Employment Criteria

-> See the list of figures

Human Capital and Family Determinants

Table 3 presents results for the main effects of human capital investments and family-related determinants. As expected, those with higher education and experience earn considerably more per hour, accrue more benefits from work, and are more likely to have union representation. Education has little effect on the likelihood of working part-time. The results for occupation are mixed, but suggest that overall, professionals (reference category) fare better, compared to those in technical, service, or labouring occupations. Family determinants affect precarious work situations in that more time on childcare and housework decreases hourly wages and benefits, while increasing the likelihood of working part-time. These results answer my second research question, suggesting that human capital investments decrease the likelihood of working in precarious settings (Hypothesis 2a), while family-related investments increase the likelihood of working in precarious settings (Hypothesis 2b).

Hypotheses 2a and 2b also predicted that human capital and family-related determinants function similarly for men and women. The absence of a significant gender effect on precarious employment criteria after accounting for human capital and family-related investments (Table 3) suggests that these investments function similarly for men and women, and any observed gender differences in benefits or union representation is a product of different levels of investments. To better understand these gendered patterns, it is necessary to review the mean differences in men’s and women’s human capital and family determinants. These differences are highlighted in Table 1 and suggest that women, although equally likely to have high education and hold professional occupations, have less workplace experience and spend considerably more time on childcare and housework, which may explain the gender differences in men’s and women’s reported workplace benefits and union protection. The control measures may also account for these gender differences: size of workplace, hours of work per week, and being born in the U.S. positively affect benefits and increase the likelihood of union protection. Men and women differ significantly on these three factors, in that men work more hours, in larger firms, and are more likely to have immigrant status. Overall then, the gender discrepancies between men’s and women’s benefits and union representation are not necessarily a product of gender biases towards men and women in the workplace, but rather result from differences in men’s and women’s human capital, family investments, workplaces, and immigrant status.

Turning again to Table 3, results suggest that significant gender differences still ensue between men’s and women’s hourly wages and the likelihood of working part-time, even after accounting for men’s and women’s human capital and family-related investments. These results support gender stratification hypotheses 3a and 3b. Like the human capital hypotheses, these hypotheses predicted that human capital investments decrease the likelihood of working in precarious settings (hypothesis 3a), while family-related investments increase the likelihood of working in precarious settings (hypothesis 3b); however, these effects function differently for men and women. To better understand these differences, it is necessary to first review the results of the gender-interaction models, and second, to review the main effects of human capital and family-related investments separately for men and women.

Gender Interactions: Human Capital and Family-related Determinants

Table 4 presents the gender interaction for the effects of human capital and family-related determinants on hourly wages and the likelihood of working part-time. According to the results from models 1 and 2, education, labour work (compared to professional work), experience, marital status, and number of children affect hourly wages differently for men and women. The gender-interaction effects on part-time work tell a different story, suggesting that technical work (compared to professional), experience, and time on housework affect the likelihood of working part-time differently for men and women. Each set of interaction terms improves the fit of the models, as suggested by the significant change in the (Pseudo) R2 / log likelihood (χ2), which compare the main effects and multiplicative models. The gender contingent effects of human capital and family-related determinants are more easily interpretable by separating the main effects model for men and women.

Focussing on the effects of human capital and family determinants on hourly wages (Table 5, models 1 and 2), it appears that for men, high school and some secondary education increase men’s hourly wage, but additional education has little effect on women. Gender differences in the effects of experience tell a similar story. Compared to women, men with more experience receive greater monetary benefits. And while the number of children men have benefits men’s overall wage, never being married can be detrimental. These two factors have no effect for women’s hourly wages.

TABLE 5

Ordinary Least Squares and Logistic Regression Results for the Effects of Human Capital Investments and Family Determinants on Precarious Employment Measures for Men (N = 835) and Women (N = 1,189)

Ordinary Least Squares and Logistic Regression Results for the Effects of Human Capital Investments and Family Determinants on Precarious Employment Measures for Men (N = 835) and Women (N = 1,189)

Note: Pseudo R2 is only reported for models 3 through 4, in accordance with logit models.

a Compared to less than high school education. b Compared to professionals. c Compared to less than 12 years experience. d Compared to married.

* p < .05; ** p < .01; *** p < .001.

-> See the list of tables

Turning to the gender specific effects for part-time work (Table 5, models 3 and 4), work experience decreases the likelihood of men working part-time by approximately 70%. Yet overall, experience has little effect for women. Men who are technical workers are also more likely to work part-time—approximately four times more likely compared to professionally employed men, and while women in this occupation are also more likely to work part-time, this effect is much less. Housework also affects the likelihood of working part-time, and this negative effect is stronger for women compared to men.

Control Measures

The size of respondents’ workplace, weekly work hours, immigrant status, and race/ethnicity also affect job quality. Table 2 reports that those working in larger organizations receive higher wages, more benefits, and are less likely to work part-time. Working more hours per week is negatively related to hourly wage, yet positively related to benefits and the likelihood of union protection (recall this variable was excluded from the model estimating the likelihood of working part-time because of high collinearity;r = .84). Immigrants are less likely to have benefits or union protection and Whites are less likely to have union protection, compared to ethnic minorities.

Discussion and Conclusions

The current study set out to answer three specific questions on: a) the gendered patterns of precarious employment; b) the effects of human capital and family-related investments on precarious employment; and c) the extent to which these investments and obligations affect precarious employment criteria differently for men and women. Based on commonly used measures of precarious employment, individuals’ hourly wage, workplace benefits, union protection, and part-time work status were considered the focal indicators of high quality and therefore non-precarious work settings.

Results from this study contribute to the literature on precarious employment in three ways: a) there is a “gender” to precarious employment in that women are more likely to work in low quality job settings, characterized by less financial rewards, fewer benefits, union protection, and part-time work status; b) women may be subjected to more precarious employment in part because of less human capital, their overriding family obligations, and other work-related criteria, which supports human capital predictions; however, c) in accordance with gender stratification theory, women’s overrepresentation in low paying, part-time work situations is partially a consequence of discriminatory practices in the workplace. Women experience disadvantages in the labour market because their human capital and family-related investments function differently compared to men’s. These results suggest that neither human capital theory, nor gender stratification theory fully explain why women are overrepresented in precarious work; rather I propose that each perspective provides a unique contribution to understanding the gendered patterns of precarious employment. Implications and explanations of these findings are discussed in the following section.

The “Gendered” Nature of Precarious Employment

The findings of this study support previous literature documenting the quality of women’s job settings, which suggest that despite women’s advances in the labour market, many are still situated in less rewarding and less secure work settings (Cranford et al., 2003; Fuller and Vosko, 2007; Kalleberg et al., 2000; Ruyter and Warnecke, 2008; Vosko, 2006). The results from the current study support human capital and gender stratification theory in that women are more likely to work in precarious settings, which is evidenced by women’s lower hourly wages, fewer benefits, absence of union protection, and over-representation in part-time positions. Yet, unlike previous studies, the current project goes beyond establishing women’s over-representation in precarious work settings by discussing why these patterns persist.

Human Capital and Family-Related Determinants: Implications for Precarious Employment

To better understand gender differences in precarious employment, this study also examined the implications of human capital and family-related determinants on precarious employment criteria. Recall that human capital theory argues that the more individuals invest in their personal capital, including work experience, education, and training in particular occupations, the more rewards they will receive in the labour market. Alternatively, the more individuals invest in family, including childcare and household chores, the less likely they are to be rewarded in the workplace, given that these investments are relatively undervalued by employers (Becker, 1994; Mincer and Polachek, 1974). Results from the current study support human capital theory in that those with more education, labour market experience, working in more prestigious professions receive greater wages, more benefits and are more likely to work full-time with union protection. Investments in family present patterns expected by human capital theory in that those who spend more time on household and childcare likely work in more precarious settings (Becker, 1994).

Gender Differences in Human Capital and Family-Related Investments

In accordance with human capital theories, this study found that gender differences in precarious employment may result, at least in part, by differences in human capital and work-related criteria. In particular, women report less labour market experience necessary to accrue greater workplace rewards, benefits, or more secure, full-time positions. Men and women also differ in terms of family-related investments, in that women spend far greater time on weekly housework, which arguably signals lower workplace commitment (Blair-Loy, 2003), which can result in fewer workplace rewards. The results from this study also suggest that gender differences in precarious employment relate to men’s and women’s work hours and the size of their work setting. Men likely work more hours in larger firms, both of which lead to greater workplace rewards especially in terms of union protection and benefits (Cranford et al., 2003). More work hours accrues additional benefits because employees are often rewarded for their commitment through promotions and extensive workplace benefits. Larger firms are more likely to be unionized because they have political power in numbers, which facilitates solidarity and union action (Hirsch and Addison, 1986).

Gender discrimination arguments help explain these differences in work experience, work-settings, and family-determinants by claiming that although some forms of human capital and family investments work similarly for men’s and women’s rewards, women may not have equal opportunity to receive jobs in larger organizations, gain labour market experience, or work the necessary hours required in highly secure, rewarding positions. Such explanations are supported by previous research in the U.S. (Britton, 2000; Reskin, 1993). These studies and others suggest that although women are meeting, and in some cases exceeding men’s educational achievements, women are still disadvantaged in the areas of labour market experience, occupational attainment (Britton, 2000), and work hours because of their overriding, and often expected family obligations (Leahey, 2007; Milkie and Peltola, 1999; Suitor et al., 2001). Such barriers explain differences in men’s and women’s human capital investments, which in turn explain why women are still overrepresented in precarious work settings.

Gender Stratification: Women’s Segregation in Precarious Employment

Women’s over-representation in precarious work settings is also a function of discriminatory practices on behalf of the employer. The current study suggests that human capital and family-related investments function differently for men and women when it comes to precarious employment. Compared to women, men with equal education and work experience receive higher hourly wages and are less likely to work part time. And, compared to men, women who spend equal time on household chores are more likely to be penalized, receiving less pay and are more likely to work part-time positions. These findings suggest that women are more likely to be employed in precarious work settings as a result of employers’ discriminatory preferences, as opposed to human capital alone. As mentioned previously, there are several reasons why employers discriminate against women in the workplace including a) organizational norms perpetuating the value of men over women in the occupation; b) the belief that women are too emotional to handle certain positions; and, c) the power of statistical discrimination in hiring decisions, which is discrimination based on stereotypes about women. Taken together, these explanations address employers’ reservations towards a particular type of worker based on logical expectations regarding the workers’ ability and commitment (England, 1992; Reskin, 1993; Wright et al., 1995). Results from the current study confirm that such workplace discrimination transcends comparable worth arguments for education, experience, and other human capital indicators, especially when it comes to women’s wages and part-time work statuses.

Despite the contributions of the current study, there are several limitations worth highlighting. First, the cross-sectional design of these data restricts confidence in causal ordering of the focal associations. For example, it is highly plausible that women who spend considerable time on household and childcare choose to work part-time rather than full-time jobs. Future research might integrate these relationships using longitudinal data that tracks women’s entry into particular precarious positions given their family-related obligations. In addition, selection effects are worth considering. It is likely that individuals with higher education choose to work for companies who offer extensive benefits and union protection. They may select more rewarding jobs, therefore avoiding precarious situations. Such considerations have implications for my estimates and study results.

Second, several measurement issues are worth noting. The four measures of precarious employment used in this study could be improved. For example, the index for benefits could be examined across types of benefits, such as family, personal, and retirement. Wages could be examined in the context of other monetary workplace rewards, such as vacation pay and sick leave. Finally, workplace security and job loss should also be considered in future research on precarious employment. These considerations were not tested in the current study given measurement restrictions of the NSCW data.

In closing, this study provides explanations for why women are continually segregated in precarious work settings. By highlighting gender differences in human capital and family-related investments, researchers can focus more on why women continue to report less work experience, more time on domestic responsibilities, and fewer work hours in smaller organizations, and document the extent to which these patterns are changing over time. Although women have made advances in the workplace in recent decades, the current study suggests that women are still disadvantaged overall in terms of human capital investments and in how these investments function in the workplace when determining placement in precarious work situations