Perceived Job and Labor Market Insecurity in the United States: An Assessment of Workers’ Attitudes From 2002 to 2014

Abstract

The distinction between perceived job insecurity (workers’ assessment of their likelihood of losing their job) and perceived labor market insecurity (workers’ assessment of their ability to find another job similar to their current position) is important because the theoretical primacy of perceived job insecurity is diminished in the context of the risk regime, presenting a need for work that identifies whether the traditional factors of security still protect workers from feeling insecure. The author addresses this need by analyzing data from the Quality of Working Life module of the General Social Survey, which was asked of respondents in 2002, 2006, 2010, and 2014. The author finds that while characteristics of secure work are still associated with lower perceived job insecurity, most tend to increase perceived labor market insecurity. The author also finds that the Great Recession produced higher levels of perceived labor market insecurity. These findings suggest that jobs previously considered to be “good” jobs such as government and union jobs may be more of a liability than an asset in the context of decreased employer–worker attachment.

Keywords

subjective employment insecurity perceived job insecurity perceived labor market insecurity General Social Survey

Given the centrality of paid employment in workers’ lives, it is not surprising that there has been an increased focus on employment insecurity in the United States. Previous research has examined employment insecurity in two distinct ways: objectively and subjectively (De Witte & Näswall, 2003; Heery & Salmon, 2000; Klandermans, Hesselink, & van Vuuren, 2010). While objective employment security tends to be measured in a more concrete sense, focusing primarily on measures such as unemployment or job retention (Diebold, Neumark, & Polsky, 1997; Farber & Valletta, 2015; Swinnerton & Wial, 1995), measures of the subjective experience of employment security are based on workers’ perceptions of their employment situation and that of others. Examining workers’ perceptions of their employment insecurity is a particularly important endeavor because unlike traditional measures of employment stability, which concentrate on the economic well-being of workers who have lost their jobs, “we can learn about how changes in job stability and job loss are affecting all workers only by looking at data on job security perceptions” (Schmidt, 2000, p. 300).¹

Perceived insecurity has also been associated with many types of negative consequences for workers, such as worse physical health (Burgard, Brand, & House, 2009; Fullerton & Anderson, 2013), worse mental health (Burgard et al., 2009; Norris, 2016), a low sense of personal control (Glavin, 2013), and low job satisfaction (Artz & Kaya, 2014; Narisada & Schieman, 2016). In fact, the anticipation of job loss can be just as stressful as unemployment (Kim & von dem Knesebeck, 2015). Also, perceived employment insecurity can be indicative of macrostructural economic transformations and often condition workers’ actions within the labor market. Higher levels of perceived insecurity lead workers to experience a greater sense of vulnerability, making them less likely to fight for higher wages and fringe benefits.

However, the existing research on employment insecurity has placed an unduly narrow focus on perceived job insecurity. The current occupational structure is characterized by shorter job tenures, a rise in nonstandard work arrangements, and fewer career ladders (Farber, 2008; Kalleberg, 2011; Kalleberg & Vallas, 2017; Osterman, 1996). It is well established that these are features of a new post-Fordist era that originated with the erosion of the social contract that once existed between capital and labor (Kalleberg, 2011; Rubin, 2014). Currently, this new era is dominated by what Beck (2000) identifies as the risk regime or “the political economy of insecurity, uncertainty and loss of boundaries” (p. 73). Under this regime, risk is increasingly assumed by workers as they are becoming more responsible for developing their own skills and managing their “boundaryless” or “nomadic” careers in what has colloquially been called the “gig economy” (Arthur, 2014; Arthur & Rousseau, 2001; Tremblay, 2008). The control workers have over their work can entice them to invest effort in their work (Occhiuto, 2017), but the uncertainty and risk often outweigh the benefits. While measures of perceived job insecurity are not obsolete per se, they fail to fully capture the economic anxiety experienced by many present-day workers who view their employment arrangements as short term from the outset.

The decreased attachment between employers and workers that characterizes the risk regime highlights the need for a broader conceptualization and operationalization of perceived employment insecurity. Namely, there is a difference between workers’ assessment of their likelihood of losing their job, or perceived job insecurity, and workers’ feelings of insecurity concerning their ability to find another job similar to their current position, or perceived labor market insecurity (Dixon, Fullerton, & Robertson, 2013; Kalleberg, 2011). In their comparative investigation of perceived employment insecurity in European countries, Dixon et al. (2013) distinguish between the two and conceptualize them in tandem as employment insecurity. I argue that the theoretical primacy of perceived job insecurity is diminished in the context of the risk regime and that it is crucial to consider perceived labor market insecurity alongside its more traditional counterpart, perceived job insecurity. The present study builds on Dixon et al.’s (2013) valuable conceptualization and applies it to the U.S. workforce to gain a more complete and nuanced understanding of workers’ attitudes.

In addition to the disproportionate attention given to job insecurity, the literature on perceived employment insecurity has other limitations. For instance, the existing work tends to concentrate on longer spans of time (Fullerton & Wallace, 2007; Schmidt, 1999, 2000) and does not adequately address the impact of the Great Recession, the most detrimental economic event since the Great Depression. The Great Recession has highlighted for many workers just how precarious their employment situation has become, and yet systematic studies of its impact on perceived employment insecurity are few and far between. Also, many of the studies that have covered perceived employment insecurity have been situated in the European context (Dixon et al., 2013; Erlinghagen, 2008; Laszlo et al., 2010). Research that takes a focused look at both perceived job and labor market insecurity exclusively in the contemporary United States is needed. However, the most notable limitation is the emphasis on the consequences perceived employment insecurity has for other outcomes, such as health (Burgard et al., 2009; Fullerton & Anderson, 2013) and participation in job training (Elman & O'Rand, 2002). Although there exists a core group of studies that highlight the patterns and determinants of perceived employment insecurity (see Erlinghagen, 2008; Fullerton & Wallace, 2007; Jacobs & Newman, 2008; Kalleberg, 2011; Kalleberg & Marsden, 2012; Schmidt, 2000), few studies have used a wide range of predictors that encompass the various aspects of security, control, and uncertainty that characterize the modern workplace in the new economy. More work that addresses how these characteristics differentially affect perceived job and labor market insecurity is required.

The purpose of this study is to explore the distinction between the two dimensions of perceived employment insecurity under the current risk regime and understand the work-related factors that shape these outcomes in the United States from 2002 to 2014. In doing so, it contributes to the existing literature on perceived employment insecurity in several ways. First, it reinforces the distinction between perceived job and labor market insecurity, which has been highlighted by recent scholarship but not fully explored (Dixon et al., 2013; Kalleberg, 2011). Second, it provides a focused look at an important period of time, examining data from before and after the Great Recession, a seminal event in the economic history of the United States. Finally, this research accounts for job characteristics not considered by previous examinations of perceived employment insecurity, examining whether traditional factors of security and control continue to alleviate workers’ perceived employment insecurity in an uncertain economy.

Background

As noted earlier, U.S. society is currently operating under a risk regime, in contrast to the Fordist regime that dominated most of the latter half of the 20th century. Beck (2000) notes that “whereas the Fordist regime brought about the standardization of work, the risk regime involves an individualization of work” (p. 70). This individualization of work has been facilitated by the structural shift of accumulation practices toward flexibility and spatialization (Kalleberg, 2011, 2012; Wallace & Brady, 2010). This individualization means that workers are responsible for shouldering a greater share of risk as they navigate an employment system in which the perceived promise of security and stability has disappeared. The result is that workers are now holding more jobs over the course of their careers than in the past (Bernhardt, Morris, Handcock, & Scott, 2001; Farber, 2008), highlighting the increasing relevance of perceived labor market insecurity. Kalleberg (2011) also presents other evidence on the growth in precarious employment relations, citing the growing importance of nonstandard and contingent work, the decline in job stability, the weakening of firm internal labor markets, and the spread of long-term unemployment. Beck (2000) claims that the distinguishing attribute of the risk regime is that it is not simply a transitional phase and that there will be no return to the standardization of work that existed under the Fordist regime. Indeed, the reality of precarious employment is structurally embedded in our institutions, being more than a long-term ebb in the business cycle (Kalleberg, 2011, 2012; Lee & Kofman, 2012). Scholars point to globalization, technological advancement, deunionization, and financialization as some of the processes responsible for the advent of the risk regime (Appelbaum, 2012; Beck, 2000; Kalleberg, 2011; Madrick, 2012; Rubin, 2014).

The factors that underpin the risk regime are not only structural, but cultural as well. Vallas and Prener (2012) identify a culture of enterprise that glorifies individual responsibility and seems to have emerged concurrently with the risk regime. They contend that this culture of enterprise represents a discursive shift that has had a considerable influence on the polarization of work while also reframing employment uncertainty as a source of liberation and an opportunity for worker agency. Navigating this culture requires an increasing amount of identity work (Vallas & Christin, 2018). While the data of the present study do not allow for an examination of the relationship between this discursive shift and the patterns of subjective insecurity, the presence of this culture of enterprise bolsters the argument made earlier for breaking perceived employment insecurity into two dimensions and examining them in tandem. That is, within the contexts of the risk regime and this culture of enterprise, it is just as important for workers to be weighing the possibility of working at some other firm as it is for them to be concerned about the job security of their current position.

While the risk regime was established over time independent of the business cycle, the business cycle can still play a prominent role in workers’ perceptions of insecurity, particularly when it comes to major recessions. In 2007, the United States experienced a substantial wave of foreclosures on subprime, adjustable rate home mortgages, which was the result of a confluence of events, including a growing housing bubble, greater financial deregulation, and increases in predatory lending practices. These foreclosures triggered a massive financial crisis known as the Great Recession, the likes of which has not been seen since the Great Depression of the 1930s. Home values plummeted, financial institutions failed, and unemployment sharply increased. Employment insecurity is situated under the broader conceptual umbrella of economic insecurity, which also includes housing-based insecurity and health care-related insecurity. Due to the nature of the American housing market and health-care system, employment insecurity often serves as the bridge to these other dimensions of economic insecurity and is arguably the most important of all. For instance, Dominitz and Manski (1997) found that, on average, expectations concerning health-care coverage and job loss were quite similar. The Great Recession is a stark exemplar of how employment and housing insecurity are inextricably linked with each other. Predictably, employment insecurity has been a major feature of the national discourse in the United States since the Great Recession. The outsized impact of this crisis and its subsequent salience in the public imagination lead to the following hypotheses:

H_1A: The Great Recession will be positively associated with perceived job insecurity, net of other covariates.

H_1B: The Great Recession will be positively associated with perceived labor market insecurity, net of other covariates.

Characteristics of Security and Control

As employment relations swing back toward flexibility and uncertainty (Kalleberg, 2011; Polanyi, 1944) and social protections for workers steadily erode (Hacker, 2008; Standing, 2002), it might seem that having a job with the traditional characteristics of security and control can mitigate the risk regime’s unfavorable institutional context when it comes to shaping workers’ perceptions of employment insecurity. However, just as the rise of the “boundaryless” career (Arthur & Rousseau, 2001) lends credence to the notion that perceived job insecurity may not be as important as perceived labor market insecurity, this trend also provides a good justification for questioning whether these traditional factors will mitigate both perceived job and labor market insecurity. The emergence of the risk regime suggests that while these traditional factors may still be beneficial for workers regarding perceived job insecurity, they may be more of a liability for their perceived labor market insecurity. This is because the risk regime represents a race to the bottom in which those jobs characterized by stability will become increasingly rare, and those workers who still have such a job may feel that it would be difficult to find another like it. Therefore, in addition to the characteristics of uncertainty that are implicit features of the risk regime, it is also important to investigate how the factors of security and control shape perceived employment insecurity.

There are several employment situations that are traditionally viewed as having more security and control than others. Being employed in federal, state, or local government is generally perceived as trading in more lucrative pay for a decreased likelihood of losing one’s job. In their examination of General Social Survey (GSS) data from 1996 and 1998, Wilson and colleagues (Wilson, Eitle, & Bishin, 2006; Wilson & Mossakowski, 2009) found a negative effect of public sector employment on perceived job insecurity. Although more recent research has shown that African Americans do not enjoy the same protections as Whites in the public sector (Byron, 2010; Wilson & Roscigno, 2016), it is likely that the findings of Wilson et al.’s previous research will be replicated in the present study when workers are analyzed in the aggregate. Thus, I offer the following hypothesis:

H_2A: Being a government employee will be negatively associated with perceived job insecurity, net of other covariates.

Despite the evidence showing that public sector employment contributes to reduced feelings of job insecurity among workers (Wilson et al., 2006; Wilson & Mossakowski, 2009), perceived labor market insecurity operates differently. While the conditions of government employees’ employment provide them with a degree of objective job security, that security has a downside when it comes to looking for new employment. Namely, the skills and knowledge required to occupy these positions tend to have limited applicability to jobs at other firms. Because Anderson and Pontusson (2007) found that public sector employment had positive effects on labor market insecurity in their analysis of 15 Organisation for Economic Co-operation and Development countries, I expect that:

H_2B: Being a government employee will be positively associated with perceived labor market insecurity, net of other covariates.

Wilson et al. also found that being a member of a union had no significant impact. However, their analysis examines only data from the 1996 and 1998 GSS, which is a small snapshot in time. Brochu and Morin’s (2012) analysis of GSS data from 1978 to 2008 finds that workers who are members of unions are more likely to feel insecure than those who do not belong to such an organization. They speculate that “union leaders may want their members to feel more insecure and therefore more dependent on the union for security” (Brochu & Morin, 2012, p. 277). Therefore, I hypothesize that:

H_3A: Being a union member will be positively associated with perceived job insecurity, net of other covariates.

Unionization has been in decline in recent decades, which has also eroded the quality of nonunion jobs. The result has been a scarcity of jobs that offer the protections and benefits that are typically associated with union jobs. Workers who are union members are likely aware of how rare these protections and benefits have become. In addition, Anderson and Pontusson discovered a positive association between being a union member and perceived labor market insecurity in their comparative study. For the United States in particular, Brochu and Morin’s (2012) analysis reveals a positive correlation between union membership and perceived labor market insecurity. Given these findings, I anticipate that:

H_3B: Being a union member will be positively associated with perceived labor market insecurity, net of other covariates.

Job tenure is a staple in the work and occupations literature and a key concept that can provide insight as to how workers’ feelings of insecurity can change over the duration of their position. The extant research has consistently shown that longer job tenures translate into workers feeling more secure about their jobs (Pedulla, 2013; Wilson et al., 2006; Wilson & Mossakowski, 2009). This is likely a result of the job-specific skills that workers have learned over the course of their tenure, which allow them to complete their work efficiently, instilling a greater sense of confidence in their ability to keep their jobs. In addition, it gives workers an insiders’ advantage over any competition, which might be hired from outside the firm. Consequently, I hypothesize that:

H_4A: Length of job tenure will be negatively associated with perceived job insecurity, net of other covariates.

However, the same reasoning also applies to the association between job tenure and perceived labor market insecurity, with previous research presenting evidence of a positive relationship (Dixon et al., 2013). Namely, workers who have occupied their current positions for an extended period of time have accrued a very specific skill set that may not easily transfer over to other positions. This will likely undermine their assessment of their ability to land another job, leading me to hypothesize that:

H_4B: Length of job tenure will be positively associated with perceived labor market insecurity, net of other covariates.

Autonomy is an important job characteristic for many workers, regardless of whether they have it or not. For those who do have some degree of job autonomy, it buffers the detrimental effect of job insecurity on health (Schreurs, van Emmerik, Notelaers, & De Witte, 2010). Contract workers with autonomy can use their control to avoid common pitfalls of standard work such as overtime (if desired) and overwork (Osnowitz & Henson, 2016). The relationship between job autonomy and perceived job insecurity has been examined by previous research with mixed results. While some studies have found that job autonomy has no significant association with perceived employment insecurity (Wilson et al., 2006; Wilson & Mossakowski, 2009), these studies use data from before or at the very start of the contemporary era. As Wallace and Brady (2001) note, the contemporary era represents a structural shift in the techniques used to accumulate capital. This new era is characterized by technocratic control, which uses information technology to manage workers (Burris, 1993, 1998). This means that those workers who have any degree of job autonomy will likely have a great deal of technical expertise or a good working relationship with technical experts (Burris, 1993). For that reason, they tend to be relatively advantaged in many respects compared with other workers and are thus more likely to feel more secure about their jobs. In their examination of data from the 2004 and 2006 GSS, Wilson and Mossakowski (2012) find that this is indeed the case. Therefore,

H_5A: Job autonomy will be negatively associated with perceived job insecurity, net of other covariates.

While the general premise of this study is that the traditional factors of security and control may be more of a problem for workers’ perceived labor market insecurity, autonomy is the exception. The expertise that is often required for jobs with high autonomy makes these workers highly employable and attractive to employers, particularly in the information economy. It is expected that this employability would translate into a high degree of confidence for workers considering searching for a comparable job.

H_5B: Job autonomy will be negatively associated with perceived labor market insecurity, net of other covariates.

Characteristics of Uncertainty

The relationship between the two dimensions of perceived employment insecurity and predictors that reflect uncertainty in the work place, the labor market, and in workers’ personal lives is no less important. For example, being recently laid off is very likely to predispose someone to a continuing anxiety of being thrust unexpectedly into the same predicament, especially if they believe (rightly or wrongly) that they were laid off for reasons specific to them. Over the course of the Great Recession, many workers who felt their positions secure were laid off. One argument made earlier is that this pivotal economic crisis has aroused a dormant fear in the U.S. workforce that has been developing within the context of the risk regime. As Rubin (2014) states, “the persistence of displacement as a normal business strategy reinforces the insecurity and tractability of the currently employed” (p. 1089). Due to this backdrop of insecurity, it is expected that:

H_6A: Having been laid off in the past year will be positively associated with perceived job insecurity, net of other covariates.

Just as perceived labor market insecurity is distinct from perceived job insecurity, the expected effect of being recently laid off is also different. Because this study focuses on the perceptions of employed workers, any who have been recently laid off have been successful in being reemployed after the separation. This success is expected to translate into greater confidence in the worker’s ability to find another job, which means that they are more likely to have a positive perception of the labor market than those who have not been recently laid off. Therefore:

H_6B: Having been laid off in the past year will be negatively associated with perceived labor market insecurity, net of other covariates.

Part-time employment is another aspect of the modern employment relationship that can be tinged with uncertainty. It is important because of the increasing role that nonstandard work arrangements are playing in the transforming employment relationship, both generally and at specific points in the occupational structure. While Fullerton and Anderson (2013) found that being a part-time worker had no significant effect on perceived job insecurity, part-time employment has been associated with poorer working conditions such as irregular hours and a lack of benefits (Kalleberg, Reskin, & Hudson, 2000), leading to the expectation that those who are working under such conditions will likely be more pessimistic about the security of their position. In addition, Fullerton and Wallace (2007) found that part-time employment status decreased perceived job security, and Schmidt (2000) found that it increased the respondents’ assessed likelihood of losing their jobs. Also demonstrating the importance of part-time employment, Farber (1999) found that workers have used alternative and part-time employment arrangements as responses to actual job loss. This means that respondents who are temporary or part-time workers are more likely to have experienced a job loss, making them more susceptible to increased feelings of job insecurity. In addition, Kalleberg and Marsden (2012) found in their examination of 1977 to 2006 data that being a full-time worker decreased perceived job insecurity, meaning that being a part-time worker will likely work in the opposite direction:

H_7A: Being a part-time worker will be positively associated with perceived job insecurity, net of other covariates.

However, it would be an oversimplification to hypothesize that being a part-time worker would have a uniform positive effect on all dimensions of perceived employment insecurity. The rationale behind the hypothesis for perceived job insecurity actually implies the opposite for perceived labor market insecurity. Because part-time workers tend to be concentrated in the low-skill secondary labor market, which is characterized by more job turnover, they have likely adjusted their expectations to conform to this reality. That is, these workers are more likely to have made lateral movements into other low-skill jobs with similar compensation and will have a more favorable view of their prospects in the labor market. Also, Kalleberg and Marsden (2012) found that being a full-time worker increased perceived labor market insecurity. Therefore,

H_7B: Being a part-time worker will be negatively associated with perceived labor market insecurity, net of other covariates.

Data and Methods

The data for this analysis come from the GSS, a multistage, stratified sample of the adult population in the United States that has been administered since 1972. Specifically, this research uses the Quality of Working Life (QWL) module that provides comprehensive, cross-sectional data from 2002, 2006, 2010, and 2014 and asks respondents about the conditions under which they worked and their experiences in the workplace.² In fact, many of the questions are taken from or inspired by the 1977 Quality of Employment Survey. This module is especially suited for understanding the factors that shape perceived job and labor market insecurity in the contemporary period because it contains a rich array of data not found in the cumulative file of the GSS.

Like many other surveys of its kind, the GSS contains a substantial amount of missing data. This necessitates the use of imputation techniques so that the sample sizes for each analysis can be maximized. All missing data were handled using the multiple imputation protocol in Stata 13.1. Using multiple imputation techniques offer distinct advantages over alternative methods of handling missing data such as listwise deletion or mean replacement. The point estimates that are produced by multiple imputations have been shown to be approximately unbiased when compared with other methods (Allison, 2000; Rubin, 1987, 1996). However, although this approach is useful for producing unbiased estimates, it does have one significant drawback. Because the reported coefficients are derived not from a singular point estimate but from several estimates, log-likelihood information is not generated. This precludes any attempt to calculate model fit statistics such as R-squared and pseudo R-squareds. Twenty imputations were used for the descriptive and multivariate analyses (Royston, Carlin, & White, 2009). The final imputed dataset is composed of 6,970 cases, with 53.4% of cases being imputed on one or more measures in the perceived job insecurity models and 53.7% of cases being imputed on one or more measures in the perceived labor market insecurity models.³

Dependent Variables

The two outcomes that are the focus of this research are perceived job insecurity and perceived labor market insecurity. The measure of perceived job insecurity from the QWL module is captured using a simple but effective measurement that asks workers to provide a succinct assessment of their job security: “The job security is good—1 = Very true, 2 = Somewhat true, 3 = Not too true, and 4 = Not at all true” (Smith, Marsden, Hout, & Kim, 2017, p. 1530). Due to the small percentages in the third and fourth categories, I follow the precedent established by previous research (Fullerton & Anderson, 2013; Fullerton & Wallace, 2007) and combine them. The QWL measure for perceived labor market insecurity is measured with the following question:

How easy would it be for you to find a job with another employer with approximately the same income and fringe benefits as you have now?—1=Very easy to find similar job, 2=Somewhat easy to find similar job, or 3=Not easy at all to find similar job. (Smith et al., 2017, p. 437)

As noted earlier, the conceptual distinction between perceived job and labor market insecurity is significant in that one is asking respondents to assess their current job versus their current prospects. There has also been some debate as to how these concepts should be measured. Kalleberg (2011) notes that while often conflated with each other, the perceived job and labor market insecurity measures discussed earlier are capturing two separate dimensions of what he calls perceived employment insecurity. Schmidt (2000) uses the options from the perceived labor market insecurity measure to split the perceived job insecurity measure into two dichotomous variables, likely to lose job and costly job loss. Schmidt’s approach is problematic because it conflates the concepts that are actually being captured by these measures. While Kalleberg (2011) downplays this distinction and follows Schmidt’s (2000) precedent, I argue that ignoring this difference is not an ideal approach. Alternatively, Fullerton and Wallace (2007) use a reverse-coded version of the perceived job insecurity measure as an ordinal dependent variable and refer to it as perceived job security. Nevertheless, their framework does not allow for the inclusion of perceived labor market insecurity. Although perceived job and labor market insecurity are conceptually similar, they are distinct dimensions of the overarching concept of subjective economic insecurity, and they warrant independent treatment. Finally, there is an empirical basis for this distinction. The correlation between these outcomes is very weak, with a correlation coefficient of .082. In addition to the conceptual issues discussed earlier, this evidence provides justification for distinguishing between these two dimensions of perceived employment insecurity.

Table 1 contains the percentage distributions by year for both of the dependent variables. For the perceived job insecurity measure, a majority of the surveyed workers feel secure in their jobs (52%), with a much smaller percentage (15%) feeling very insecure in their jobs. For the perceived labor market insecurity measure, the modal response of the surveyed workers is that it would not be easy at all to find another job with approximately the same income and fringe benefits (40.6%), with a smaller but still sizeable percentage (21.6%) feeling that it would be very easy to find another job. Furthermore, the percentage of workers who experience high levels of perceived labor market insecurity rises sharply in 2010 compared with the other years. This indicates that the Great Recession may have been responsible for greater levels of economic anxiety for potential job seekers. A rough comparison of the most insecure categories of each outcome shows that more of the workers included in this sample feel more precarious about their prospects in the labor market than they do their current job insecurity. This suggests that there may be something distinct about the post-2000 period generally that is shaping these attitudes differently than in the past.

Table 1.

Distributions for Dependent Variables, GSS: Quality of Working Life Module, 2002, 2006, 2010, 2014.

Dependent variables	2002	2006	2010	2014	All years
Perceived job insecurity
The job security is good.
1 = Very true	53.8%	51.0%	48.6%	54.6%	52.0%
2 = Somewhat true	31.2%	33.6%	34.8%	33.9%	33.3%
3 = Not too true or not at all true	15.0%	15.4%	16.7%	11.5%	15.0%
	100%	100%	100%	100%	100%
Perceived labor market insecurity
How easy would it be for you to find a job with another employer with approximately the same income and fringe benefits as you have now?
1 = Very easy	24.1%	25.1%	15.2%	17.4%	21.6%
2 = Somewhat easy	35.8%	39.1%	34.1%	40.8%	37.8%
3 = Not easy at all	40.2%	35.8%	50.8%	41.7%	40.6%
	100%	100%	100%	100%	100%
Frequency totals	1,701	2,643	1,174	1,452	6,970

Note. Estimates are adjusted for complex survey design with Stata’s svy prefix. Percentages are weighted.

GSS = General Social Survey.

Independent Variables

There are several job characteristics that serve as key independent variables in the multivariate analysis. There are four measures of security and control. Government employee is a dummy variable, coded as 1 if the respondent is employed by the federal, state, or local government and 0 if they are not. Union member is a dummy variable, coded as 1 if the respondent belongs to a union and 0 if the respondent does not. Job tenure is measured as the number of years that the respondent has worked in their present job. Job autonomy is a scale comprising four variables from the QWL module (α = .71). The measures that compose the autonomy scale capture whether the respondent has a lot of freedom to decide how to do their job, if they have a lot of say in their work, how often the respondent takes part in decisions in their work environment, and how often they are able to set the way things are done. In addition to the factors of security and control, this analysis also accounts for two characteristics of uncertainty. Laid off is coded as 1 if the respondent was laid off from their main job at any time within the last year and 0 if not. Part-time is a dichotomous variable, coded 1 if the respondent works less than 35 hours per week and 0 if the respondent works 35 hours or more per week.

Each model in this analysis includes a set of dummy variables indicating the region in which the respondent lives (New England, Mid-Atlantic, South Atlantic, East North Central, West North Central, East South Central, West South Central, Mountain, and Pacific), with New England as the reference category. Time is captured using dummy variables for the years 2002, 2006, 2010, and 2014. Year 2002 is used as the reference category in the multivariate analyses. Urban resident is a dummy variable, coded as 1 if the respondent lives in the central city of a metropolitan area and 0 if they live in a suburb or rural area. Female is a dummy variable, coded as 1 if the respondent is female and 0 if they are male. Race has also been found to be related to perceived job insecurity (Wilson et al., 2006). Race is captured using the variable non-White, with White being the reference category. Age is measured as the respondent’s age in years. Because Fullerton and Wallace (2007) found that middle-aged workers tend to feel less secure about their jobs, age is included in quadratic form to account for its curvilinear relationship to the dependent variables. Married is a dichotomous variable for marital status, coded as 1 if the respondent is married and 0 if they are not. Years of education is measured as the number of years the respondent has attended school. Income (ln) is measured as the natural log of the respondent’s yearly gross income. The analysis also includes a series of dummy variables representing nine occupational categories: professional, managerial, sales, clerical, craft, operative, service, agriculture, and laborer, the last of which is the reference category. These measures are derived from the OCC80 and OCC10 variables in the GSS. Table 2 presents the means for the independent variables by year, adjusted by survey sampling design. The sample is 50.8% female, and 27.1% non-White. The mean age is 41.186, the average years of education is 13.552, and 33.7% of the workers are employed in a professional or managerial occupation.

Table 2.

Means for Independent Variables, GSS: Quality of Working Life Module, 2002, 2006, 2010, 2014.

	2002	2006	2010	2014	All years
Control variables
Urban resident	0.278	0.296	0.339	0.335	0.307
Female	0.515	0.495	0.525	0.509	0.508
Non-White	0.223	0.309	0.242	0.282	0.271
Age	40.074	40.879	42.217	42.207	41.186
Married	0.520	0.541	0.517	0.498	0.523
Years of education	13.510	13.383	13.693	13.796	13.552
Income (ln)	8.024	7.527	7.396	7.858	7.693
Occupation
Professional	0.200	0.203	0.199	0.243	0.210
Managerial	0.139	0.132	0.144	0.091	0.127
Sales	0.112	0.086	0.106	0.084	0.095
Clerical	0.130	0.132	0.131	0.138	0.133
Craft	0.095	0.106	0.078	0.070	0.091
Operative	0.099	0.100	0.080	0.102	0.097
Service	0.152	0.164	0.180	0.225	0.176
Agriculture	0.012	0.016	0.021	0.006	0.014
Laborer	0.044	0.044	0.024	0.029	0.037
Security and control
Government employee	0.170	0.195	0.170	0.194	0.184
Union member	0.147	0.132	0.095	0.125	0.128
Job tenure	6.435	6.660	6.796	7.574	6.819
Job autonomy (α = .71)	0.036	−0.063	−0.040	−0.032	−0.029
Uncertainty and conflict
Laid off	0.078	0.081	0.105	0.064	0.081
Part-time worker	0.206	0.184	0.217	0.186	0.195

Note. Adjusted by survey sampling design.

GSS = General Social Survey.

Method and Analytic Strategy

Because the outcome measures are ordinal, the multivariate analysis of this study uses generalized ordinal logistic regression to fit partial proportional odds models. The partial proportional odds model offers significant advantages over other methods such as ordinary least squares regression, the proportional odds model, and multinomial logit models. In regard to ordinary least squares regression, many of its assumptions (such as equal spacing between categories) do not hold for ordinal outcomes, making it an inappropriate tool for examining the dependent variables described earlier (Long, 1997; Winship & Mare, 1984). The proportional odds model, on the other hand, assumes an underlying latent variable and estimates logistic regressions at specific cut points in its distribution. The problem with this approach is that the parallel regression assumption, or the assumption that the coefficients are the same at each cut point, is often violated (Long, 1997). Lastly, although multinomial logit models provide a valid and thorough framework for analyzing ordinal outcomes, these models lack parsimony and are not readily interpretable. The partial proportional odds model offers an appropriate and easily understood framework for understanding ordinal outcomes that does not rely on often violated assumptions (Liu & Koirala, 2012; Williams, 2006). It does this by relaxing the parallel regression assumption, “… allowing the effect of each explanatory variable to vary across different cut points of the ordinal outcome variable without data restructuring” (Liu & Koirala, 2012, p. 243). This study uses Williams’s (2006) gologit2 command to identify any independent variables that violate the assumption and relax the parallel regression constraint for them, as shown by the Brant test.⁴

All of the models in this chapter are weighted using the WTSSALL sampling weight, which adjusts for selection of one adult per household and the subsampling of initial nonrespondents after 2004. Finally, this analysis uses the svy option of the gologit2 command (Williams, 2006) to use the survey design variables provided by the National Opinion Research Center to account for changes in the sampling design over time.

Findings

The multivariate analysis uses odds ratios calculated from the logistic regression coefficients for each model. The advantage of presenting odds ratios is that they are easily interpretable and provide a useful framework for comparing associations. Because both outcomes of this study have three categories, the result for each outcome is a table containing two estimations. The column on the left presents an estimation of the odds of being in the two higher categories of each outcome versus the lowest one. The column on the right presents an estimation of the odds of being in the highest category of each outcome versus the two lowest ones. The odds ratios in the right-hand column are presented only for those variables that violated the parallel regression assumption. The other variables that did not violate the assumption have a dash, as the odds ratios in the left-hand column are the same for both columns.

Perceived Job Insecurity

Table 3 presents the odds ratios generated from a generalized ordinal logit regression predicting perceived job insecurity. Being female has a significant negative effect, reducing the odds of feeling very insecure by 15.5%. Being non-White is significantly and positively associated with perceived job insecurity versus being White, while the relationship with age is curvilinear, indicating greater levels of insecurity for individuals in the middle of their careers. In addition, each year of education increases the odds of feeling very insecure by 3.1%. Notably, the time variables are not significant, failing to support H_1A. Being an urban resident, being married, income, and all of the occupation variables are also not significant.

Table 3.

Odds Ratios From Generalized Ordinal Logit Regressions for Perceived Job Insecurity, GSS: Quality of Working Life Module, 2002, 2006, 2010, 2014 (N = 6,970).

	(1) Somewhat true/ Not at all true vs. (0) Very true		(1) Not at all true vs. (0) Somewhat true/Very true
	exp(b)	SE	exp(b)	SE
Control variables
2006	1.050	(0.084)	–
2010	1.189	(0.111)	–
2014	0.943	(0.086)	–
Urban resident	0.977	(0.069)	–
Female	0.845*	(0.059)	–
Non-White	1.229**	(0.094)	–
Age	1.084***	(0.018)	–
Age squared^a	0.999***	(0.019)	–
Married	0.890	(0.059)	–
Years of education	1.031*	(0.015)	–
Income (ln)	0.997	(0.009)	0.973	(0.014)
Occupation^b
Professional	0.935	(0.169)	–
Managerial	1.001	(0.181)	–
Sales	0.950	(0.169)	–
Clerical	0.871	(0.148)	–
Craft	0.989	(0.172)	–
Operative	0.806	(0.156)	–
Service	0.741	(0.124)	–
Agriculture	0.687	(0.284)	–
Security and control
Government employee	0.655***	(0.059)	–
Union member	0.907	(0.112)	–
Job tenure	0.979***	(0.005)	–
Job autonomy (α = 0.71)	0.532***	(0.027)	0.504***	(0.030)
Uncertainty
Laid off	2.174***	(0.287)	–
Part-time worker	1.027	(0.106)	1.163	(0.149)
Constant	0.194		0.034

Note. Standard errors are in parentheses. Analyses are adjusted for sampling design. Multiple imputations for missing cases are used (m = 20). Controls for region are included in all models, but the coefficients are not shown. GSS = General Social Survey.

Standard error for age squared is multiplied by 100.

Laborer is the reference category.

p < .05. **p < .01. ***p < .001 (two-tailed).

On the whole, factors of security and control are associated with lower levels of perceived job insecurity. Being a government employee decreases the odds of feeling very insecure by 34.5%, which is consistent with the expectations of H_2A. However, being a union member is not significant, providing no evidence to support H_3A. For every year increase in job tenure, the odds of feeling very insecure decrease by 2.1%, supporting H_4A. Higher levels of job autonomy reduce the odds of feeling that it is somewhat true or not at all true that the job security is good by 46.8%. This negative and significant effect persists in the other model, lending support to H_5A. Unsurprisingly, there is support for H_6A, as being laid off from one’s job in the past year increases the odds of feeling very insecure by 117.4%. Finally, the odds ratio for being a part-time worker is not significant for perceived job insecurity, providing no support for H_7A despite the parallel regression constraint being relaxed for this variable.

Perceived Labor Market Insecurity

Table 4 presents the generalized ordinal logit odds ratios for the models predicting perceived labor market insecurity. Most of the time variables for this outcome are significant, supporting H_1B. Relative to year 2002, year 2006 is negative, and year 2010 is positive. Year 2014 is positive as well, but only for the respondents who felt that it would be not easy or somewhat easy to find another job. Respondents surveyed in 2014 had 47.3% higher odds of being in those two categories compared with respondents asked in 2002. Age is positive as well, indicating greater levels of labor market insecurity for older workers. The significant negative odds ratio for years of education persists in both models, contrary to the analysis of perceived job insecurity. Being an urban resident, being female, race, marital status, and income are all insignificant in this model.

Table 4.

Odds Ratios From Generalized Ordinal Logit Regressions for Perceived Labor Market Insecurity, GSS: Quality of Working Life Module, 2002, 2006, 2010, 2014 (N = 6,970).

	(1) Somewhat easy/ Not easy vs. (0) Very easy		(1) Not easy vs. (0) Somewhat easy/Very easy
	exp(b)	SE	exp(b)	SE
Control variables
2006	0.857*	(0.060)	–
2010	1.566***	(0.134)	–
2014	1.473***	(0.154)	1.089	(0.102)
Urban resident	0.955	(0.063)	–
Female	1.028	(0.068)	–
Non-White	1.007	(0.071)	–
Age	1.033*	(0.015)	–
Age squared^a	1.000	(0.017)	–
Married	0.940	(0.064)	–
Years of education	0.954**	(0.014)	–
Income (ln)	1.005	(0.008)	—
Occupation^b
Professional	0.613*	(0.116)	–
Managerial	1.030	(0.211)	0.734	(0.138)
Sales	0.688*	(0.122)	–
Clerical	0.845	(0.158)	–
Craft	0.642*	(0.122)	–
Operative	0.821	(0.159)	–
Service	0.554**	(0.099)	–
Agriculture	0.630	(0.203)	–
Security and control
Government employee	1.351**	(0.125)	–
Union member	1.090	(0.167)	1.554***	(0.178)
Job tenure	1.040***	(0.005)	–
Job autonomy (α = .71)	0.786***	(0.037)	–
Uncertainty
Laid off	1.474**	(0.202)	–
Part-time worker	0.768*	(0.078)	0.913	(0.088)
Constant	3.176		0.479

Standard error for age squared is multiplied by 100.

Laborer is the reference category.

p < .05. **p < .01. ***p < .001 (two-tailed).

The odds ratios for the occupation variables are much less monolithic than those for the perceived job insecurity models. Unsurprisingly, the professional, sales, craft, and service categories are negative and significant, meaning that the odds of feeling insecure about obtaining another job are lower for workers in these occupations than laborers.

Generally speaking, the factors of security and control are associated with a greater sense of labor market insecurity. Being a government employee increases the odds of feeling very insecure by a factor of 1.351, supporting H_2B. Being a union member increases the odds of feeling that it is not easy to get another job by a factor of 1.554, supporting H_3B. However, this effect is not significant for the model predicting the odds of being in the two highest categories. In addition, for every year increase in job tenure, the odds of feeling very insecure increase by a factor of 1.040, supporting H_4B. The exception to this trend of positive coefficients is job autonomy, whose negative effect supports H_5B. Unexpectedly, being laid off in the past year increases the odds of workers’ feelings of labor market insecurity by a factor of 1.474, providing no support for H_6B. On the other hand, being a part-time worker reduces perceived labor market insecurity by about 23.2%, supporting H_7B.

Discussion

The hypotheses and findings of the multivariate analysis are summarized in Table 5. The findings indicate that, within the context of the risk regime, there are several key divergences between perceived job and labor market insecurity. First, the impact of the Great Recession, the most seminal economic event of the 21st century so far, is different for each outcome. There is no support for H_1A, as workers were not more concerned about their job insecurity than before the Great Recession. This is despite unemployment levels still being high as a result of the downturn in 2010. However, this period did heighten workers’ concern about the labor market and the transferability of their skills, supporting H_1B. Because year 2006 is negative and significant and year 2010 is positive and significant for the perceived labor market insecurity models, it seems that the Great Recession had a major impact on workers’ feelings of insecurity about obtaining a position similar to their current one, which was declining prior to its onset. This suggests that workers will feel more vulnerable during times of economic crisis when faced with the possibility of seeking a new position.⁵ These findings speak to the distinctiveness of the Great Recession even in the uncertain context of the risk regime.

Table 5.

Hypotheses and Results of Analyses for Perceived Job and Labor Market Insecurity.

		Perceived job insecurity		Perceived labor market insecurity
Hypothesis	Variable	Expected	Result	Expected	Result
1	Great Recession	+	ns/ns	+	+ / +
2	Government employee	−	− / −	+	+ / +
3	Union member	+	ns / ns	+	ns / +
4	Job tenure	−	− / −	+	+ / +
5	Job autonomy	−	− / −	−	− / −
6	Laid off in past year	+	+ / +	−	+ / +
7	Part-time worker	+	ns / ns	−	− / ns

Note. + = positive coefficient, − = negative, ns = not significant. Results are from Tables 3 and 4.

These results present other interesting variations between the two dimensions of perceived employment insecurity, with the most conspicuous differences found between the odds ratios for government employee, union member, and job tenure. While these three factors ostensibly reflect security and control, they do not translate into a universal reduction in perceived insecurity in the risk regime. Being a government employee is negatively associated with perceived job insecurity but positively associated perceived labor market insecurity, supporting both H_2A and H_2B. The support for H_2A makes sense, given the perception that government employment comes with a great deal of job security. However, when government employees consider their prospects of finding an equivalent job elsewhere, the picture is much bleaker as they view their jobs as being good relative to others. It may be that the uncertainty of the risk regime contributes to a greater appreciation of their current position and heightened feelings of labor market insecurity, although the data cannot definitively show whether this is the case.

Being a union member has no significant association with perceived job insecurity, failing to support H_3A. This is likely related to the changing employment relationship that has characterized the risk regime, leading union members to have no greater feelings of job security than nonunion workers. However, it is positively associated with perceived labor market insecurity in the model for not easy versus somewhat/very easy in Table 4, providing limited support for H_3B. While unions are still able to afford some protections to workers who fall under their purview, unionization is in decline and workers recognize the difficulty in finding another job that offers those types of protections. Unions also educate their members about the nature of the job market and perhaps even use fears of job loss to stoke member interest and enthusiasm for union activities.

The findings also support both H_4A and H_4B, with job tenure representing a common thread that ties together these results. Government workers and union members tend to experience longer job tenures, which reflect a long-term accretion of firm- and job-specific skills and knowledge. This is both a strength and a vulnerability for workers. It strengthens the position of workers in the firm relative to those outside, reducing their perceived job insecurity. However, it also makes them vulnerable because these skills have a limited applicability to other organizations and positions, heightening their perceived labor market insecurity. This vulnerability is problematic for traditionally oriented workers who started working during the Fordist period and are now forced to navigate the risk regime with little to no skill or aptitude for doing so.

H_5A and H_5B are both supported, with the negative odds ratio for job autonomy persisting through all models for both outcomes. As noted earlier, workers with higher levels of job autonomy will likely have a great deal of technical expertise or a good working relationship with technical experts (Burris, 1993, 1998), which makes them highly employable and attractive to employers. Workers with greater autonomy are also likely to be very proficient at managing their careers even under the risk regime and therefore are less likely to feel insecure about finding another job. In addition, those who are more educated and employed in a professional occupation are more likely to have more job autonomy, which possibly bolsters this effect to some degree. Job autonomy will become increasingly important in the future as a mark of status and power in an occupational structure characterized by uncertainty and a technocratic system of control (Wallace & Brady, 2001).

The story for the factors of uncertainty is somewhat less compelling than that of the security and control factors, but not any less important. While there is support for H_6A, the model for perceived labor market insecurity shows a positive association with laid off that runs counter to expectations, failing to support H_6B. I had hypothesized that, because the workers in my sample were reemployed within a year of losing their job, their outlook concerning the prospect of finding a similar job would be more optimistic. However, it is possible that these workers found the process of finding new work to be particularly daunting under the risk regime and therefore feel very pessimistic about their chances of finding a similar position again.

Perhaps the most mysterious feature of the models for perceived job insecurity is the lack of support for H_7A. People working in part-time positions tend to be in the secondary labor market, where job insecurity is practically a feature of the position. Therefore, the lack of an association with perceived job insecurity is unexpected and difficult to explain.⁶ However, the negative odds ratio in the first perceived labor market insecurity model partially supports H_7B and is not so inexplicable. These jobs tend to be low-skill positions, and the reality of insecurity means that these workers have likely attuned their expectations accordingly and are confident in their ability to find a similar job.

One notable result that emerged from this analysis is the positive association between education and perceived job insecurity. While this seems to go against the conventional wisdom that higher levels of education should foster a greater sense of job security, education still plays a role in allaying workers’ fears in regard to labor market insecurity. Also, in contrast to the findings of previous longitudinal research, occupation does not influence perceived job insecurity. It is particularly telling that professionals’ feelings of job insecurity are not significantly different from those of laborers. The odds ratios of education, professional, and managerial in this model indicate that white-collar professions are no longer the buffer against feelings of job insecurity that they once were. This resonates with the findings of Jacobs and Newman (2008), who discovered that feelings of insecurity were becoming more pronounced among the so-called economic winners. However, the story is different in the models for perceived labor market insecurity. While white-collar workers feel no more secure than other workers in the course of their daily work experiences, they do have a greater sense of career security. These findings for education and occupation suggest that the changes brought about by the risk regime are uneven and support the idea that traditional markers of security may no longer offer the sort of protection that they once did in the Fordist period.

Conclusion

The purpose of this article was to highlight the primacy of perceived labor market insecurity and to use data from the QWL module of the GSS to determine whether the traditional factors of security and control continue to ease workers’ anxieties in the new risk regime, particularly after the Great Recession. My analysis reveals that while these factors are associated with reduced feelings of job insecurity, most of these once-coveted job characteristics actually exacerbate workers’ feelings of labor market insecurity, which I argue is the more important measure under the risk regime. That is, jobs that were previously considered to be “good” jobs such as government and union jobs may be more of a liability than an asset in the context of decreased employer–worker attachment. The only exception is job autonomy, which is universally associated with lower levels of perceived employment insecurity. When compared with previous longitudinal research, the protective functions of other factors such as occupation seem to have diminished as well. Regarding the Great Recession, it had no impact on workers’ perceived job insecurity, but it did increase their feelings of perceived labor market insecurity.

This study has serious implications for a society in which people’s economic and therefore physical well-being is tied to their employment situation, which is often subjected to the volatility of markets outside of workers’ immediate control. In the United States especially, risk has been shifted away from institutions in all spheres of society and toward individuals (Beck, 1992; Hacker, 2008; Smith, 2001). Namely, workers are individually responsible for financing higher education and many different types of job training, supplementing paltry social security benefits, and shopping for as well as purchasing health insurance if their employer does not offer coverage. Furthermore, these findings have implications for the relations between workers themselves, as well as for the relations between labor and capital more generally. Given the weakening attachment between workers and employers that is a defining feature of the risk regime, it has become more important for workers to cultivate networks outside of their firm. This has the potential to weaken workers’ social solidarity in the workplace, encouraging them to be more self-interested as they manage their own careers that are no longer entirely spent in a single firm. In the aggregate, this individualization of interest may also undermine the ability of labor to organize and resist the demands of capital. From a policy perspective, moving past established factors of security and trying to give all workers greater job autonomy will not address these problems, particularly in periods marked by economic disruption. Rather, the focus should be on the structural factors that have shaped this context of insecurity. The existence of a gig economy does not necessitate the existence of a risk regime, which undermines the protective functions of the traditional factors of security and control.

Although this research contributes to the literature in several ways, there are some limitations. First, although this study makes the case that the onset of the risk regime has profound implications for perceived employment insecurity, the limitations of the public use QWL module prevent this context from being adequately modeled. For example, information concerning respondents’ location beyond the region is not available. Future research should examine this context using a multilevel research design that accounts for structural factors such as the unemployment rate and deindustrialization. Second, this study does not fully explore differences in the outcomes between genders, racial groups, and age cohorts. Now that the importance of perceived labor market insecurity and the erosion of protections that might allay it have been established, future work should be devoted to these questions. Third, this study relies on cross-sectional data that does not capture changes in worker attitudes over the course of an individual’s career. This type of data would be very useful for understanding insecurity using a life course perspective, which is important given the significance of age in all of the models in this study’s analysis. Finally, this study examines outcomes that rely on singular measurements that may not fully capture the intricacies of perceived job and labor market insecurity. Social researchers should strive to formulate surveys with a more nuanced approach to measuring these dimensions. The meaning of what is generally called job security is changing. My research suggests that perceived labor market insecurity, while not necessarily displacing perceived job insecurity as a topic for investigation, is more relevant under the risk regime. As scholars continue to explore the impact that the risk regime is having on workers’ experiences and attitudes, perceived labor market insecurity is a dimension of perceived employment insecurity that demands deeper investigation.

Footnotes

Acknowledgments

The author would like to thank Michael Wallace for his helpful comments on previous versions of this article. The article also benefitted from the feedback of anonymous reviewers, for which the author is grateful.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Travis Scott Lowe is an assistant professor of Sociology at the University ot Tulsa. His research is situated within the areas of the sociology of work, urban sociology, and inequality. His current research examines perceived employment insecurity and urban racial inequality.

References

Allison

P. D.

(2000) Multiple imputation for missing data: A cautionary tale. Sociological Methods and Research 28(3): 301–309.

Anderson

C. J.

Pontusson

(2007) Workers, worries, and welfare states: Social protection and job insecurity in 15 OECD countries. European Journal of Political Research 46(2): 211–235.

Appelbaum

(2012) Reducing inequality and insecurity: Rethinking labor and employment policy for the 21st century. Work and Occupations 39(4): 311–320.

Arthur

M. B.

(2014) The boundaryless career at 20: Where do we stand, and where can we go? Career Development International 19(6): 627–640.

Arthur

M. B.

Rousseau

D. M.

(2001) The boundaryless career: A new employment principle for a new organizational era, New York, NY: Oxford University Press.

Artz

Kaya

(2014) Job insecurity and job satisfaction in the United States: The case of public sector union workers. Industrial Relations Journal 45(2): 103–120.

Beck

(1992) Risk society: Towards a new modernity, Newbury Park, CA: Sage Publications.

Beck

(2000) The brave new world of work, Malden, MA: Polity Press.

Bernhardt

Morris

Handcock

M. S.

Scott

M. A.

(2001) Divergent paths: Economic mobility in the new American labor market, New York, NY: Russell Sage Foundation.

10.

Brochu

Morin

L.-P.

(2012) Union membership and perceived job insecurity: Thirty years of evidence from the American general social survey. Industrial and Labor Relations Review 65(2): 263–285.

11.

Burgard

S. A.

Brand

J. E.

House

J. S.

(2009) Perceived job insecurity and worker health in the United States. Social Science & Medicine 69(5): 777–785.

12.

Burris

B. H.

(1993) Technocracy at work, Albany, NY: State University of New York Press.

13.

Burris

B. H.

(1998) Computerization of the workplace. Annual Review of Sociology 24: 141–157.

14.

Byron

R. A.

(2010) Discrimination, complexity, and the public/private sector question. Work and Occupations 37(4): 435–475.

15.

De Witte

Näswall

(2003) ‘Objective’ vs ‘subjective’ job insecurity: Consequences of temporary work for job satisfaction and organizational commitment in four European countries. Economic and Industrial Democracy 24(2): 149–188.

16.

Diebold

F. X.

Neumark

Polsky

(1997) Job stability in the United States. Journal of Labor Economics 15(2): 206–233.

17.

Dixon

J. C.

Fullerton

A. S.

Robertson

D. L.

(2013) Cross-national differences in workers’ perceived job, labour market, and employment insecurity in Europe: Empirical tests and theoretical extensions. European Sociological Review 29(5): 1053–1067.

18.

Dominitz

Manski

C. F.

(1997) Perceptions of economic insecurity. Public Opinion Quarterly 61(2): 261–287.

19.

Elman

O'Rand

A. M.

(2002) Perceived job insecurity and entry into work-related education and training among adult workers. Social Science Research 31(1): 49–76.

20.

Erlinghagen

(2008) Self-perceived job insecurity and social context: A multi-level analysis of 17 European countries. European Sociological Review 24(2): 183–197.

21.

Farber

H. S.

(1999) Alternative and part-time employment arrangements as a response to job loss. Journal of Labor Economics 17(S4): S142–S169.

22.

Farber

H. S.

(2008) Short(er) shrift: The decline in worker-firm attachment in the United States. In: Newman

K. S.

(ed.) Laid off, laid low: Political and economic consequences of employment insecurity, New York, NY: Columbia University Press, pp. 10–37.

23.

Farber

H. S.

Valletta

R. G.

(2015) Do extended unemployment benefits lengthen unemployment spells? Evidence from recent cycles in the U.S. labor market. Journal of Human Resources 50(4): 873–909.

24.

Fullerton

A. S.

Anderson

K. F.

(2013) The role of job insecurity in explanations of racial health inequalities. Sociological Forum 28(2): 308–325.

25.

Fullerton

A. S.

Wallace

(2007) Traversing the flexible turn: U.S. workers’ perceptions of job security, 1977–2002. Social Science Research 36(1): 201–221.

26.

Glavin

(2013) The impact of job insecurity and job degradation on the sense of personal control. Work and Occupations 40(2): 115–142.

27.

Hacker

J. S.

(2008) The great risk shift: The new economic insecurity and the decline of the American dream, New York: Oxford University Press.

28.

Heery, E., & Salmon, J. (Eds). (2000). The insecure workforce. New York, NY: Routledge.

29.

Jacobs

Newman

K. S.

(2008) Rising angst? Change and stability in perceptions of economic insecurity. In: Newman

K. S.

(ed.) Laid off, laid low: Political and economic consequences of employment insecurity, New York, NY: Columbia University Press, pp. 74–101.

30.

Kalleberg

A. L.

(2011) Good jobs, bad jobs: The rise of polarized and precarious employment systems in the United States, 1970s–2000s, New York, NY: Russell Sage Foundation.

31.

Kalleberg

A. L.

(2012) Job quality and precarious work: Clarifications, controversies, and challenges. Work and Occupations 39(4): 427–448.

32.

Kalleberg

A. L.

Marsden

P. V.

(2012) Labor force insecurity and U.S. work attitudes, 1970s–2006. In: Marsden

P. V.

(ed.) Social trends in American life: Findings from the general social survey since 1972, Princeton, NJ: Princeton University Press, pp. 315–337.

33.

Kalleberg

A. L.

Reskin

B. F.

Hudson

(2000) Bad Jobs in America: Standard and nonstandard employment relations and job quality in the United States. American Sociological Review 65: 256–278.

34.

Kalleberg, A. L., & Vallas, S. P. (2017). Probing precarious work: Theory, research, and politics. In A. L. Kalleberg & S. P. Vallas (Eds), Research in the sociology of work (Vol. 31), 1–30.

35.

Kim

T. J.

von dem Knesebeck

(2015) Is an insecure job better for health than having no job at all? A systematic review of studies investigating the health-related risks of both job insecurity and unemployment. BMC Public Health 15(985): 1–9.

36.

Klandermans

Hesselink

J. K.

van Vuuren

(2010) Employment status and job insecurity: On the subjective appraisal of an objective status. Economic and Industrial Democracy 31(4): 557–577.

37.

Laszlo

K. D.

Pikhart

Kopp

M. S.

Bobak

Pajak

Malyutina

Marmot

(2010) Job insecurity and health: A study of 16 European countries. Social Science & Medicine 70(6): 867–874.

38.

Lee

C. K.

Kofman

(2012) The politics of precarity: Views beyond the United States. Work and Occupations 39(4): 388–408.

39.

Liu

Koirala

(2012) Ordinal regression analysis: Using generalized ordinal logistic regression models to estimate educational data. Journal of Modern Applied Statistical Methods 11(1): 242–254.

40.

Long

J. S.

(1997) Regression models for categorical and limited dependent variables, Thousand Oaks, CA: Sage.

41.

Madrick

(2012) The deliberate low-wage, high-insecurity economic model. Work and Occupations 39(4): 321–330.

42.

Narisada

Schieman

(2016) Underpaid but satisfied: The protective functions of security. Work and Occupations 43(2): 215–255.

43.

Norris

D. R.

(2016) Job loss, identity, and mental health, New Brunswick, NJ: Rutgers University Press.

44.

Occhiuto

(2017) Investing in independent contract work: The significance of schedule control for taxi drivers. Work and Occupations 44(3): 268–295.

45.

Osnowitz

Henson

K. D.

(2016) Leveraging limits for contract professionals: Boundary work and control of working time. Work and Occupations 43(3): 326–360.

46.

Osterman

(1996) Broken ladders: Managerial careers in the new economy, New York, NY: Oxford University Press.

47.

Pedulla

D. S.

(2013) The hidden costs of contingency: Employers’ use of contingent workers and standard employees’ outcomes. Social Forces 92(2): 691–722.

48.

Polanyi

(1944) The great transformation: The political and economic origins of our time, Boston, MA: Beacon Press.

49.

Royston

Carlin

J. B.

White

I. R.

(2009) Multiple imputation of missing values: New features for mim. The Stata Journal 9(2): 252–264.

50.

Rubin

B. A.

(2014) Employment insecurity and the frayed American dream. Sociology Compass 8(9): 1083–1099.

51.

Rubin

D. B.

(1987) Multiple imputation for nonresponse in surveys, New York, NY: John Wiley & Sons.

52.

Rubin

D. B.

(1996) Multiple imputation after 18+ years. Journal of the American Statistical Association 91(434): 473–489.

53.

Schmidt

S. R.

(1999) Long-run trends in workers’ beliefs about their own job security: Evidence from the general social survey. Journal of Labor Economics 17(S4): S127–S141.

54.

Schmidt

S. R.

(2000) Job security beliefs in the general social survey: Evidence on the long-run trends and comparability with other surveys. In: Neumark

(ed.) On the job: Is long-term employment a thing of the past?, New York, NY: Russell Sage Foundation, pp. 300–331.

55.

Schreurs

van Emmerik

Notelaers

De Witte

(2010) Job insecurity and employee health: The buffering potential of job control and job self-efficacy. Work & Stress 24(1): 56–72.

56.

Smith

T. W.

Marsden

Hout

Kim

(2017) General social surveys, 1972–2016, Chicago, IL: National Opinion Research Center.

57.

Smith

(2001) Crossing the great divide: Worker risk and opportunity in the new economy, Ithaca, NY: Cornell University Press.

58.

Standing

(2002) Globalisation: The eight crises of social protection. Indian Journal of Labour Economics 45(1): 17–46.

59.

Swinnerton

K. A.

Wial

(1995) Is job stability declining in the U.S. economy? Industrial and Labor Relations Review 48(2): 293–304.

60.

Tremblay

D.-G.

(2008) From casual work to economic security: The paradoxical case of self-employment. Social Indicators Research 88(1): 115–130.

61.

Vallas

Christin

(2018) Work and identity in an era of precarious employment: How workers respond to “personal branding” discourse. Work and Occupations 45(1): 3–37.

62.

Vallas

Prener

(2012) Dualism, job polarization, and the social construction of precarious work. Work and Occupations 39: 331–353.

63.

Wallace

Brady

(2001) The next long swing: Spatialization, technocratic control, and the restructuring of work at the turn of the century. In: Kalleberg

A. L.

Berg

(eds) Sourcebook on labor markets: Evolving structures and processes, New York, NY: Plenum Press, pp. 101–133.

64.

Wallace

Brady

(2010) Globalization or spatialization? The worldwide spatial restructuring of the labor process. In: McDonough

Reich

Kotz

D. M.

(eds) Contemporary capitalism and its crises: Social structure of accumulation theory for the 21st century, New York, NY: Cambridge University Press, pp. 121–144.

65.

Williams

(2006) Generalized ordered logit/partial proportional odds models for ordinal dependent variables. The Stata Journal 6(1): 58–82.

66.

Wilson

Eitle

T. M.

Bishin

(2006) The determinants of racial disparities in perceived job insecurity: A test of three perspectives. Sociological Inquiry 76(2): 210–230.

67.

Wilson

Mossakowski

(2009) Fear of job loss: Racial/ethnic differences in privileged occupations. DuBois Review 6(2): 357–374.

68.

Wilson

Mossakowski

(2012) Job authority and perceptions of job security: The nexus by race among men. American Behavioral Scientist 56(11): 1509–1524.

69.

Wilson

Roscigno

V. J.

(2016) Public sector reform and racial occupational mobility. Work and Occupations 43(3): 259–293.

70.

Winship

Mare

R. D.

(1984) Regression models with ordinal variables. American Sociological Review 49(4): 512–525.