Stressors and work-family conflict among female teachers in urban public schools in Kenya

In the study reported on here we sought to identify teaching and general stressors, and how they contributed to work-family-conflict among female teachers in urban public schools in Kenya. A total of 375 female teachers with at least their youngest child not yet in primary school completed a survey of closed and open-ended questions. Role theory, stressors, and work-family conflict literature informed the study. The data were analysed using hierarchical regression and content analysis. Findings yielded some contradictory results suggesting that, while the work-family conflict is a universal construct, culture and socioeconomic factors likely influence the nature and the impact of stressors. That is, these teachers identified stressors not mentioned in previous Western literature and not captured by quantitative measures. For instance, the greatest stressor cited was domestic workers not being reliable. The study also indicated that disruptions in the normal flow of work was regarded as a stressor. In addition, a sick child, rather than its age, was regarded as a stressor. Theoretical and practical contributions of the study are provided.


Introduction
Today's organisations expose workers to increased occupational stress. This is due to the changing nature of the globalised economy in which these organisations operate, characterised by work intensification as reported in Australia (Fein, Skinner & Machin, 2017), Europe (Green & McIntosh, 2001), Malaysia (Ismail & Teck-Hong, 2011), and other parts of the world. The globalised economy is characterised by changing technology, job designs, and cost-cutting measures leading to intensification of work. Organisations tend to have a lean staff who need to work more hours. In addition, there is a blurred boundary between work and life, and increased job insecurity (Bamberger, Larsen, Vinding, Nielsen, Fonager, Nielsen, Ryom & Omland, 2015;Moen, Lam, Ammon & Kelly, 2013) causes employees in various occupations to experience stress from work and non-work sources (Bamberger et al., 2015;Cinamon & Rich, 2005;Janik & Rothman, 2015;Tiyce, Hing, Cairncross & Breen, 2013). Some level of stress is beneficial as it acts as motivation to employees. However, in many parts of the world high levels of stress at work are associated with negative consequences such as job dissatisfaction, high turnover, burnout, marital dissatisfaction, ill health, et cetera (Allen, Herst, Bruck & Sutton, 2000;Eby, Casper, Lockwood, Bordeaux & Brinley, 2005). In essence, stress or a stressor is any environmental, social, or internal demand which requires the individual to adjust his/her usual behaviour patterns (Holmes & Rahe, 1967). It follows that some professions, such as teaching, are considered more stressful than others (Clunies-Ross, Little & Kienhuis, 2008;Ng'eno, 2007;Shernoff, Mehta, Atkins, Torf & Spencer, 2011;Sichambo, Maragia & Simiyu, 2012). One source of stress is work-family conflict (Greenhaus & Beutell, 1985). Incidentally, role theory is one of the underpinning theories in work-family conflict literature (Kahn, Wolfe, Quinn, Snoek & Rosenthal, 1964).
Research shows that employees, especially females with caregiving responsibilities and young children, face more work-family conflictespecially if these women are not adequately supported (Cinamon & Rich, 2005). In this study we attempted to identify the nature of stressors that female teachers with young children in urban public schools face, and how these stressors relate to work-family conflict. The Kenyan education system has experienced massive expansion with the introduction of free primary education in 2003, however, it lacks the necessary human and other teaching resources (Sichambo et al., 2012).
Over the years, the extended family has provided the needed support with regard to childcare and housework, with domestic workers primarily drawn from the informal networks supplementing it. However, the extended family support is shrinkingespecially in urban areas of sub-Saharan Africa (Mokomane, 2014). Moreover, domestic workers have become increasingly unreliable (Muasya & Martin, 2016). In Kenya, the government has continued to advocate for better domestic workers' remuneration making this form of support fairly expensive (Muasya, 2014). A few studies in Kenya have sought to identify stressors that lead to burnout and turnover among teachers in Kenya (Muasya, 2017b;Ng'eno, 2007;Sichambo et al., 2012). However, few researchers have attempted to identify stressors with regard to gender, young children, working in a highly stressful occupation such as teaching, and urban settings where extended family support is dwindling. In this study we explored the role of culture, how it impacts on stressors at work and home, and how these stressors differ from those identified in the predominantly Western literature.

Role Theory, Work-family Conflict and Stressors
This study was underpinned by role theory. Role theory assumes a theatrical metaphor where each actor in a play knows his/her role (Biddle, 1986). In society, people fulfil many roles. As social actors, people should know the roles, social position and expectations (norms, beliefs and preferences) of these roles. Managing multiple roles can lead to positive outcomes (role expansion) e.g. high selfesteem and life satisfaction, or negative outcomes (role stress) due to role overload and role conflict (Nordenmark, 2004). Kahn et al. (1964) extended role theory to organisations. They posit that organisations have different demands or norms which people should follow. Failure to follow these norms leads to role conflict and strain.
Biddle (1986) notes that role theory is hailed for its simplicity and heuristic qualities and has been applied in many social disciplines to solve social problems. However, it has its own limitations. Role theory has many extensions such as role conflict, role taking, and role-playing, which are not theoretically grounded; additionally, these extensions have conflicting sets of assumptions.
Role theory posits that constrained resources and incompatibility among roles make it hard for people to achieve the expectations of the different roles (Kahn et al., 1964). One application of role conflict is the inter-role conflict between work and non-work roles (family, leisure). Based on role theory, Greenhaus and Beutell (1985) developed a work-family conflict construct which has three major groups of antecedents or stressors: time based, strain based and behaviour based. Workfamily conflict (WFC) is "a form of inter-role conflict in which the role pressures from the work and family domains are mutually incompatible in some respects" (Greenhaus & Beutell, 1985:77). Most of the work-family conflict literature is informed by role theory (Rantanen, Kinnunen, Mauno & Tillemann, 2011). In this study, role theory and workfamily conflict provide a framework for identifying stressors.
Role theory assumes that home and work are mutually exclusive. The theory is criticised for upholding the bifurcation between work and family (Golden, Kirby & Jorgenson, 2006). Interestingly, the Kenyan formal employment sector seems to follow the model of separation of work and family that was adopted by the West during the Industrial Revolution (Golden et al., 2006). However, the division of housework and childcare has, like in other sub-Saharan countries, followed the traditional collectivistic model where females within the extended family shared these chores and men worked outside the home (Tsikata, 2009).
When individuals are faced with limited resources, they struggle to achieve multiple roles, which creates a sense of role overload. Role overload occurs when an individual feels that s/he has too much to do within the available time and with the available resources (Bolino & Turnley, 2005). For a worker, role overload may manifest in the form of him/her bringing work home and working beyond normal working hours. Bolino and Turnley (2005) found a relationship between work overload and initiative. Workers who excel and benefit their organisations with new ideas, programmes and projects may work overtime, which is an antecedent for work-family conflict.
In addition, work-family conflict can be related to and may be exacerbated by the place where one works (Muasya, 2017b). According to Okonkwo (2014), despite having young children, teachers in Enugu, a small city in Nigeria, did not experience strain-based work interference with family. This was due to the presence of the extended family and domestic workers who acted as resources to offset the effects of work-family related strain. However, this scenario is not the same in Lagos, Nigeria's capital. Epie and Ituma (2014) found that professionals in Lagos spent long hours commuting due to traffic jams and poor road conditions. These issues contributed to work-family conflict, which led to professionals quitting their jobs. Long commutes reduce time and energy resources required for fulfilling work and family roles. This study was conducted in an urban setting, Nairobi (three million inhabitants) and three rural towns Eldoret (280,000 inhabitants), Machakos, (150,000 inhabitants) and Wote (50,000 inhabitants) (Kenya National Bureau of Statistics, 2009). Large cities and large towns in sub-Saharan Africa experience traffic congestions. Moreover, in cities, extended families are not available to assist with household chores and childcare, resulting in reliance on domestic workers (Muasya, 2014).
Some studies in Western and Kenyan contexts have attempted to identify teacher-specific stressors, which include excessive teaching workloads, teacher involvement in student misbehaviour, a lack of basic resources and personnel, and accountability policies (Clunies-Ross et al., 2008;Shernoff et al., 2011;Sichambo et al., 2012). In the Kenyan context, researchers have linked these stressors with burnout but not with work-family conflict. In this study we investigated whether these stressors also contributed to work-family conflict.
Interestingly, research conducted with female Israeli teachers found that novice teachers experienced more work-family conflict than experienced teachers (Cinamon & Rich, 2005). Besides studies on burnout conducted in Kenya, no study has attempted to relate years of teaching experience to the work-family conflict in Kenya. Apart from specific stressors, general and family stressors also influence work-family conflict. Cinamon and Rich (2005) found that in Israel, the number and ages of children were stressors. However, this was not the case in Enugu, Nigeria, due to the support from extended family and domestic workers (Okonkwo, 2014). We investigated stressors and how they contributed to work-family conflict among female urban teachers. The study was guided by two research questions: • Research question 1: From the multiple roles that female teachers perform, what are the teachers' stressors in Kenya? • Research question 2: What is the relationship between these stressors and work-family conflict?
These questions were answered using both quantitative and qualitative data.

Sample and Procedure
In this study I used a mixed method design. Quantitative and qualitative data were collected using closed and open-ended questions. The four study areasthe three towns of Machakos, Wote, Eldoret, and Nairobi citywere purposively selected, but schools were randomly chosen. Female teachers teaching at primary and secondary public urban schools and who had at least one child in primary school and below, were recruited to participate in the study. Data were collected using a piloted questionnaire distributed to participating teachers via a designated contact person in each of the schools. Of the 472 questionnaires, 375 were completed in full. Scales that measure work-family conflict (developed in prior studies but adapted for this study) were included in the closed-ended questions. This paper was compiled from the author's doctoral dissertation, and only data relevant to this paper is included. Results from the closed-ended questions triangulated that of the open-ended questions. Results in the qualitative section are supported with verbatim quotes of the participants' responses. I sought consent for their participation from the specific teachers and approval from the school administration. This study was also approved by Arizona State University and the National Council of Science and Technology in Kenya.

Measures
I adapted work-family conflict scales from Carlson, Kacmar and Williams (2000). These scales consist of two constructs: work interference with family (WIF), and family interference with work (FIW). The two constructs are further divided into four sub-constructs: time-work interference with family -time WIF (α =.83), strain-work interference with family -strain WIF (α =.77), time-family interference with worktime FIW (α = .74), and strainfamily interference with work -strain FIW (α = .87). WIF statements like "My work keeps me from my family activities more than I would like" were included in the questionnaire and were assessed on a 5-point Likert scale, where 5 represented strong agreement, and 1 strong disagreement.

Data Analysis
I used the IBM Statistical Package for the Social Sciences (SPSS) version 20 to analyse quantitative data from the closed-ended questions. I tested the scales for reliability using Cronbach alpha and all the work-family scales had a score of 0.7 and above, which is recommended. I combined the different construct with a number of Likert-type questions into composites variables and calculated their measures of central tendency. I further conducted a Pearson correlation (see Table 1). I used the variables which had significant values in the regression model. I used content analysis and inductive coding proposed by Tracy (2013) to analyse the qualitative data. I used 40 respondents for each question to develop the initial categories, which I used to code the rest of the data. Tracy recommends two coding cycles. The first cycle is to describe data, while in the second cycle is to use the identified codes, organise them into categories and identify the underlying relationships. Furthermore, I sought thick descriptions to support the categories. My doctoral supervisors supervised the process to ensure its credible.
The quantitative and qualitative results triangulated each other.

Quantitative Data
The results of correlations between variables are depicted in Table 1. Time WIF correlated with the number of children in the household (r = .13, p < .05); time FIW correlated with time taken to school (r = -.13, p < .05) and urban area (r = .14, p < .01). Strain WIF correlated with urban area, (r = .12, p <.05), number of children in the household (r = .11, p <.05) and age of the oldest child (r = .11, p < .05). Finally, strain FIW correlated with urban area (r = .12, p < .05). For regression results, I used the Enter Regression Method, with the aim of exploring the relationship between antecedents/stressors. There were four dependent variables: time FIW, time WIF, strain WIF and strain FIW. The predictors were: urban area (the work location, whether town or city), marital status (married vs single), age of the participant, number of children in the household under 18 years, age of the oldest child, number of years of teaching experience, time taken to work (in min), average class size, number of hours spent at home doing school-related work in a week, and whether the teachers had a domestic worker or not.

Qualitative Data
I posed the following question to explore the stressors that teachers experienced and how they contributed to work-family conflict: "Describe in detail the challenges/difficulties you encounter every day as you combine schoolwork, childcare and housework responsibilities." Analysis of this question generated three primary categories: 1) manifestation of work-family conflict, 2) stressors, and 3) effects of work-family conflict (see Table 1).

Manifestations of work-family conflict
The qualitative findings show that teachers experienced work-family conflict and time pressures in their roles at work and at home. Additionally, teachers experienced strain due to home-related roles as well as work-related roles (see Table 3).

Stressors
The main stressors were due to a lack of adequate support at home. Another stressor was interruptions in the normal flow of work and family schedules, which meant that the teachers had to seek a new work rhythm. These interruptions could be due to a child falling ill, long commutes, the teacher attending her own child's school-related activities. such as games (see Table 4). There were 400 mentions of the effect of work-family conflict. Table 4 below shows the sub-categories, concerns raised, the percentage of mentions, and examples.

Effects of work-family conflict
This category had three sub-categories: the effects of work-family conflict at home, at school and both at home and school. It led to concerns on how children were raised, the level to which housework was attended to, and marital satisfaction. The effects of WFC at school led to a poor working relationship with supervisors and co-workers, as the teacher was considered less committed to her work. Generally, the teachers reported low productivity at school, increased incidences of being late, and absenteeism. Teachers felt that they did not excel in their roles at home or at school. There were 200 mentions of effects of work-family conflict (see Table 5).

Summary of Findings
The findings show that the key stressors that predicted WFC were location of the school, the number of children in the household, the time spent to school-related work at home during the week, and average class size. Marital status, the age of participant, the age of the oldest child, years of teaching practice, and the presence of a domestic worker were not significant regression predictors of WFC.
In the qualitative component, women experienced time and strain pressures. The main stressors were unreliable domestic support, interruptions in the flow of work and home routines. These time and strain pressures were due to a lack of finances, a lack of enough sleep, the workload at home and at work. The teachers felt that WFC led to concerns about their efficiency and effectiveness at home and at work. The location of the school contributed to WFC, a finding similar to that of Epie and Ituma (2014). Large cities have more traffic congestions forcing parents to leave very early and come home very late (Muasya, 2014(Muasya, , 2016. Furthermore, unlike in rural areas, support from extended families is missing in large cities and cannot be relied on. The age of the oldest child and years of teaching experience did not predict WFC, which differs from the findings of Cinamon and Rich (2005). There is no difference in WFC between married and single teachers. However, women experience more time pressures at home than at work (Greenhaus & Beutell, 1985).
The analyses of the quantitative and qualitative data yielded inconsistent and somewhat contradictory results. The qualitative analysis revealed more and different stressors. Teachers indicated the long distance to work, time, a lack of finances, un-reliable domestic help and sick children as stressors. In fact, the unreliability of domestic support was the most cited stressor in the qualitative data but it was not a significant predictor in the quantitative data. Furthermore, the lack of spousal support, which is cited in many Western studies (Cinamon & Rich, 2005), was not a stressor at all in either the quantitative or the qualitative data. Even the lack of extended family support was not mentioned at all in this study. However, the negative effects of WFC such as reduced productivity at work, strife with supervisors, tardiness at work, marital conflict, children concerns, fatigue, et cetera, agree with meta-analyses of WFC in Western studies (Allen et al., 2000;Eby et al., 2005).
The age of the children was not a concern (Aryee, 2005; Huffman, Culbertson, Henning & Goh, 2013) in the quantitative data, but the number of children was a significant predictor of WFC. However, in qualitative data the welfare of the childrengood health, proper feeding, and development of good moralswas a stressor. Some of the illnesses were linked to poor feeding. Likewise, high domestic expectations of the female led to high workloads at home. Muasya (2017a) found that immigrant women from sub-Saharan Africa in the United States of America (USA) had high expectations of cleanliness at home, which forced them to constantly clean. However, with time, this expectation waned as they acculturated in the USA.

Contribution to Theory, Practical Implications, and Direction for Future Research
The results seem to suggest that work-family conflict is a universal construct. That is, women reported experiencing time pressures and strain as they attempted to balance work and family demands, which agree with results from studies con-ducted in Western and non-western contexts (Cinamon & Rich, 2005;Eby et al., 2005;Epie & Ituma, 2014). However, what drives the conflict, the stressors, could be different or have different impacts on teachers in Kenyan and Western contexts due to different cultural norms. Role theory states that cultural expectations can lead to role conflict (Kahn et al., 1964).
In this study we found that stressors were influenced by cultural norms and socio-economic conditions, such as the lack of support by domestic workers. The lack of reliable domestic support was the most reported stressor among the teachers. This could be because domestic workers substituted the collectivistic support from females within the extended family who shared childcare and household chores. This was not reported as a stressor in Western research. However, a lack of spousal support was not considered a stressor in this study. This is probably because in the Kenyan culture men are not socialised to do housework. In the Western culture the lack of spousal support is regarded as a stressor (Cinamon & Rich, 2005).
Despite the study focusing on female teachers with young children, the qualitative component of the study showed that women did not report the presence of children or the age of the children as a stressor, as suggested in previous Western research, but rather the well-being and the needs of the children, for example good health and good moral upbringing were the stressors. Generally, anecdotal evidence in Kenya indicates children are valued and regarded as a sign of wealth and prosperity. This study implies that the culture of the place dictates what are regarded as stressors and their impact, which agrees with role theory. Role theory postulates that social expectations may influence the role conflict experienced by an individual. Some stressors were influenced by the socioeconomic conditions in which these women found themselves, which are beyond cultural norms. Examples are the lack of adequate finances, and long commuting distances and time, which may also be experienced by other low-income families in other parts of the world. In Western contexts the lack of finances impacts working-class women more than those in the middle class (Hennessy, 2009). This implies that the working class in the West could face the same experiences as these teachers. This is a limitation of the role theory as it focuses on the socio-cultural environment and ignores the economic aspects.
This was a seminal study and suggests several directions for further research. To start with, longitudinal studies with more representative data should be conducted using both qualitative and quantitative methods. This research showed that there could be a different set of stressors not fully explored in the Western literature, which causes work-family conflict in Kenya. Thus, a qualitative study should be carried out as a follow-up study which includes the perceptions of other school stakeholdersmanagement, ministry of education, other female teachers with no children, and male teachers.
This study can have some application outside sub-Saharan Africa and the teaching profession. For instance, due to the changing nature of work in the globalised economy, it follows that the work environment will keep on changing with more work demands (Bamberger et al., 2015;Moen et al., 2013). This will make some groups of workers more vulnerable to work-family conflict than others. For instance, in many parts of the world, old people were taken care of mostly by women within the extended family. As more women migrate to urban areas and join formal employment, the elderly could become another group of people that may need to be cared for. Currently, this is a challenge in the Western world but has not yet become an issue in sub-Saharan Africa and Asia (Allen et al., 2000;Eby et al., 2005). Employees who need to take care of elderly parents will need to be accommodated within the current work structures.
Similar parallels of this study can be drawn with regions that have relied on extended families for childcare responsibilities such as China and Asia. As more women in these regions take formal jobs outside the home, the workplace will need to accommodate them and offer measures to reduce work-family conflict as noted among mothers in Malaysia (Mahpul & Abdullah, 2011) and China (Xiao & Cooke, 2012).
This study has policy implications for schools and society in general. Childcare remains a dilemma for young working mothers in urban areas who rely on domestic workers for childcare support. Thus, a need exists for alternative childcare support systems to solve these dilemmas.

Limitations of the Study
There was a possibility of common variance method bias. However, the study included a survey with open and closed-ended questions, which reduced these effects. Another limitation was that the study focused on a few towns and Nairobi city in Kenya, which may not be representative of experiences of teachers in other towns and cities in Kenya and elsewhere, limiting generalisation of the findings.

Conclusion
In this study we sought to discover the stressors that teachers with young children face and how they influence WFC. From this study, it is apparent that WFC is a universal stressor and that role theory can apply to non-Western contexts. This study reveals that the level of conflict experienced at urban and rural school locations was different. Additionally, some stressors such as general workload at home and at work are universal, while others such as the age and number of children or their wellbeing are culturally influenced.