Beneath the glass ceiling: Explaining gendered role segmentation in call centres

Abstract

Although the call centre is much researched, the literature on gender remains surprisingly undeveloped given the importance of this setting for women’s employment. This study of role segmentation in four call centres demonstrates women’s disproportionate representation in more routinized mass production roles, as opposed to higher status or managerial grades. It also analyses three explanations – human capital, domestic status and supervisor career support. The evidence shows that women face a ‘glass ceiling’, first, on entry to the call centre in terms of human capital disadvantage and levels of domestic constraint and, second, within the call centre in their ability to secure supervisor support for career opportunities. We argue that even for women with similar career aspiration and human capital to men, domestic responsibilities create obstacles before they reach the glass ceiling, especially for managerial roles, and contribute thereafter to reinforcing their concentration in more intensive, lower status work.

Keywords

call centres careers gender glass ceiling gender in organizations

Introduction

This article analyses the predictors of gendered work role segmentation in call centres. The call centre is an important organizational setting, for despite being a major locus of female employment, men appear over-represented in higher-level positions. Through a study of four call centres, we examine the influence of human capital, domestic status and supervisor career support upon the roles occupied by men and women.

Although the call centre has been the subject of extensive study, both as a phenomenon in its own right and as a research site for testing diverse theories (Deery and Kinnie, 2004), the literature on gender remains surprisingly limited (Russell, 2008) and early promising studies (Belt, 2002; Belt et al., 2000, 2002) have not been built upon. Many questions remain unanswered leaving contradictory representations of the opportunities call centres provide for women. Positively, they are held to offer flexible employment and career possibilities in ‘female-friendly’ workplaces (Russell, 2008); to be the site of semi-professionalized, knowledge work (Frenkel et al., 1999); or the information economy’s version of semi-skilled clerical and service work (Russell, 2008). A less optimistic portrayal is the call centre as ‘female ghetto’ (Belt, 2002), of demanding service work, flat career structures, and glass ceilings that limit progression (Belt et al., 2002; Durbin, 2006). Men are likely to occupy higher status technical or business-related positions, while women dominate the more ‘socially skilled’, yet more intensive and routinized customer service roles (Lloyd and Payne, 2009). Such characteristics suggest that call centres represent a continuation of feminized, low-status clerical labour (Crompton and Jones, 1984).

This article provides an analytical focus and empirical substance lacking in the limited call centre gender literature. Several major difficulties are manifest in these studies (Belt, 2002; Belt et al., 2000, 2002; Durbin, 2006), not least that they are essentially descriptive accounts and, often based on selective qualitative data, frequently speculative in their conclusions. Belt’s (2002: 64) self-critical observation that she had been able only to ‘scrape the surface’ in explaining women’s career progression is apposite. More substantively, the focus is almost entirely on managerial practices and women’s experiences at workplace level, largely in relation to recruitment, skill utilization, career opportunities and gender stereotyping. The studies implicitly assume that the workplace by itself can explain gendered segmentation and women’s relative disadvantage. Of course, organizational and worksite dimensions are important but, we argue, women’s treatment in employment must be analytically integrated with their human capital and domestic lives.

An additional criticism is that these studies make the mistake, common to much of the early call centre literature (see Taylor et al., 2002), of treating call centres and their workflows as if they were homogeneous. Belt and Durbin understate the fact that the numerically dominant Customer Service Representative (CSR) roles are variegated according to diverse characteristics, including the degree of standardization or customization of customer interaction, the simplicity or complexity of calls, call length, the amount of remuneration, the extent of training and discretion over task performance. Gendered dimensions may indeed map onto, or underpin, such differentiation, with women concentrated in particular CSR roles.¹ Thus, the existing literature, in concentrating on career, focuses exclusively on progression from CSR to team leader and to managerial level. Although undoubtedly important, this preoccupation with sequential progression neglects the significance of segmentation within the ranks of the CSRs for a gendered analysis. This weakness is addressed methodologically, empirically and analytically in this study.

Beyond the call centre, this study contributes more generally to debates on women’s careers and gender-based job segmentation. First, despite substantial research on the glass ceiling to women’s advancement within professions and senior executive levels (Lyness and Thompson, 1997; Tharenou, 1999, 2001), far fewer studies examine women’s work in contemporary service occupations (Kerfoot and Korczynski, 2005). By revealing the ways in which women’s success (or lack thereof) continues to depend on different factors from men, the present study contributes theoretically to understanding the persistence of women’s disadvantage in newer, more flexible, and numerically female-dominated organizations.

A second contribution is to the debate about whether gendered job segregation is determined by individual choice, or more structural features of the labour market, household division of labour or organizations. The call centre provides a prism for examining the claim that for some women, work is desirable as a flexible, low commitment, secondary source of household income, and that differences in preference may explain both vertical (status) and horizontal (job role) segregation (Hakim, 2000). By examining the causes of segmentation for men and women within a single (albeit differentiated) occupation, this study provides evidence in a controlled setting of the relative contributions of choice and structural factors affecting women’s advancement.

The article begins with a categorization of call centre work that addresses the lack of differentiation in previous literature and its potential implications for understanding gender segmentation. It then evaluates the contributions of existing call centre and gender research in order to identify predictors relevant to advancement. In so doing, we construct hypotheses that are tested through a survey of four call centres. Following the presentation of findings, we return to the theoretical contributions of the study for understanding gender equality in careers within call centre and service work more generally.

Categorizing call centre work

As a unique hybrid of clerical/office and interactive service work (Boreham et al., 2008), the call centre has provided a lean, spatially flexible and low-cost means of delivering high-volume customer contact, but the work is not homogeneous. For Frenkel et al. (1999) and, more explicitly, Korczynski (2002) it is the tension between the twin logics of cost-efficiency and customer orientation that gave rise to the call centre’s hybrid customer-orientated bureaucracy formation. Batt and Moynihan (2002) provided a sophisticated attempt to understand variation, relating product market segment to organizational strategy, HR policy and degree of worker discretion. The outcome was their distinction between three call centre models: mass production (MP), mass customization (MC) and professional services.

Jobs may be distinguished further by the extent to which they are characterized by volume and value – or quantity and quality (Taylor et al., 2002) – the outcome of job design characteristics and managerial priorities. Contingent factors impacting such variability include firm/centre size, industry/sector, market conditions, and whether services are inbound, outbound, outsourced or in-house. At one extreme are the most volume-driven mass market operations, where call throughput is prioritized, call times are minimal and customer contact is most standardized and ‘transactional’, while at the opposite extreme, quality considerations predominate, call times are extended and customer interaction is more complex and relational. Some centres do deliver professional services (Collin-Jacques and Smith, 2005) and bespoke higher value or more technical customer interaction. However, potential variability does not mean equivalence for, as Lloyd and Payne (2009: 620) remind us, the high volume, MP model prevails.

Consequently, the call centre labour market for women has been dominated by jobs within this MP category, but Taylor and Anderson (2008) have suggested sexual divisions of labour within call centre work, which reflects broader employment trends (Webb, 2010). According to their survey, males were concentrated (69%) in IT technical help-desk work and in higher-end financial services such as stockbroking (59%), while in the numerically dominant basic customer service, information provision and sales roles females were most disproportionately represented (80%). Notable female concentrations were in travel, transport and holiday operations (70%) and local government (79%).

Informed by these heuristic constructs and by empirically-established differences, the present study follows Batt and Moynihan (2002) in separating work roles according to mass production and mass customization. The former is characteristic of lower status, routinized work roles, and the latter consists of roles in which greater customization, complexity and variety inform customer encounters. A third category comprises team leaders and managers. Although women may form a majority in all three, we propose that men are disproportionately represented in MC and team leader/manager roles and women in MP roles. Given the limitations of existing call centre gender literature, which neglects role segmentation, we present the first hypothesis:

Hypothesis 1 (H1): Gender predicts type of work role, such that being female increases the likelihood of being in a mass production role, as opposed to either mass customization or team leader/manager roles.

Explaining gender segmentation

Although service work attracts significant proportions of women, it remains unclear whether lack of progress within these occupations is attributable to women’s choices, human capital or more particular barriers either within or outside the workplace that interact to prohibit access to career opportunities. The present study focuses on three potential explanations that have been linked to women’s careers (Judge et al., 1995; Melamed, 1995; Metz and Tharenou, 2001): human capital, domestic status and, in the workplace, supervisor support for development. How each of these is dealt with in gender call centre research provides a point of departure for the broader literature and for the construction of appropriate hypotheses. First, we present the case for each as factors that differentiate the male and female workforce and have been shown to be linked to career success. In particular, we suggest that women are disadvantaged at entry to call centres by their overall lower levels of human capital and greater domestic responsibilities, as well as by the effects of gender role stereotyping once employed. Second, building on evidence relating women’s human capital disadvantage to domestic status, we propose that these factors play out differently in the workplace for women and men to influence both the work roles they occupy and the degree of supervisor career support they are offered. These hypotheses are diagrammatically represented in Figure 1 and explicated below.

Human capital

The call centre gender literature considers the influence of human capital only from the perspective of recruitment and skills utilization and not from ‘supply-side’ characteristics. During recruitment, employers emphasize social and communications skills through literacy tests or competency-based interviews (Belt, 2002). However, sectoral expansion has meant competition for transferable service and IT skills, which places a premium on certain kinds of work experience. Firm-specific product or information systems knowledge may provide an internal ladder towards higher paid and more technical positions (Batt et al., 2002; Sieben and de Grip, 2004). These considerations apart, the existing literature (Belt, 2002; Belt et al., 2000, 2002; Durbin, 2006) contains no systematic evidence of how human capital inputs relate to work roles and how they may differ by gender or lead to gender segmentation. Promotion criteria are ambiguous. Longer tenure may indicate better recognition of company-specific knowledge and experience and increase promotion chances through internal labour markets (Bielby and Baron, 1984), but employers may, equally, disregard ‘career maturity’ when faced with cost-cutting pressures (Golsch, 2007).

More generally, ‘supply-side’ explanations locate gender career differences in women’s own decisions about education and employment. Human capital theory emphasizes the rational choices made by women whether to invest in formal qualifications and skill development (Becker, 1985) since, spending less time in the labour force, women have less incentive to invest in their marketable capital (Warren et al., 2001). Men’s progress at managerial levels benefits from greater mobility across jobs and within organizations (Stewart and Gudykunst, 1982), while women’s careers are more likely to be interrupted with periods of non-wage and non-standard (e.g. flexible hours) work that interfere with accumulated service. These factors, in turn, are related to fewer promotions (Zeytinoglu and Cooke, 2008) and segregation into low-wage jobs (Tomlinson et al., 2009). Thus, as human capital is related to career success (e.g. Becker, 1993; Metz and Tharenou, 2001; Ng et al., 2005; Tharenou et al., 1994; Wayne et al., 1999), men’s over-representation in senior positions results from greater accumulation of these inputs.

This study examines two indicators of human capital: educational qualifications and organizational tenure. We consider whether women’s lower qualifications and tenure (a consequence of more frequent career interruptions) explain their segmentation into MP roles. It is also possible that women have longer average tenure than men because of lower overall job mobility (Alvesson and Billing, 1997), but we frame the hypothesis in terms of women’s lower human capital in general, as we propose its mediation of the relationship between gender and work role. Thus, we hypothesize:

Hypothesis 2 (H2): (a) Women have less human capital than men; and (b) less human capital increases the likelihood of being in mass production as opposed to either mass customization or team leader roles.

Domestic responsibility

The call centre gender literature does gesture towards the role of domestic and household circumstances, but references are scattered and observations speculative, so that the influence of such circumstances remains unincorporated. For example, Belt (2002: 54) cites a management consultant who referred to a substantial ‘female underclass’ in call centres composed of working-class women ‘with childcare commitments remaining solidly at the bottom of career ladders’.

Belt further suggests that call centres provide the flexibility to accommodate complex household dynamics and work timings, implying, consistent with preference theory (Hakim, 2000), that women with greater domestic responsibilities prefer low responsibility call centre work. The household division of labour is a critical factor in determining these choices. Longitudinal studies of family time-use and labour market participation over the life-course (Gershuny, 2004; Scott, 2010) reveal that women lose their bargaining power relative to men in human capital accumulation by specializing in non-waged work or entering non-standard employment (e.g. reduced hours), sometimes involuntarily (Nardone, 1995; Walsh, 1999). In addition to reducing women’s labour market competitiveness, men are likely to remain primary wage earners and women to base further decisions about paid work around family priorities or partners’ pursuit of greater earnings or promotion (Powell and Greenhaus, 2010).

Notwithstanding scattered insights, there is no systematic examination of gender role segmentation in call centres that incorporates household and domestic factors with human capital influences. In attempting to overcome this weakness, the effects of domestic status are examined through a measure of family and household structure, including marital status, dependent children/care responsibilities, and degree of responsibility for caring. We express the hypothesis again in terms of mediation, such that women are expected to have more domestic responsibilities than men, which in turn is related to role segmentation:

Hypothesis 3 (H3): (a) Women report more domestic responsibilities than men; and (b) more domestic responsibilities increase the likelihood of being in mass production as opposed to either mass customization or team leader roles.

Supervisor career support

Regardless of individuals’ gender, human capital or domestic status, the structural characteristics of call centres generate problems for career progression. Flat hierarchies create job queuing, particularly in the absence of internal job hierarchies that are mainly confined to larger centres or organizations (Gorjup et al., 2008). For women, though, additional barriers may emerge because of the absence of formalized promotion procedures, which are argued to increase the likelihood of managerial stereotyping (Bielby, 2000). The few studies to examine progression in call centres highlight a lack of agreement about which skills are important and should be rewarded and indicate wide variation in formal Human Resource Management (HRM) implementation (Belt et al., 2002; Lloyd and Payne, 2009; Russell, 2009). Social competence, understood as encompassing cognitive, interpersonal, emotion management and self-mastery skills, and especially suited to women (Lloyd and Payne, 2009), often remains ‘invisible’, not formally appraised or matched by rewards or job promotions (Belt et al., 2002; Russell, 2009). While women have achieved some success in progressing to people management, team leader roles (Belt, 2002; Durbin, 2006), gender role stereotyping is evident, with Belt (2002) citing female agents who claimed their managers considered them lacking the more ‘masculine’ skills required for strategic managerial positions.

Despite concentrations of women in middle management, evidence suggests subtle stereotyping that allows men easier access to necessary human and social capital for internal advancement. Selection for training beyond basic product and systems knowledge seems based on managers’ stereotypical notions of women’s suitability (Belt, 2002; Belt et al., 2002). Wider cultural cues may reinforce stereotypes. For example, the absence of female role models at senior levels (Durbin, 2006), conceivably, perpetuates the effects of male managerial hierarchies, such as gender-similarity effects in selection and promotion (Cohen et al., 1998; Tharenou, 2001) and curbs women’s aspirations because of lack of congruence with managers (Litzky and Greenhaus, 2007). Men are also more likely than women to identify lack of advancement opportunity as a reason for exit (Russell, 2009) suggesting a more easily appeased female workforce with respect to development opportunities.

Career mentoring by supervisors can provide access to the skill development necessary for promotion (Wayne et al., 1999) but women are generally shown to be less effective at accessing this kind of social capital (Tharenou et al., 1994). This disadvantage has also been shown to exist in female-dominated contexts. In particular, the ‘glass escalator’ effect (Hultin, 2003; Maume, 1999) describes men’s accelerated career paths in traditionally ‘female’ occupations, such as teaching (Williams, 1992), based on ‘grooming’ by supervisors. This has not been directly tested in call centres. Thus, we propose that women will experience less supervisory career support than men, and this helps explain their role segmentation:

Hypothesis 4 (H4): (a) Women experience lower levels of supervisory career support than men; and (b) supervisory career support increases the likelihood of being in mass customization or team leader rather than mass production roles.

Interaction of gender, human capital and domestic status

The argument thus far is that gendered role segmentation is explicable by the profile of women attracted to and employed by call centres (women are generally less qualified and have less firm-specific experience than men as well as greater domestic responsibilities) and by their experience of career support once employed. We add the proposal that these initial human capital and domestic conditions present barriers for women that do not exist for men. Generally, women receive fewer returns on their human capital than men, which leave them in jobs that underutilize their potential (Tomlinson et al., 2009). Women with high levels of human capital may be less likely to advance than equivalent men. One reason is the unequal effect of domestic life, several aspects of which can interfere with women’s, but not men’s career paths once employed. Family structure (e.g. being married or having domiciled children) is related to advancement or career continuity for both sexes (Stroh et al., 1992; Tharenou, 1999) but more negatively for women. Having a spouse and children has no impact on, or may even benefit, men’s careers (Kirchmeyer, 2006; Landau and Arthur, 1992), while women who advance are less likely to be married or have children than men who advance (Tharenou and Conroy, 1994). Women also continue to bear a greater burden of domestic labour, irrespective of employment status or educational level (Breen and Cooke, 2005; Dex, 2003) compared with men in similar domestic circumstances.

These findings suggest that gender moderates the relationships between human capital, domestic status and work role proposed in Hypotheses 2 and 3. We refine these, therefore, to propose that the relationship between high human capital and work role will be stronger for men than women, and that high domestic responsibility will negatively affect women’s, but not men’s chances of rising out of MP roles. We also examine the three-way interaction between these variables to test whether domestic responsibility is more likely to have a limiting effect on women’s human capital than men’s. These gender moderation effects are expressed in the following hypotheses:

Hypothesis 5 (H5): Women with high human capital (education, tenure) are less likely to be in mass customization or team leader roles than men with high human capital.

Hypothesis 6 (H6): Domestic responsibility reduces the likelihood of being in mass customization or team leader/manager roles for women but not men.

Hypothesis 7 (H7): For women the positive effects of education and tenure on the likelihood of being in mass customization or team leader roles will be weakened by domestic responsibilities.

Our final hypothesis attributes women’s lower levels of supervisory career support to domestic status. Durbin (2006: 131) tentatively suggests that caring responsibilities impinge on women’s ability to take external courses and work the long hours that will enable them to be employed in higher status roles or to progress within the call centre. Linking this to managerial stereotyping of women, the job-matching model (Barron et al., 1989, 1993) explains women’s lower earning power and training disadvantage by employers’ reluctance to invest in women’s development given the perceived likelihood of them leaving the organization (Lazear and Rosen, 1990). Knote and Ishio’s (1998) longitudinal event-history analysis showed that women experience a training disadvantage that widens following entry regardless of human capital, occupation and organization. This highlights the effects of home status and family stage in women’s ability to pursue opportunities at work; for example, women with dependent children will experience more barriers in accessing opportunities than men with dependent children (Metz, 2005). Consistent with a job matching explanation of gender role stereotyping, we propose that women with domestic responsibilities will have lower supervisory career support than men in comparable situations:

Hypothesis 8 (H8): Gender moderates the relationship between domestic responsibility and supervisory career support, such that for women but not men, greater domestic responsibilities reduce supervisory career support.

Control variables

The study controlled for career aspiration to account, first, for the possibility that role segmentation reflects women’s personal preferences for low commitment work (Belt, 2002) and, second, for ‘recruitment effects’ created by employers targeting women with fewer career aspirations, such as ‘mature returnees’ (Callaghan and Thompson, 2002). We controlled for age, given its importance for career success (Van der Heijden et al., 2009), and for time in current work role in order to counteract bias inherent in presenting prior or current workplace experiences as a predictor of current work role. This allows greater confidence in the hypothesized relationships if they apply to recent as well as more experienced role-occupants. Differences between call centres were considered when examining the effects of workplace practices, particularly supervisory career support.

Method

Data collection

The data come from four large call centre workplaces (200+ employees) in four industries/sectors: financial services (Moneyflow), outsourced services (Thejobshop), telecoms/entertainment (Entcomm) and holiday/travel (Holstravel). Data gathering took place during 2002. Qualitative data, which provided evidence of the HR policies and practices and roles within each centre, came from 251 exploratory interviews with managers and employees and 77 semi-structured interviews with employees across gender, age, job type and level. Centres were comparable in terms of HR policies (e.g. Performance Development and Reward systems) and job types (Table 1). Following qualitative data gathering, 1476 questionnaires were distributed to all CSRs, including team leaders but excluding back office and support staff, across the four centres with sealable envelopes to be returned to designated post boxes. Anonymity was ensured and socially desirable responses minimized. Face-to-face distribution and follow-up resulted in a high average response rate of 60 percent (855 questionnaires) and ensured representation across shifts, demographics and sections. Missing gender data left 847 respondents (252 men, 595 women).

Table 1

Data triangulation for the distinction between MP workflows, MC workflows and team leaders/managers

Company-Section	Call centre data				Survey data
	Starting pay^a	Initial training	Average call times	Attrition (%)^b	N^c	Job control		Women (%)
						Mean	(SD)
Mass production
M-Unregulated Loans	£9950	4 weeks	5 min 30 sec	50	41	2.24	(.77)	76
T-Energyco	£11,000	2 weeks	3 min 40 sec	47	39	2.18	(.95)	62
T-Gen Business	£10,000	2 weeks	< 3min	44	61	2.41	(.99)	68
E-Service/Installation	£9500	4 weeks	3 min 40 sec	>centre ave.	158	2.17	(.89)	76
E-First Line Retention	£9500	4 weeks	3 min 40 sec	>centre ave.	8	1.59	(.42)	87
E-Moves and Transfers	£10,750	4 weeks	3 min 30 sec	=centre ave.	17	1.81	(.60)	82
E-2^nd Line Retention	£10,750	5 weeks	4 min	=centre ave.	14	1.95	(.94)	71
H-Teletext	£10,200	3 weeks	3 min 30 sec	59	91	2.21	(.71)	77
H-Main Reservations	£10,800	4 weeks	5 min 45 sec	49	92	2.25	(.75)	85
Survey total/means					521	2.09	(.78)	76
Mass customization
M-Regulated Loans	£13,620	6 months	None	12	57	2.77	(.89)	77
T-Bluechip	£14,000	6 weeks	6–30 min	9–11	54	2.70	(1.09)	46
T-Gamesco	£13,500	2 months	5–60 min	10	35	3.05	(.93)	52
T-Carexec	£14,500	2 months	4 min	9	25	2.75	(.64)	60
E-Service Performance	£12,670	Experienced staff + 4 wks	None	<centre ave.	19	4.08	(.14)	33
H-Complaints	£12,500	Experienced staff + 2 wks	5–30 min	15	29	2.51	(.81)	75
Survey total/means					219	2.98	(.76)	57
Team leaders/Managers
Moneyflow	N/A	N/A	N/A	N/A	14	3.84	(.73)	55
Thejobshop	N/A	N/A	N/A	N/A	35	4.01	(.84)	48
Entcomm	N/A	N/A	N/A	N/A	28	3.62	(.68)	64
Holstravel	N/A	N/A	N/A	N/A	30	4.22	(.57)	80
Survey total/means					107	3.93	(.71)	61

Notes: N/A denotes Not Available; ave. denotes average; M Moneyflow, T Thejobshop, E Entcomm, H Holstravel

Base starting salaries are the simplest comparative measurement for CSR roles, notwithstanding salary ranges, which involve incremental increases for experience and length of service. Evidence comes from diverse sources: Incomes Data Services (IDS, 2000, 2001, 2002), company sources and management reports, and confirmation through employee interviews.

Senior and HR management supplied mean annual company-wide attrition levels for Moneyflow (34%), Entcomm (34%), Holstravel (47%) and Thejobshop (26%). Detailed figures for each section/workflow are available for Moneyflow, Thejobshop and Holstravel. In Entcomm, managers were only prepared to report whether levels were greater or smaller than the centre average.

Total survey respondents N = 847.

Measures

Control variables

In addition to gender (women coded 1) and call centre, we measured age (1 = 16–30, 2 = 31–40, 3 = 41 and over), time in current role (months), and career aspiration with regard to call centres at employment commencement. The career aspiration variable was drawn from open-ended responses to reasons for present job choice coded as a ‘career decision’ (e.g. advancement, challenge, interest) (1) or ‘decision based on convenience/non-career reasons’ (e.g. location, better pay/conditions, first one offered) (0). The Scottish call centre workforce is homogeneous with respect to other potentially influential demographics (e.g. ethnicity) (Taylor and Anderson, 2008).

Human capital

Highest educational qualification was scaled according to Scottish qualifications frameworks (1 = less than high-school; 2 = high-school and equivalent, 3 = degree, either undergraduate or postgraduate). Current tenure was measured in months.

Domestic responsibility

Household status and responsibility were captured in a single scale based on that of Orser and Leck (2010). We added whether respondents bore primary responsibility for caring as a domestic restriction on labour market participation. Childcare was differentiated by age groups for pre-school and school-age children. The scale was constructed by summing seven dichotomous variables each measured as 0/1 with a minimum score of 1 indicating no variable was present and a higher score greater domestic responsibility: presence of spouse/partner; children (age 0–4); children (age 5–11); children (age 12–17); other care responsibilities; primary responsibility for care of young children; and primary responsibility for other care.

Supervisory career support

Degree of career support from immediate supervisor was measured as the mean of five items appropriate for call centre work based on Greenhaus et al.’s (1990) perceived supervisory support scale (e.g. ‘My immediate boss gives me helpful feedback about my performance/provides training opportunities which will allow me to get a better job in this company’; five-point scale of agreement; α = .82). An additional question on training activity ensured meaningful questions about supervisor support. There were no significant differences across call centres for agents − 75 percent or more received at least on-the-job training (χ²(3) = 4.93) - or for team leader/managers, of whom over 56 percent reported off-the-job training (χ²(3) = 3.71).

Dependent variable

Each respondent was classified according to one of three roles: mass production (MP) (coded 1), mass customization (MC) (coded 2), or team leader/manager (TL/M) (coded 3). While TL/Ms were identified by responses to an open-ended request for job title, the other two were achieved by triangulating multiple sources of evidence to differentiate between MP and MC workflows. The descriptor ‘workflow’ derives from Frenkel et al.’s. (1999: 36) definition as ‘a structured set of tasks (work) leading to a specified output … oriented towards a particular market’, which helpfully acknowledges that a call centre might contain activities of differing complexity, contrasting prioritization of quantity and quality, or of standardization and customization. Thus, a single site might contain both MP and MC workflows and commensurate work roles.² The most significant differentiation may lie not so much between entire centres, but between distinct workflows, so that discrete activities in a particular centre may correspond more to those in another centre rather than with others in their own workplace.

Data sources for triangulation (Table 1) included, first, two HR measures, starting pay and length of initial training. In all instances, pay was lower for new starts for MP roles. Similar distinctions emerge for training duration. At E-Service Performance and H-Complaints Handling, staff were exclusively recruited internally and underwent initial and developmental training, so that the ‘initial’ training cited for these MC roles was additional. Second, average call times are indicative of the complexity of customer encounters. Care is taken, though, not to conflate complexity with skill as the latter concept is much contested (Lloyd and Payne, 2009). What stands out is the short average handling times (AHTs) of the MP relative to MC workflows. In M-Regulated Loans and E-Service Performance, no call times were prescribed given potentially extended encounters. On two Thejobshop contracts (Bluechip and Gamesco) and at H-Complaints Handling, calls varied from simple resolution to protracted problem solving. Third, attrition rates, long acknowledged (Russell, 2008) as indicative of, inter alia, the extent of task pressure and work intensity, levels of remuneration and job satisfaction, provide evidence of differentiation. Insofar as data permit, and having standardized the measurement criteria, centre-wide annual percentages disaggregated by workflow indicate real differences between MP and MC roles.

Finally, survey data indicate employees’ perceived job control, the number of survey respondents and the proportion of women in each section. Job control was measured using the mean of four items consistent with Sprigg et al.’s (2007) measure for call centres (five-point scale from 1 ‘not at all’ to 5 ‘a great deal’; α = .78). In all cases, control is greater for MC roles and there is a statistically significant difference between mean control for MC (M = 2.98, SD = .76) relative to MP roles (M = 2.09, SD = .78) (t(738) = 6.81, p < .001). Mean job control for TL/Ms is, in turn, significantly higher than for MC roles (M = 3.93, SD = .75) (t(324) = 10.43, p < .001). Although some sections diverge within a role (e.g. the high job control score for E-Service Performance), the triangulation of several data sources about each section and overall mean differences between categories provides a sound basis for combining data across centres.

Examination of written job descriptions further supported the distinction between MP and MC roles, as similar job titles (e.g. CSR) could conceal differences in the nature of tasks performed. There were differences too in the regularity and stringency of performance management measures and procedures. Summarizing the data, those in MP roles were more likely to face regular performance reviews and the possibility of performance improvement measures. Those in MC roles, while still requiring appraisals, received fewer performance checks of lesser stringency, characteristics that additionally demonstrated their comparatively enhanced discretion.

The final pooled survey sample comprised 62, 26 and 12 percent in MP, MC and TL/M roles, respectively. Although some centres had a greater proportion of MP roles (Holstravel and Entcomm) and the UK industry at the time was primarily operating a MP business model (CM Insight, 2004), oversampling of MC and TL/M roles allowed sufficient observations for hypothesis tests. We applied sample weights based on the totals within each call centre for the three roles to adjust for this.

Analysis

Hypotheses 1 to 7 relate to two dependent variables: (1) the likelihood of being in a mass customization (MC) rather than mass production (MP) role, and (2) the likelihood of being in a team leader/manager (TL/M) rather than MP role. These hypotheses were examined through multinomial logistic regression (MLR).³ MLR runs logistic regressions between the reference category (MP roles) and the other categories, producing three equations, with coefficients for the reference group normalized to zero. These provided relative risk ratios (exponentiated logit coefficients), sometimes referred to as odds ratios (ORs), which can be interpreted as the probability of being a member of either the MC or TL/M groups. Values over one indicate positive relationships such that the independent variable increases the probabilities of being in that group. Values less than one mean the independent variable decreases the probabilities of being in the group.

Three models were estimated: Model 1 included gender and control variables; Model 2 added human capital, domestic responsibility and supervisory career support as possible mediators of gender effects in Model 1; and Model 3 added the interaction terms created after centring variables to reduce the mean to zero and avoid multicollinearity. For the purposes of the MLR, all predictors were recoded as dichotomous variables using the median value as the cutpoint in order to minimize problems with empty cells resulting from the use of covariates and enhance interpretability of interaction terms. For H8, a univariate generalized linear model (GLM) was used to determine the effect of gender and domestic responsibility on supervisor career support, adjusting for time in role, age and career aspiration, and controlling for workplace to account for possible differences in workplace HR practice.

We were unable to include the individual call centre as a fixed factor in the MLRs, given restrictions in the sample size representing MC and TL/M categories of the dependent variables. This was caused by typically flat hierarchies, which mean fewer specialized and team leader roles, despite our over-sampling of these categories. Including call centre as a fixed factor would have resulted in a large proportion of missing data in the subpopulations of interest. We, nevertheless, acknowledge possible differences in HR/management practice across centres by conducting univariate tests across and within call centres. We highlight differences in firm practice that may have impacted some predictors of interest (e.g. supervisory career support). We were also able to control for call centre in the analysis involving supervisory career support as the dependent variable.

Results

Table 1 shows the greater overall numerical concentration of respondents in MP roles, and, more specifically, the high proportion of women in MP (76%) relative to MC roles (57%) and TL/Ms (61%) (χ²(2) = 16.87, p < .001). Table 2 summarizes the survey sample means, standard deviations and intercorrelations by gender. Men and women were similar in age (59% below 30 years) and career aspirations, with even proportions indicating the call centre was/was not a career choice. As expected for three of the proposed independent variables in H2 and H3, women had lower qualifications, one-third below high school level compared with 20 percent of men; scored higher on domestic responsibility; and reported lower supervisory career support. However, contrary to expectation women had longer tenure. These differences were confirmed in a binary logistic regression regressing gender (women coded 1) on all control and independent variables (χ²(7) = 52.75, p < .001, −2 Log Likelihood = 937.56) and providing the following coefficients: education (β = –.35, p < .01); supervisory career support (β = –.25, p < .05); domestic responsibility (β = .32, p < .001); and tenure (β = .02, p < .05). Separate analyses carried out for each centre and each work role (analysis not shown) replicated these findings, except in the case of TL/Ms where men and women differed only in education level (t = 2.82, p < .05). With no evidence of multicollinearity among the independent variables (absolute correlation coefficients were .40 or less) all predictors were retained for further analysis.

Table 2

Means, standard deviations and intercorrelations (men and women)^a

	Men												Women
	M	(SD)	1	2	3	4	5	6	7	8	9	10	M	(SD)	t^c
1. Work role: MP^b	.49	(.50)		–	–	−.01	−.21	−.11	−.17	−.12	.11	−.18	.67	(.47)	5.01 **
2. Work role: MC^b	.35	(.48)	–		–	−.05	.10	.04	.19	.02	−.05	.06	.22	(.42)	3.95 **
3. Work role: TL/M^b	.17	(.37)	–	–		.08	.19	.11	.00	.15	−.15	.18	.11	(.31)	2.01 *
4. Age: >30	.40	(.49)	−.02	−.11	.16		.08	.09	−.32	.14	.40	−.03	.41	(.49)	−.21
5. Career job choice	.46	(.50)	−.37	.16	.29	.11		.05	−.11	.04	.04	.25	.41	(.49)	1.28
6. Time in role	10.94	(11.32)	−.23	.15	.11	.03	.12		.00	.40	.08	−.05	11.81	(12.17)	−.99
7. Education	2.04	(.68)	−.18	.10	.11	−.08	−.11	−.10		−.06	−.19	.01	1.84	(.72)	3.89 **
8. Tenure	17.96	(19.90)	−.18	.04	.20	.11	.10	.40	−.07		.15	−.04	22.70	(21.24)	−3.01 **
9. Domestic resp	1.84	(.99)	−.08	−.05	.11	.35	.08	.00	−.10	.12		−.14	2.22	(1.33)	−4.66 **
10. Supervisor support	3.01	(.67)	−.17	.06	.16	.07	.26	.16	−.07	.16	.18		2.90	(.68)	2.23 *

Notes: N(men) = 252, N(women) = 595; Age and Career job choice coded 0/1; Time in role and Tenure measured in months; Education scaled 1–3; Domestic responsibility scaled 1–8; Supervisor career support scaled 1–5.

Correlations below diagonal are for men and above for women. Absolute values of .12 and above (for men) and .06 and above (for women) are significant at p < .05.

MP ‘Mass Production’, MC ‘Mass Customization’, TL/M ‘Team Leader/Manager. Dummy variables coded 0/1 (correlations not reported).

Gender differences examined with independent-samples t test.

p < .05, ** p < .01.

Table 3 presents the odds ratios for the factors affecting the probability of being in either MC or TL/M, compared with MP, roles. Model 1 shows that being a woman decreased the odds of being in either role by about 50 percent, supporting H1. This gender effect occurred despite the strong influence of career preference, which made it over two times as likely for respondents to be in MC rather than MP roles, and over three times as likely to be TL/Ms. Model 2 added the mediators proposed as explanations of gendered segmentation (H2–H4). Having above high school qualifications increased the odds by about four times of being in an MC or TL/M role. Tenure was unrelated to being in an MC role, but for TL/M roles, there was a two percent gain in the odds for every unit change in months employed in the company. Women had longer tenure than men, however, contrary to expectation. Thus, H2b was supported only in the case of education. More domestic responsibility led to a 60 percent decrease in the odds of being in TL/M roles but was unrelated to MC roles. As women reported more domestic responsibility than men, this supported H3b in the case of TL/M roles. Supervisory career support, of which women had significantly less, also increased the odds of being in both MC and TL/M roles, thereby supporting H4b.

Table 3

Odds ratios for multinomial logistic regressions predicting probability of being in MC or team leader/manager roles^a

	Model 1		Model 2		Model 3
	MC v MP	TL/M v P	MC v MP	TL/M v MP	MC v MP	TL/M v MP
Women	0.50 ***	0.51 **	0.51 **	0.48 **	0.56 *	0.50 *
Age: >30	0.75	1.56 *	0.79	1.33 *	0.89	1.67 *
Career job choice	2.44 ***	4.60 ***	2.92 ***	4.85 ***	2.31 ***	3.69 ***
Time in role	2.06 ***	2.56 ***	2.52 ***	2.31 **	1.04 ***	1.03 *
Education: above HS^b			4.32 ***	3.61 ***	0.07 *	0.15 **
Tenure			0.91	2.38 ***	1.01	1.77 ***
Domestic responsibility			0.56	0.40 **	0.86	0.77 *
Supervisor support			1.59*	2.45 **	0.51 **	2.22 ***
Interaction terms^c
Women/above HS^d					0.25	0.56 *
Women/high domestic resp^e					0.63	0.55 *
−2 Log Likelihood	731.89		639.41		624.97
BIC	798.49		759.28		674.62
Nagelkerke Pseudo-R²	.16		.28		.30

Notes: N = 724 Intercept not shown. BIC: Bayesian Information Criterion; All predictors were recoded to dichotomous variables.

Reference category is ‘mass production’ role.

HS: High School. Reference category is HS qualifications or lower.

Significant interaction terms only included in final model.

Reference category is men/above HS education.

Reference category is men/high domestic responsibility.

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

For Model 2, overall, the −2 Log Likelihood and other model comparison criteria (Bayesian Information Criterion), which estimate the unexplained variance remaining, were reduced from Model 1, indicating better overall model fit. However, the significant main effects of education, domestic responsibility and supervisory career support did not remove the gender effect, suggesting only partial mediation for these variables.

Model 3 reports the model adding significant interaction terms. The interaction of gender with tenure was not significant for either role, and only in the case of TL/Ms were the interactions with education and domestic responsibility significant. Women with higher qualifications had 44 percent lower odds of occupying TL/M roles than men with equivalent qualifications. Similarly, women with high domestic responsibility had 45 percent lower odds than men with high levels of responsibility of occupying TL/M roles. These significant interactions supported H5 and H6 in the case of TL/Ms, but not for MC roles. The three-way interaction between gender, human capital (education or tenure) and domestic responsibility was not significant in either case. Thus, H7 was not supported. Running the MLRs again for men and women separately (Table 4), the different effects of human capital and domestic responsibility for female TL/Ms are apparent. While education has the same benefits for men and women in MC roles, and predicts TL/Ms roles for men, it plays no role for women in predicting TL/M role. In addition, high domestic responsibility reduces the likelihood of women occupying TL/M roles by 44 percent compared with women with low domestic responsibility, but has no effect for men. An unexpected finding was that supervisory career support predicted work role for women but not men. Thus, although women reported less supervisory career support than men (H4a), its effects when present were to increase their likelihood of being in MC and TL/M roles.

Table 4

Odds ratios for multinomial logistic regressions predicting probability of being in MC or team leader/manager roles (men and women)^a

	Men		Women
	MC v MP	TL/M v MP	MC v MP	TL/M v MP
Age: >30	0.69	1.67	1.12	1.57
Career job choice	4.83***	12.51***	2.42***	3.73***
Time in role	5.78***	3.11*	1.69*	2.27*
Education: above HS^b	4.51***	10.41***	5.08***	1.91
Tenure	0.75	4.11*	0.98	1.90*
Domestic responsibility	0.79	1.30	0.89	0.56*
Supervisor support	0.99	1.33	1.63*	2.78**
−2 Log Likelihood	250.28		378.77
Nagelkerke Pseudo-R²	.42		.30
N	243		595

Notes:

Reference category is ‘mass production’ role.

HS: High School. Reference category is HS qualifications or lower.

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

Finally, there was a significant interaction between gender and domestic status in the prediction of supervisor career support (F(1846) = 3.66, p < .05), while controlling for call centre age, career aspiration and time in role. The estimated marginal means and standard errors from this analysis (Table 5) showed no gender difference in supervisor career support at lower levels of domestic responsibility but significant divergence for men and women at high levels. This supports H8 although we did not hypothesize that domestic status would actually increase supervisory career support for men. This is shown in Figure 2.

Table 5

Estimated marginal means for supervisor career support

	Low domestic responsibility		High domestic responsibility
	M(S.E.)	95% CI	M(S.E.)	95% CI
Men	2.98(.05)	2.88,3.08	3.08(.08)	2.92,3.24
Women	2.94(.04)	2.87,3.02	2.83(.04)	2.75,2.91

Discussion

The findings confirm our hypotheses that women were disproportionately segregated into mass production (MP) as opposed to mass customization (MC) and team leader/manager (TL/M) roles, and that this was related to their lower educational qualifications, greater domestic responsibility and lower levels of reported supervisory career support. The effects of education and domestic responsibility were moderated by gender for TL/M roles only, indicating further disadvantage in these but not MC roles. Specifically, men with above high-school qualifications were more likely than women with the same qualifications to occupy TL/M roles, while women with high domestic responsibility were less likely than men of equivalent domestic status to occupy these roles. For MC roles, human capital was equally predictive for both men and women, and there was no differential effect of domestic status.

Our proposed three-way interaction between gender, human capital and domestic status was not supported, suggesting that women’s domestic responsibilities did not interfere with the positive effects of human capital on work role. However, domestic responsibilities did have a negative impact on supervisory career support for women, while for men, this effect was reversed – supervisor support increased as men’s domestic responsibilities increased. This was all the more significant given the additional finding that supervisory career support was more important for women than for men, as it increased women’s chances of being employed in MC and TL/M roles but had no effect for men.

Theoretical implications

Although existing research has shown women’s success in progressing to call centre management roles (Belt, 2002; Durbin, 2006), the present study shows that women continue to be over-represented in mass production CSR roles. Our aim was to provide some explanation for this gendered role segmentation as well as revealing potential barriers facing women’s progression to less routinized CSR and managerial roles. These explanations were located both outside the workplace (in women’s lower human capital and the effects of domestic division of labour) and within the workplace (in the influence of supervisors’ career support). Several theoretical implications emerge from these findings that contribute to understanding both call centre employment and women’s careers in newer service contexts.

First, there is some support for human capital theory (Becker, 1985). Qualifications explained both women’s segmentation into MP roles and women’s progress to MC roles. There was also an internal labour market based on acquired company-specific skills, measured by organizational tenure, in the case of TL/Ms, which applied equally for men and women. Our finding that women had longer tenure than men may point to a potential route out of MP roles for women. Qualitative evidence suggests that in the context of relatively high labour turnover, men and women who demonstrate their capabilities as call handlers and accumulate competencies from different departments have an increased likelihood of becoming team leaders in MP activities, if only because they have remained while many colleagues have departed. One agent frustrated by the few openings for team managers explained:

[My manager] took me aside and she said sooner or later [the centre manager] will decide to send someone on a project or make a new team, and then there will be team manager positions. She says it will happen. (CSA, Unregulated Loans, Moneyflow)

Thus, education and tenure are related to women’s career success in MC and TL/M roles, respectively, similar as for men. This confirms findings from other work contexts on the importance of human capital for women’s careers (Metz and Tharenou, 2001; Ng et al., 2005).

Nonetheless, these workplaces attract women with lower qualifications than men. Hopes that call centres, which are often located in areas of high unemployment (Richardson and Belt, 2001), might remove barriers to career opportunities for those with fewer formal qualifications are not supported; instead, it seems likely that women with less educational capital will remain ‘trapped’ in inferior jobs within ‘female ghettos’ (Belt, 2002) unless they can accumulate sufficient tenure for when team leader roles become available. As well as exacerbating growing inequalities in employment between women with less and more education (Konietzka and Kreyenfeld, 2010), the question also arises as to why call centres attract large numbers of less educated women, but not less educated men. Unemployment and labour market insecurity may be worse for men than women (Oppenheimer, 1994) suggesting that these women are actually better off than less educated men. More sceptically, to encourage workforce stability in the face of high attrition and to enhance organizational flexibility, employers may intentionally recruit from ready supplies of women with lower qualifications and conducive domestic situations, who are willing to accommodate flexible shifts (Holtgrewe, 2006). Whether men with low qualifications gravitate to better or worse employment situations could not be evaluated, but assuming that call centres will continue to attract women with low qualifications, their inability to escape these roles means that inequalities are likely to persist.

A second contribution is to show that, although human capital is sufficient to explain men’s call centre work roles and plays a role for women, the domestic division of labour still shapes women’s careers. Those assuming more of the domestic burden were less likely to occupy TL/M roles than men in the same position or women with lower domestic responsibilities. This was not just because of decisions not to invest in human capital, as the ‘domestic penalty’ extended to women with higher qualifications as elsewhere (Tomlinson et al., 2009). Nor can we assume that women in these circumstances prefer less responsible jobs (Hakim, 2000) as we controlled for career aspiration. Women continue to bear a greater domestic burden (Breen and Cooke, 2005; Dex, 2003), which leads to career decisions around household dynamics, such as partners’ higher earnings (Powell and Greenhaus, 2010), accessibility of childcare (McRae, 2003), or the hours demanded in management roles (Harkness, 1999). Female managers described increased pressures associated with supervisory roles. For example:

You take work home, you can’t avoid it. They told me when I got promoted that I was now working beyond contract. (Team Leader, Holstravel)

Thus, women in managerial roles conform to a ‘male breadwinner career model’ (Kelan, 2008) by removing domestic interference in order to progress. Although call centres have been thought of as flexible workplaces able to accommodate family demands (Belt, 2002) this does not seem to apply to women who achieve more senior roles.

Third, the study suggests that the ‘glass escalator’ effect observed in other female-dominated contexts, which accelerates men’s progress (Maume, 1999; Williams, 1992), seems to exist in call centres. Women with domestic commitments, regardless of career aspiration, age or time in role, received less supervisor support for training and developmental opportunities than men, who received more support with greater domestic commitments. This was despite the fact that supervisor support was an important predictor of being in a MC and TL/M role for women, confirming other evidence that women respond better to organizational supports, perhaps because of restricted opportunities elsewhere (Tharenou and Conroy, 1994). The finding for men is consistent with other evidence that domestic commitments seem to benefit men’s careers, in contrast to its effects on those of women (Kirchmeyer, 2006).

The source of disadvantage here lies within the workplace, with supervisors’ inability to accommodate career or training opportunities to personal demands. Our findings support the proposal that supervisors base decisions about allocation of resources on an implicit job matching model (Barron et al., 1989, 1993) applying stereotypes, for example, about women’s commitment, in the way they stream men and women towards opportunities (Bielby and Baron, 1986; Heilman, 2001). Women with dependents are less able to access training (Metz, 2005), and interrupted employment and reduced hours (Knote and Ishio, 1998) are likely to enhance this job matching effect. An illustration from our qualitative data was the irritation several women felt with shift patterns that clashed with domestic demands and created barriers for accessing training opportunities, without any allowances from management. This frustration is perceptible in the following narrative:

The better jobs are just out of reach, because the training you need to give yourself a chance to get them is often during the day when I’m on back shift. They say that I could come in early and do the training in the afternoon but that’s when I’m catching up on housework, picking up the kids. Even when I’m on days I miss training because they’re short on the phones and they won’t let you go. I want to get ahead so it’s all really frustrating. (CSA, Energyco, Thejobshop)

Thus, we show that women, who already assume more domestic responsibility regardless of career aspiration, are disadvantaged further by supervisor stereotyping because of their domestic status. Previous findings that women in call centres do not receive the training required to advance their careers (Belt, 2002; Durbin, 2006) did not consider how domestic life interacted with this workplace experience. The present study shows that supervisors are either unwilling or unable to counteract structural obstacles to development opportunities that reside both in domestic life and the constraints of call centre work organization.

Practical implications

To the extent that there is a ‘success story’ for these women, it seems to depend on qualifications for technical or business-related CSR (mass customization) roles and conducive domestic situations or tenure for managerial roles. Progress in the latter may involve a reduced burden of domestic labour, such as a supportive partner or support networks (Hyman et al., 2005), combined with dual breadwinner arrangements. The importance of organizational tenure also suggests some recognition of firm-specific knowledge, equally, for both men and women.

However, it is the implications for supervisors that should be emphasized, given their importance for women in both types of roles and the apparent stereotyping of domestic constraints experienced relative to men. Consistent with Webster (2004), these findings stress that supervisors should enable women’s access to formal training for higher level skills and work processes if they are to advance to more responsible call centre roles. Thus, positive discrimination, still regarded as necessary for ‘transformative’ change (Noon, 2010), should encourage women’s uptake of developmental opportunities. This has already been recognized as an issue within call centres. All four organizations espoused formally progressive equal opportunities policies, and one (Moneyflow) was a founding member of ‘Opportunity 2000’, an initiative aimed at improving women’s representation at management levels, promoting open access to training and development, ‘career break’ schemes holding positions open, job-sharing, crèche facilities and flexible working arrangements (e.g. returner’s bonus of £100/month to offset childcare costs). However, this centre also excluded job sharers from becoming team leaders and was unsympathetic to the need to care for sick dependents. The other three centres were even less enlightened than this ‘industry leader’ displaying similarly ‘macho’ management cultures in their accommodation of domestic demands.

It seems that call centres are a long way from realizing these commitments to equal access. Tight staffing levels compounded by high turnover in MP roles and relatively few senior roles compel managers to keep agents on the phones, emphasizing targets rather than personal development. Given that perceptions of women’s domestic commitments may be the source of supervisors’ actions, more should be done to encourage them to implement family-friendly career policies and to resist sacrificing equality initiatives to the tide of pressures to increase targets and tightly manage shifts, as frequently occurs (Noon, 2007).

Conclusions

This article demonstrates that for many women in call centres, employment is shaped by a pre-existing disadvantage resulting from constrained educational capital and domestic freedom, which is then perpetuated in their workplace experience of supervisory support for development. Our distinctive contribution is to analyse the combined and interactive effects of these factors, as depicted in Figure 1, in a newly emerged service context with high female concentrations. Call centre work has been heralded as offering unhindered advancement opportunities for women based on experience and regardless of qualifications (Russell, 2008). The conclusion is that these putatively less bureaucratic workplaces apparently have failed to dissolve gender differences as hoped for by new economy optimists (Beck, 2000; Castells, 2000). With the exception of MC roles where women’s presence was unrelated to domestic status and depended on human capital, as for men, the findings suggest continuity with the gendered segmentation found in earlier studies of women’s white collar work (e.g. Crompton and Jones, 1984). Our findings also seem to support Huws’ (2003) observations that the progenitors of the female call centre workers are indeed the routine bank clerks and telephone operators of the past.

This is the first large-scale study of gender in call centres that applies a multivariate framework to career issues. A major limitation of descriptive and qualitative studies (Belt, 2002; Belt et al., 2000, 2002; Durbin, 2006) is that they have concentrated on the gendered practices and experiences within the workplace as if this self-contained organizational dimension were able by itself to explain discrimination. The workplace, we argue, provides a necessary but insufficient contribution. In order to understand women’s position, their household experiences and constraints must perforce be integrated with human capital characteristics and their treatment at work, and this analytical approach should be adopted for a large sample systematically identified as occupying distinct work roles. This separation of call centre work roles to allow the analysis of segmentation is itself a novel contribution. Avoiding dualistic representations of the call centre as either semi-skilled knowledge work for women or a cul-de-sac, the study shows that it is not a female ghetto for all women, particularly those who may aspire to MC roles. Nevertheless, women do experience a relative disadvantage compared with men, which cannot be explained by the conventional wisdom that women in call centres prefer to prioritize domestic commitments over career progression (Belt, 2002: 54). Indeed, a criticism of Belt’s work is that the failure to incorporate systematically the influence of domestic and household constraint and of human capital implies an explanatory default to preference theory. This observation provides a wider contribution to debates about choice versus structural constraint in explaining women’s career outcomes (Crompton and Harris, 1998; Hakim, 2000; Gash, 2008).

While the call centre may not be archetypical of all routinized jobs in the contemporary service sector, by virtue of its high concentration of women, non-standard shift patterns, role segmentation and female disadvantage, this study may nevertheless have important general implications. Within the context of sustained sectoral growth (Lloyd and Payne, 2009), women still experience the ambiguity of opportunity. Inevitably, some women have achieved ‘better’ jobs; but equally, women’s progression has been hindered by the complex interaction of low human capital, domestic responsibilities and difficulties accessing social capital that confine many for long periods to more routinized roles.

Limitations and future research

The study is limited by its cross-sectional design and its focus on a single employment sector, albeit that it addressed a significant gap within previous research on call centres and service work more broadly as sources of opportunity for women. The research design was unable to locate with accuracy the decision and transition points for men and women in progressing to better roles. Future research could longitudinally track women over the life course of work, studying cohorts from their entry point into call centres and charting their respective treatment by managers, development activities, career decisions and progression, while controlling for human capital, domestic circumstance and local labour market conditions. These transition effects, along with the emergence of gender disadvantage over time, were captured by Hull and Nelson’s (2000) panel study on women’s legal careers. Further research should also attempt to unravel the nature of supervisor gender role stereotyping in call centres, building on existing qualitative evidence as well as the present findings that link women’s access to social capital and their domestic lives.

The title of the article, ‘beneath the glass ceiling’, captures the continued presence of a glass ceiling within call centre work while adding the idea of pre-existing barriers for women before they enter the workplace. Removal of the glass ceiling should attend not only to workplace discrimination but also to these initial disadvantages; for example, in educational capital and the implications of a continuing domestic division of labour, which appear to filter into the workplace and continue to impact on women’s careers.

Footnotes

This study received financial support from the ESRC Future of Work Programme – Employment and Working Life Beyond the Year 2000, grant number L212252006.

Notes

References

Alvesson

Billing

(1997) Understanding Gender and Organizations. London: SAGE.

Barron

Black

Loewenstein

(1989) Job matching and on-the-job training. Journal of Labor Economics 7(1): 1–19.

Barron

Black

Loewenstein

(1993) Gender differences in training, capital, and wages. The Journal of Human Resources 28(2): 343–364.

Batt

Moynihan

(2002) The viability of alternative call centre production models. Human Resource Management Journal 12(4): 14–34.

Batt

Hunter

Wilk

(2002) How and when does management matter? Job quality and career opportunities for call center workers. In: Appelbaum

Bernhardt

Murnane

(eds) The Labor Market for Low Wage Workers. New York: Russell Sage Foundation, 270–312.

Beck

(2000) The Brave New World of Work. London: Polity Press.

Becker

(1985) Human capital, effort, and the sexual division of labor. Journal of Labor Economics 3(1, pt. 2): 33–58.

Becker

(1993) Human Capital: A Theoretical and Empirical Analysis with Special References to Education (3rd ed.). Chicago: The University of Chicago Press.

Belt

(2002) ‘A female ghetto?’ Women’s careers in call centres. Human Resource Management Journal 12(4): 51–66.

10.

Belt

Richardson

Webster

(2000) Women’s work in the information economy. The case of telephone call centres. Information, Communication & Society 3(3): 366–385.

11.

Belt

Richardson

Webster

(2002) Women, social skill and interactive service work in telephone call centres. New Technology, Work and Employment 17(1): 20–34.

12.

Bielby

(2000) Minimizing workplace gender and racial bias. Contemporary Sociology 29(1): 120–129.

13.

Bielby

Baron

(1984) A woman’s place is with other women: Sex segregation within organizations. In: Reskin

(ed.) Sex Segregation in the Workplace: Trends, Explanations, Remedies. Washington, DC: National Academy Press, 27–55.

14.

Bielby

Baron

(1986) Men and women at work: Sex segregation and statistical discrimination. American Journal of Sociology 91(4): 759–799.

15.

Boreham

Parker

Thompson

Hall

(2008) New Technology@Work. London: Routledge, 59–82.

16.

Breen

Cooke

(2005) The persistence of the gendered division of domestic labour. European Sociological Review 21(1): 43–57.

17.

Bryk

Raudenbausch

(1992) Hierarchical Linear Models: Applications and Data Analysis Methods. Newbury Park, CA: SAGE.

18.

Callaghan

Thompson

(2002) ‘We recruit attitude’: The selection and shaping of routine call centre labour. Journal of Management Studies 39(2): 233–253.

19.

Castells

(2000) The Rise of the Network Society. Oxford: Oxford University Press.

20.

CM Insight (2004) The Truth about Contact Centre Performance: Mass Customisation versus Mass Production, part 1. London: CM Insight.

21.

Cohen

Broschak

Haveman

(1998) And then there were more? The effect of organizational sex composition on the hiring and promotion of managers. American Sociological Review 63(5): 711–727.

22.

Collin-Jacques

Smith

(2005) Nursing on the line: Experiences from England and Quebec (Canada). Human Relations 58(1): 5–32.

23.

Crompton

Harris

(1998) Explaining women’s employment patterns: ‘Orientations to work’ revisited. The British Journal of Sociology 49(1): 118–136.

24.

Crompton

Jones

(1984) White-Collar Proletariat: Deskilling and Gender in Clerical Work. London: Macmillan.

25.

Deery

Kinnie

(2004) (eds) Call Centres and Human Resource Management: A Cross-National Perspective. Basingstoke: Palgrave.

26.

Dex

(2003) Families and Work in the Twenty-first Century. York: Joseph Rowntree Foundation.

27.

Durbin

(2006) Gender, skills and careers in UK call centres. In: Burgess

Connell

(eds) Developments in the Call Centre Industry. London: Routledge, 117–135.

28.

Frenkel

Korczynski

Shire

Tam

(1999) On the Front Line. Ithaca: Cornell University Press.

29.

Gash

(2008) Preference or constraint? Part-time workers transition in Denmark, France and the United Kingdom. Work, Employment & Society 22(4): 655–674.

30.

Gershuny

(2004) Time through the life-course in the family. In: Scott

Treas

Richards

(eds) The Blackwell Companion to Sociology of Families. Oxford: Blackwell, 158–178.

31.

Golsch

(2007) Women’s employment in Britain. In: Blossfeld

Hofmeister (eds) Globalization, Uncertainty and Women’s Careers: An International Comparison. Cheltenham: Edward Elgar, 275–298.

32.

Gorjup

Valverde

Ryan

(2008) Promotion in call centres: Opportunities and determinants. Journal of European Industrial Training 32(1): 45–62.

33.

Greenhaus

Parasuraman

Wormley

(1990) Effects of race on organizational experiences, job performance evaluations, and career outcomes. Academy of Management Journal 33(1): 64–86.

34.

Hakim

(2000) Work-Lifestyle Choices in the 21st Century: Preference Theory. Oxford: Oxford University Press.

35.

Harkness

(1999) Working 9 to 5. In: Gregg

Wadsworth

(eds) The State of Working Britain. Manchester: Manchester University Press, 90–108.

36.

Heilman

(2001) Description and prescription: How gender stereotypes prevent women’s ascent up the organizational ladder. Journal of Social Issues 57(4): 657–674.

37.

Holtgrewe

(2006) Restructuring gendered flexibility in organisations: A comparative analysis of call centres in Germany. In: Gottfried

Walby

Gottschall

Osawa

(eds) Gendering the Knowledge Economy: Comparative Perspectives. Basingstoke: Palgrave, 250–272.

38.

Hull

Nelson

(2000) Assimilation choice or constraint? Testing theories of gender differences in the careers of lawyers. Social Forces 79(1): 229–264.

39.

Hultin

(2003) Some take the glass escalator, some hit the glass ceiling? Career consequences of occupational sex segregation. Work and Occupations 30(1): 30–61.

40.

Huws

(2003) The Making of a Cybertariat: Virtual Work in a Real World. New York: Monthly Review Press.

41.

Hyman

Scholarios

Baldry

(2005) Getting on or getting by? Employee flexibility and coping strategies for home and work. Work, Employment & Society 19(4): 705–725.

42.

Judge

Cable

Boudreau

Bretz

Jr (1995) An empirical investigation of the predictors of executive career success. Personnel Psychology 48(3): 485–519.

43.

Kelan

(2008) Gender, risk and employment insecurity: The masculine breadwinner subtext. Human Relations 61(9): 1171–1202.

44.

Kerfoot

Korczynski

(2005) Gender and service: New directions for the study of ‘front-line’ service work. Gender, Work and Organization 12(5): 387–399.

45.

Kirchmeyer

(2006) The different effects of family on objective careers success across gender: A test of alternative explanations. Journal of Vocational Behavior 68(2): 323–346.

46.

Knote

Ishio

(1998) The gender gap in company job training. Work and Occupations 25(2): 141–167.

47.

Konietzka

Kreyenfeld

(2010) The growing educational divide in mothers’ employment: An investigation based on the German micro-censuses 1976–2004. Work, Employment & Society 24(2): 260–278.

48.

Korczynski

(2002) Human Resource Management in Service Work. Basingstoke: Palgrave.

49.

Landau

Arthur

(1992) The relationship of marital status, spouse’s career status, and gender to salary level. Sex Roles 27(11–12): 665–668.

50.

Lazear

Rosen

(1990) Male-female wage differentials in job ladders. Journal of Labor Economics 8(1): S106–S123.

51.

Litzky

Greenhaus

(2007) The relationship between gender and aspirations to senior management. Career Development International 12(7): 637–659.

52.

Lloyd

Payne

(2009) ‘Full of sound and fury, signifying nothing’: Interrogating new skill concepts in service work – the view from two call centres. Work, Employment & Society 23(4): 617–634.

53.

Lyness

Thompson

(1997) Above the glass ceiling? A comparison of matched samples of female and male executives. Journal of Applied Psychology 82(3): 359–375.

54.

McRae

(2003) Constraints and choices in mothers’ employment careers: A consideration of Hakim’s Preference Theory. British Journal of Sociology 54(3): 317–338.

55.

Maume

(1999) Glass ceilings and glass escalators: Occupational segregation and race and sex differences in managerial promotions. Work and Occupations 26(4): 483–509.

56.

Melamed

(1995) Barriers to women’s career success: Human capital, career choices, structural determinants, or simply sex discrimination. Applied Psychology: An International Review 44(4): 295–314.

57.

Metz

(2005) Advancing the careers of women with children. Career Development International 10(3): 228–245.

58.

Metz

Tharenou

(2001) Women’s career advancement: The contribution of human and social capital. Group and Organization Management 26(3): 312–342.

59.

Nardone

(1995) Part-time employment: Reasons, demographics and trends. Journal of Labor Research 16(3): 275–292.

60.

TWH

Eby

Sorensen

Feldman

(2005) Predictors of objective and subjective career success: A meta-analysis. Personnel Psychology 58(2): 367–408.

61.

Noon

(2007) The fatal flaws of diversity and the business case for ethnic minorities. Work, Employment & Society 21(4): 773–784.

62.

Noon

(2010) The shackled runner: Time to rethink positive discrimination? Work, Employment & Society 24(4): 728–739.

63.

Oppenheimer

(1994) Women’s rising employment and the future of the family in industrial societies. Population and Development Review 20(2): 293–342.

64.

Orser

Leck

(2010) Gender influences on career success outcomes. Gender in Management: An International Journal 25(5): 386–407.

65.

Powell

Greenhaus

(2010) Sex, gender and decisions at the family-work interface. Journal of Management 36(4): 1011–1039.

66.

Richardson

Belt

(2001) Saved by the bell? Call centres and economic development in less favoured regions. Economic and Industrial Democracy 22(1): 67–98.

67.

Russell

(2008) Call centres: A decade of research. International Journal of Management Reviews 10(3): 195–219.

68.

Russell

(2009) Smiling Down the Line: Info-Service Work in the Global Economy. Toronto: University of Toronto Press.

69.

Scott

(2010) Quantitative methods and gender inequalities. International Journal of Social Research Methodology 13(3): 223–236.

70.

Sieben

de Grip

(2004) Training and expectations on job mobility in the call centres sector. Journal of European Industrial Training 28(2–4): 257–271.

71.

Sprigg

Stride

Wall

Holman

Smith

(2007) Work characteristics, musculoskeletal disorders, and the mediating role of psychological strain: A study of call center employees. Journal of Applied Psychology 92(5): 1456–1466.

72.

Stewart

Gudykunst

(1982) Differential factors influencing the hierarchical level and number of promotions of males and females within an organization. Academy of Management Journal 25(3): 586–597.

73.

Stroh

Brett

Reilly

(1992) All the right stuff: A comparison of female and male manager’s career progression. Journal of Applied Psychology 77(3): 251–260.

74.

Taylor

Anderson

(2008) Contact Centres in Scotland – The 2008 Audit. Glasgow: SDI.

75.

Taylor

Hyman

Mulvey

Bain

(2002) Work organization, control and the experience of work in call centres. Work, Employment & Society 16(1): 133–150.

76.

Tharenou

(1999) Is there a link between family structures and women’s and men’s managerial career advancement? Journal of Organizational Behavior 20(6): 837–863.

77.

Tharenou

(2001) Going up: Do traits and informal social processes predict advancing in management? Academy of Management Journal 44(5): 1005–1017.

78.

Tharenou

Conroy

(1994) Men and women managers’ advancement: Personal or situational determinants? Applied Psychology: An International Review 43(1): 5–31.

79.

Tharenou

Latimer

Conroy

(1994) How do you make it to the top? An examination of influences on women’s and men’s managerial advancement. Academy of Management Journal 37(4): 899–931.

80.

Tomlinson

Olsen

Purdam

(2009) Women returners and potential returners: Employment profiles and labour market opportunities. European Sociological Review 25(3): 349–363.

81.

Van der Heijden

BIJM

De Lange

Demerouti

Van der Heijde

(2009) Age effects on the employability – career success relationship. Journal of Vocational Behavior 74(2): 156–164.

82.

Walsh

(1999) Myths and counter-myths: An analysis of part-time female employees and their orientations to work and working hours. Work, Employment & Society 13(2): 179–204.

83.

Warren

Rowlingson

Whyley

(2001) Female finances: Gender wage gaps and gender assets gaps. Work, Employment & Society 15(3): 465–488.

84.

Wayne

Liden

Kraimer

Graf

(1999) The role of human capital, motivation and supervisor sponsorship in predicting career success. Journal of Organizational Behavior 20(5): 577–595.

85.

Webb

(2010) Gender and the post-industrial shift. In: Scott

Crompton

Lyonette

(eds) Gender inequalities in the 21^st Century. New Barriers and Continuing Constraints. Cheltenham: Edward Elgar Publishing, 85–108.

86.

Webster

(2004) Digitising inequality: The cul-de-sac of women’s work in European services. New Technology, Work and Employment 19(3): 160–176.

87.

Williams

(1992) The glass escalator: Hidden advantages for men in the ‘female’ professions. Social Problems 39(3): 253–267.

88.

Zeytinoglu

Cooke

(2008) Non-standard employment and promotions: A within genders analysis. Journal of Industrial Relations 50(2): 319–337.