Work–Family Balance Choices of Women Working in Kenyan Universities

The proportion of women working in the formal sector in Sub-Saharan Africa has increased in recent years. The kinship networks are weakening, leading to a decline in the traditional forms of support for child care and housework. This study examined the work–family balance options of women working in Kenyan universities within the context of changing national domestic workers’ legislation. Data were collected by use of surveys in two universities. Results showed that as the cost of hiring domestic workers increased, women became indifferent in their choice between employing domestic workers and using daycare centers. Women with older children who employed day domestic workers were more likely to use daycare centers than women with younger children who employed live-in domestic workers. Women with young children in preschool and primary school found their universities less accommodating in helping them balance work and family demands. Employers perceived that the domestic workers’ legislation led to a drop in morale among domestic workers, and demands of pay raises as they became choosier and more inclined to search for better paying employers. It also resulted in a shift of work–family balance strategy for women who opted to hire domestic workers on an “as-needed” or “weekend basis.” Some women stopped hiring them altogether and instead started taking their young children to daycare centers. Cost and affordability determined the use of domestic workers. These women suggested that their employers should increase their job flexibility and put up subsidized daycare centers.


Introduction
Universities are an important source of formal employment for women in Kenya. Most universities in Kenya are located in towns and cities. The number of universities in Kenya has grown since the country's independence in 1963, from one to 22 public universities, 29 private universities, and 13 university substituent (affiliated) colleges by June 2013 (Commission for University Education, 2013). Transportation in most Kenyan urban areas is a challenge; most employees rely on public transport to travel to and from work. The road infrastructure in most urban areas in Kenya is poor and prone to vehicle congestion, especially during rush hours. Thus, employees have to leave their houses for work at dawn, and usually arrive after 6:00 p.m. in the evening due to traffic delays (Muasya, 2014). This has exacerbated the work-family challenges that women face, particularly those who have young children. These women have to leave for work before the children wake up; by the time they return from work, they find the children already asleep.
The relatively inexpensive extended family support for childcare in Kenya and other Sub-Saharan African countries has been declining; with urbanization, kinship networks that women have typically relied on for child care support are weakening rapidly (Miller, Gruskin, Subramanian, Rajaraman, & Heymann, 2006), eventually diminishing the traditional family support for child care (Cassirer & Addati, 2007;International Labour Organization [ILO], 2004). Traditionally, in Sub-Saharan Africa, women relied on other women in the extended family and close neighbors for childcare and housework support. As families pursue formal employment in the urban areas, they lose or cannot afford work-family support, and they adopt the nuclear model of family (Aryee, 2005;Noyoo, 2014). Most women in the formal sector are part of dual working couples and have fixed job schedules that go from 8:00 a.m. to 5:00 p.; therefore, many families hire domestic workers (Muasya, 2014).
However, in the last few years, there has been worldwide tightening of legislation with regard to the employment of 630772S GOXXX10.1177 1 Arizona State University, Tempe, USA domestic workers. This is after ILO (2011) passed Convention 189 on "Decent Work for Domestic Workers." Although currently Kenya is not a signatory of Convention 189, it now enforces the Employment Act (2007), which stipulates the minimum wage and other statutory dues for domestic workers. This came as a result of landmark victory of Robai Musinzi v. Safdar Mohamed Khan in 2012(Kenya Law Reporting [KLR], 2012. This has made the hiring of domestic workers expensive in big towns. In this article, I examine first the work-family balance options of women working in Kenyan universities regarding their use of domestic workers, and daycare centers in the context of tightening domestic workers' legislation. Second, I examine the extent to which university employers assist their workers in balancing work and family demands. Clark (2000) defined work-family balance as "the extent to which individuals are equally engaged in and equally satisfied with work and family roles" (p. 513).
This study was motivated by my comparative experiences of working and studying in three universities in different countries. The first university was in Zimbabwe, where as a faculty member, I lived on the university premises and relied on the university shuttle for transportation to and from house to work. I had my two children in this country. I had a flexible work schedule and easy access to transport to go back home during the day. Zimbabwe had a breastfeeding policy that required institutions to allow mothers to leave for home early to breastfeed their children, if necessary. I employed a day (domestic) worker to take care of my children when I was away.
While my children were still young, I moved back to Kenya, my home country, where I worked as a faculty member at a private university in Nairobi. In the new university, I had no access to university housing, and I had to commute a long distance from my rented house; thus, every day I had to endure the inconvenience of traffic jams, typical of Nairobi, Kenya's largest city. I had to leave for work at dawn to arrive at my workplace early and returned home late in the evening; by the time I got home, my children were asleep. Here, I had less job flexibility, and so I employed a live-in domestic worker to assist me.
My third university was in the United States, where I was a doctoral student. In the United States, childcare services are unaffordable to most graduate students with children, particularly those from third world countries. Thus, I could not afford daycare services or a nanny to look after my young children. I had to share childcare responsibilities with my husband. These experiences stirred a desire in me to investigate how other Kenyan women working in universities juggle between work and home responsibilities.

Overview of Kenya Work-Family Balance Practices
A recent study called on Kenya's employers to institute policies that accommodate employees faced with work-family balance conflicts (Strathmore Business School, 2011).
Organizations in Kenya have not instituted telecommuting options that allow employees to work from home, or policies that permit work flexibility. Kenyan employers still have a long way to reach international standards in enacting familyfriendly work policies and practices (Strathmore Business School, 2011).
Family-friendly work policies are "a formal or informal set of terms and conditions which are designed to enable an employee to combine family responsibilities with employment" (Simkin & Hillage, 1992, p. 13). These policies are subdivided into three categories: (a) leave arrangements which include maternity leave, paternity leave, and compassionate leave; (b) flexible working arrangements such as part-time arrangements and/or a compressed workweek; and (c) workplace facilities such as subsidized childcare, crèches, counseling, and so on (Simkin & Hillage, 1992).
In Kenya, the 2007 Employment Act enforces statutory leaves such as 3 months paid maternity leave [Section 29(1)], 2 weeks paid paternity leave [Section 29(8)], 21 days paid annual leave [Section 28(1)], and at least one day off each week [Section 27 (2)]. The provision of flexible working arrangements and workplace facilities such as crèches are left at the discretion of the employer; there is no special provision for child care in the Kenya labor laws (AfricaPay.org/ Kenya, n.d.; Employment Act, 2007).
My assumptions about the causes of this limited support for family-friendly policies are as follows: (a) urbanization (and resulting decline of traditional extended family support) is a recent experience in Kenya; (b) work-family research is still at a nascent stage in Sub-Saharan Africa. Most workfamily studies today are based on Western contexts and this gives the false impression that work-family balance issue is not a problem in Sub-Saharan Africa but rather a Western issue. Therefore, organizations and government departments lack sufficient evidence to create and implement work-family policies (Aryee, 2005); (c) most studies in Kenya have assigned work-family balance issues a peripheral position as one of the factors that cause career ceiling in universities for women (Onsongo, 2006) and hinder women in political ascendancy (Kiamba, 2008) and managerial ascendancy (Lyness & Thompson, 2000;Mangatu, 2010). Research that focuses on work-family balance issues of female university employees contributes to knowledge on how institutions are shaping the work-family outcomes for women and families.

Work-Family Conflict (WFC)
Due to limited research from Kenya, I will draw most of my literature review from research in Western countries. Previous research (and experience) shows us that as employees seek to meet both family and work obligations, they often experience WFC. WFC is "a form of interrole conflict in which the role pressures from work and family domains are mutually incompatible in some respect" Greenhaus & Beutell, 1985, p. 77). Work can interfere with family responsibilities-WFC-or family may interfere with work responsibilities-FWC (family to work conflict). Previous research findings show that WFC is more prevalent than FWC in Western countries (Greenhaus & Beutell, 1985). The opposite of WFC is workfamily balance, where the conflict between the two spheres of home and work is minimized and is described by Clark (2000) as "satisfaction and good functioning at work and at home, with minimum role conflict" (p. 751).
It is a challenge for many employees, especially female employees, to achieve a balance between work and family roles. Most of the studies in work-family balance and WFC are based on role theory (Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964). The role theory assumes that the work and home spheres are separate domains, and each has multiple roles. The roles are influenced by the norms and expectations of the society or organizations, and each individual knows the societal expectations of what he or she should do (Biddle, 1986).
Thus, if an individual does not conform to these expectations or norms, it can lead to role conflict. WFC is also informed by the scarcity hypothesis which assumes that individuals have finite resources in terms of energy and time and trying to fulfill them leads to conflict (Goode, 1960). WFC can be time based, strain based, or behavior based (Greenhaus & Beutell, 1985). Time-based conflict occurs when participation in one domain makes it hard for an individual to participate in another domain; for example, this conflict may arise due to number of hours worked and inflexible work schedule (Aryee, 2005).
Strain-based conflict occurs when the strain in one domain makes it hard to fulfill the demands of the other domain. It can be caused by stressors such as role overload, role ambiguity, and lack of support at work and at home. Finally, behavior-based conflict occurs when the behaviors of one domain are not appropriate in the other domain. For example, most work contexts do not permit emotionality that is experienced (and accepted) in family home contexts.
WFC has been associated with negative organization outcomes (Greenhaus & Beutell, 1985). For instance, there is a negative relationship between WFC and job satisfaction. WFC was negatively associated with work-related outcomes and depression, which led to substance abuse such as alcoholism, decreased marital satisfaction, decreased job and family satisfaction, and an intention to quit work (Eby, Casper, Lockwood, Bordeaux, & Brinley, 2005). Studies have shown that academic female employees experience more stress from work-family-related issues than their male counterparts (O'Laughlin & Bischoff, 2005); for example, female employees at Michigan State University reported that they could balance work and family only at the expense of their sleep (Damiano-Teixeira, 2006). What about Kenyan women?

Effects of Legislation
Most families employ domestic workers as a work-family balance strategy in urban Kenya. The use of domestic workers (house helps) depends on their reliability, that is, availability when needed, and affordability (De Regt, 2009). In 2011, the ILO passed Convention 189 (Decent Work for Domestic Workers). This convention required countries to set minimum wage and ensure access to health, paid leave, and other social benefits to domestic workers. To be in line with these requirements in 2011, the Kenya government issued legal notice No. 64, which stipulated the minimum wages for domestic workers (Kenya Subsidiary Legislation, 2011). Although these provisions were already in the Kenya employment Act 2007, they were hard to enforce due to the informal nature of the work contract between domestic workers and their employers. But with the landmark case of Robai Musinzi v. Safdar Mohamed Khan (KLR, 2012), the industrial court in Nairobi ruled that verbal contracts were binding. In this case, the domestic worker Robai Musinzi was compensated for unlawful dismissal and termination of employment. These developments led to tightening of the legislation in regard to domestic workers and increasing the cost of employing domestic workers, as it enforced minimum wages as well as other statutory dues such as social security dues (KLR, 2012).
These new statutory regulations were met with mixed reactions in Kenya (Muiruri, 2011). Some people thought it would result in high unemployment among domestic workers (Kariuki, 2011) as families might not be able to afford them in the first place (Karambu, 2011); others speculated that new work-family arrangements would emerge to fill the gap (Juma, 2011). This article seeks to investigate the effects of the changes of labor laws on the use of domestic workers among Kenyan university women with children.

Role of Workplace Support
Apart from government policies, organizational policies can affect work-family balance outcomes of families too (e.g., life satisfaction). For example, research shows that workfamily-friendly policies result in committed workers who are loyal to the organization (Chiu & Ng, 1999), as these types of policies reduce employee absenteeism, stress, WFC, and high turnover (Halpern, 2005). However, research also suggests that although an organization offers family-friendly policies, employees may refrain from using them (e.g., parental leave) because they think it may negatively affect their career prospects (Lyness & Thompson, 2000).
Therefore, for an organization to be viewed as supportive (beside the family-friendly programs), it must also show its commitment in how much it cares about an employee's family. A supervisor should understand when an employee is facing a family crisis. They should create an atmosphere where employees can freely talk about their personal and family needs, especially if they are direct supervisors, as the welfare of the family is in their hands (Anderson, Coffey, & Byerly, 2002). The results of a recent study investigating the relationship between work-family-friendly programs and commitment to work and work withdrawal revealed the importance of transformational leadership in organizations. The study found employees were resentful toward a supervisor who did not support the use of work-family-friendly benefits, and showed more commitment if they had supervisors who were empathetic, inspiring, and challenging (Wang & Walumbwa, 2007).
Apart from supervisors' support, many organizations in Western countries have implemented programs to improve the well-being of their workers; these include job sharing, job protected parental leave, daycare centers, dependent care assistance, telecommuting, part-time return to work options, canteens, sport facilities, laundry facilities, and unpaid family leave (Lobel & Kossek, 1996).
Some studies have found a positive relationship between supervisor, organizational support, and employee satisfaction with work-family balance and thus encourage implementation of family-friendly policies and practices (Ezra & Deckman, 1996). However, there are few studies in the banking sector; the exception is Wang and Walumbwa (2007), who compared family-friendly work policies and transformational leadership, among bank employees in Kenya, Thailand, and China. There are no case studies of the role of institutional support in the Kenyan universities, and therefore this study will be exploratory in nature.
The current study seeks to identify the effects of change in domestic worker legislation and lack of institutional policies (programs and supervisory) in work-family outcomes of women working in universities in Kenya. The study posed the following research questions: Research Question 1: What factors influence women's choice of domestic workers versus daycare centers as a work-family balance strategy? Research Question 2: What are the effects of domestic workers legislation on the use of domestic workers as work-family strategy? Research Question 3: How accommodating are Kenyan universities in work-family policies?

Participants and Procedure
This article uses self-report data from a larger study carried out in one public and one private university in Kenya, where 100 questionnaires were distributed to female employees with children in July 2011. The surveys were given to key contacts in the departments to distribute to their colleagues. Seventy surveys were returned completed, with a completion rate of 70%. Two sevenths of the participants were from a private university, and five sevenths from public university 1 completed the surveys. The study was approved by the Institutional Review Board at Arizona State University and approved by the Kenyan National Council of Science and Technology (NCST). Participants were women with children in primary school, and below.

Measures
The study used a survey which had both closed-ended questions (quantitative) and open-ended questions (qualitative). The survey scales were generated by the researcher, have not been used in prior studies, and may not be generalizable to other sectors. The sample included women of all pay scales from administrative and academic divisions of the university.
Participants' demographics. Demographic questions asked for participants' age, number of children and their ages, department where they work, years of work experience, marital status, and whether they employ a domestic worker, and if so, the domestic worker living arrangements (live-in vs. day worker).
Satisfaction with domestic worker. This was measured by a researcher-developed five-item scale listed below, where participants rated each item on a Likert-type scale of 1 = strongly agree to 5 = strongly disagree. The internal consistency of domestic worker satisfaction scale was α = .75. Daycare services. This scale measured the extent to which participants preferred the use of daycare center versus use of domestic workers for care of young children. It had five items, measured on a Likert-type scale of 1 = strongly disagree to 5 = strongly agree. Some items were reverse coded. The internal consistency of the scale was α = .67.
Namely, 1. Taking a child to the daycare is the same as leaving him or her with the domestic worker; 2. Daycare provides an atmosphere like home; 3. Prefer leaving the child at home with domestic worker rather than in a daycare; 4. Daycare is more expensive than the use of domestic worker; and 5. Daycare is more reliable than domestic workers.
Legislation effects. This was measured with three questions assessing the level of importance that participants assigned to the impact they experienced from the implementation of government requirements for domestic workers. These were as follows: (a) my ability to pay the domestic worker minimum wage, (b) my ability to pay the domestic worker overtime, and (c) loss of flexibility in my domestic worker work arrangements. Participants were asked to rate the importance of each effect on a Likert-type scale, ranging from 1 = very unimportant to 5 = very important. This scale had internal consistency of α = .70.
Accommodation. This measure sought to find out the extent to which the institution accommodated (helped) the employees to meet their work-family balance issues. Participants were asked to indicate their responses on a Likert-type scale, ranging from 1 = very unaccommodating to 5 = very accommodating. The internal consistency of this scale was α = .82, and the items included the following: (a) the institution has daycare facilities, (b) provide teaching/work schedules that accommodate women with young children, (c) allow me to bring my child to work if necessary, (d) allow me to swap schedules with my colleagues, and (e) provide a flexible working schedule.

Quantitative Data
Statistical analyses were conducted using SPSS-21 software. I used descriptive and regression analysis for the data generated from responses to the quantitative questions.

Qualitative Data
To capture the opinions of these Kenyan women more in depth, I included two open-ended questions in my survey.
The first question asked participants to describe the challenges they experienced since the introduction of the domestic workers legislation. The second question asked participants about their opinions on ways their employer (university) can improve the female work-family balance issues. It should be noted that, as I had worked in Kenyan university before, I had to exercise self-reflexivity in designing and implementing the study (Leavy, 2014). What we study tends to be influenced by our predispositions and values, and in designing and implementing my study, I had to bracket my experiences and tacit knowledge about work-life balance issues (Parse, Coyne, & Smith, 1985). One way to get the participant's point of view is by use of open-ended questions in which the participants described their own experiences, and these are captured using thick descriptions (Geertz, 1973).
I used inductive data analysis methods to analyze the responses to the open-ended questions (Charmaz, 2006). That is, I read the participant responses many times and conducted open coding on the data (Glaser & Strauss, 1967). I searched for common themes among the responses in relation with the study objectives to create initial categories (Glaser, 1978). I then used constant comparison method to refine the initial categories (Charmaz, 2006;Tracy, 2013). The results of these analyses are presented in the following section. Constant comparison involves comparing incident by incident to identify dimensions and properties in the data (Glaser & Strauss, 1967). After developing my initial categories, I shared them with my academic advisor for review. I did conceptual mapping to identify the core categories (Charmaz, 2006;Strauss & Corbin, 1998).

Results and Discussion
Participants' Characteristics Table 1 shows the correlations between the study items, means, and standard deviations. The percentage of women aged 40 years and younger was 41.2%. A significant proportion of women-71%-had a spouse. They had worked in the university for an average of 11.89 years (SD = 1.51), and 55% lived in their own homes. The average number of children per female was 3.09 (SD = 1.51). Most participants-53 (83%)-employed a domestic worker, of which 75% were day workers. The remaining 12 (17%) did not have a domestic worker, but a majority had employed a domestic worker in the past. Their children were slightly older, past daycare age.

The Use of Domestic Worker Versus Daycare Services
The first research question sought to identify the factors that influence women's choice of domestic workers versus daycare centers as a work-family balance strategy. To answer this question, I did two regressions on "daycare services" and "satisfaction with domestic worker". In addition, I analyzed the responses on the first open-ended question in which the participants described the challenges they had experienced since the introduction of the domestic workers' legislation.
Daycare services. Use of daycare services was negatively correlated with the living arrangement of the domestic worker (r = −.31, p < .05) and the age the participant was comfortable to take her child to daycare center (r = .49, p < .01). That is, women with live-in domestic workers and with younger children were less likely to use daycare centers. Possible reasons were that participants were more comfortable taking older rather than younger children to the daycare centers. Furthermore, women with day workers were more ready to take their child to daycare than those with a live-in domestic worker.
A hierarchical regression was conducted after controlling for marital status of the participants. The age women were comfortable to take their young children to daycare, and domestic worker living arrangement, contributed significant variance to daycare services. The regression of total score for daycare services on marital status was not significant, R 2 = .05, F(1, 35) = 1.69, p = .20; in the second model, two variables-the age the participant was comfortable to take her child to daycare center, and living arrangement of domestic worker-accounted for significant variance on the use of daycare services, ΔR 2 = .28, F(2, 33) = 5.31, p = .01. The mean for reference group-the day workers was 3.80 and that of live-in domestic workers was 3.04-was significant (β = −.38, p = .03); the overall regression model was significant, R 2 = .28, F(3, 330) = 4.24, p = .01. Day worker and live-in domestic workers' living arrangements offers different experiences. Employers of day workers are more satisfied with daycare services than employers of live-in domestic workers. The older the child, the more women would be comfortable to take him or her to daycare services.
Satisfaction with domestic worker. Satisfaction with domestic worker was positively correlated with domestic worker duties (r = .59, p < .01) and the length of time a domestic worker stayed with the respondent (r = .31, p < .05). This means that the more a female employer was satisfied by the domestic workers' services, the more likely she was to retain the domestic worker. Participants in the study were more satisfied with domestic workers who stayed a longer period than those who stayed for a short period. On average, a domestic worker stayed for 3 years (M = 3.06, SD = 2.46). This suggests that for many, being a domestic worker was a temporary job.
Besides daycare services, a hierarchical regression was conducted, after controlling for marital status, to determine whether the domestic worker length of stay and domestic worker duties had significant variance in satisfaction with the domestic worker. The regression of the total score for satisfaction with domestic worker on marital status did not have significant variance, R 2 = .001, F(1, 45) = 0.04, p ≤ .84. Inclusion of domestic worker's length (centered) and domestic worker duties accounted for significant variance in domestic worker satisfaction ΔR 2 = .30, F(2, 43) = 9.23, p < .001. The final regression model was significant, R 2 = .30, F(3, 43) = 6.18, p = .001. The domestic worker length centered (β = .30, p = .03) and domestic worker duties (β = .43, p = .003) were positive predictors of domestic worker satisfaction.
The respondent satisfaction with domestic worker depended on the length of time the worker stayed with her. For instance, if the domestic worker did not stay long, the participants expressed lower satisfaction with the domestic worker. Besides the length of stay, participants were satisfied with the domestic worker if she performed her assigned duties well. Anecdotal evidence shows that high turnover of domestic workers may affect the participants' performance at her workplace as well.
The study found that some attributes of the domestic worker affected the use of domestic worker as a work-family balance strategy, including the length of period a domestic worker was willing to stay, where she resided, and her work performance. Older research participants tended to keep the same domestic worker longer, either perhaps because they have some experience with handling domestic workers or they have no young children (anecdotal evidence seems to indicate that domestic workers prefer women with older children). The execution of domestic worker's duties was correlated with satisfaction with the domestic workers (r = .59, p < .01).
To summarize, the factors that determine the satisfaction in the use of a domestic worker are: the length of time she was willing to stay and how well she executed her duties. These factors relate to the issue of worker reliability, a major concern in the use of domestic workers. De Regt (2009) found that Yemeni women preferred foreigners as they were more reliable than the locals or Somali women. This is because the foreigners stayed longer, as their families were far away. In addition, the locals had their own families to care for. In case of emergencies, this could cause inconvenience for the families who employed them.
The factors that influence the mothers' decisions to use either daycare or domestic worker were the age of their child and the living arrangement of the domestic worker. Employers of day workers were more willing to use daycare services than employers of stay-in domestic workers. In addition, women were more willing to use daycare services if the child was older (M = 2.53, SD = 1.19).

Effects of Legislation on Use of Domestic Workers as Work-Family Balance Strategy
To answer Research Question 2, about the effects of changes in legislation on the use of domestic workers, and work-family strategy, I did one regression on effects of legislation and also analyzed responses to the first open-ended question. This question asked participants to identify challenges they faced after the enforcement of minimum wage, and social security payments in regard to employment of domestic workers.
Legislation was positively correlated with the living arrangement of the domestic worker (r = .33, p < .05) and the age women are comfortable to take their children to daycare center (r = .31, p < .05). A hierarchical regression was carried out to find out if, after controlling for marital status, domestic worker living arrangement and daycare age (both centered) would account for significant variance in effects of legislation. The regression of the total score for legislation effect on marital status was not significant, R 2 = .01, F(1, 35) = 0.17, p = .68. The inclusion of domestic worker living arrangement and daycare age did not have significant variance in legislation effects, ΔR 2 = .06, F(2, 33) = 2.58. The overall regression model was not significant, R 2 = .14, F(3, 33) = 1.78, p = .17. In the next paragraph are the results of the analysis of responses to the first open-ended question.
Fifty-three women responded to the first qualitative question which asked about the challenges faced by women after the changes in domestic workers labor laws. I identified four categories of responses: affordability, domestic worker performance, change in work-family strategy, and none (see Table 2). In the first category of affordability, participants raised the concern of whether they would be able to afford the new domestic worker salary requirements due to their own salary constraints. As one participant noted, "I am unable to employ because my salary can't be able to maintain her." Some of the participants in this category felt the salary demanded was too high when they factor in food and lodging. Instead, the salary should depend on the domestic worker's job description and work performance. Thus, these women were trying to come to terms with the new wage requirements, although there were participants who hoped to be able to pay domestic worker the minimum wage because it was a humane thing to do.
The second most frequent response was that some women opted for a change in work-family balance strategy, such as coming up with different domestic working arrangements. For example, they hired domestic workers on an as-need basis-maybe just for weekends. As one participant commented, "I pay my domestic worker as per visit and she has not complained about the wages." Some participants had stopped hiring a domestic worker especially if they did not have young children, whereas other participants changed their child care practices.
One respondent was of the opinion that in fact if you have no baby, there is no need of hiring one. Many people are avoiding, instead, they hire them on weekends and pay them. It is expensive, since they eat and sleep in your house. I would prefer to take my small babies to a daycare.
This result confirms the quantitative results that if the cost of hiring domestic worker increased, women became indifferent between use of domestic workers or daycare centers, and those with older children were willing to take their children to daycare centers. A third response to the question was related to the domestic workers' job performance. Some women stated that immediate noticeable effects of the legislation were cases of domestic workers' poor job performance and display of negative attitude, rudeness, and arrogance. Of the five instances describing poor job performance, only one respondent reported that the domestic worker showed respect. Because the domestic workers now had choice, the participants seemed to feel that domestic workers became less serious about their work. As one participant posits, their "morale for work has gone down and always demanding a salary increment," and some started misusing things. Another respondent said that the domestic workers "want the new salary increment plus benefits, it is quite expensive. They keep on shifting for the right employer." Finally, almost a quarter of the women reported no challenges in the working relationship with their domestic workers after the new legislation. For two women, the minimum wage had not been affected in their geographical areas. While for the other 13 women, the domestic workers did not mind the pay they got. Some domestic workers were ignorant of the minimum requirements, whereas for some women the minimum salary was not a big deal for them to pay.
To summarize, despite the quantitative data yielding no significant results, the responses to the first open-ended question show that women had concerns of affordabilityespecially those with lower incomes (my study included university women of all income levels), see Table 2. Some women changed their working arrangement for the domestic worker to come on need basis or weekends. These women reported a sudden change in the way the domestic workers executed their duties. They noticed the domestic workers' sudden drop of morale, and domestic workers became choosier above who to work for. Some employers did not experience any change, perhaps because the domestic worker was uninformed of the legislation or the legislation was not implemented yet in their region or they could afford the stipulated salaries.
These results are confirmed by national news reports, describing how the change in legislation was met with mixed reactions in Kenya (Juma, 2011;Kariuki, 2011;Muiruri, 2011). In the De Regt (2009) study, affordability was also a concern in the employment of the domestic workers.

Female Workers' Opinion on University Accommodation
To answer Research Question 3, I did one regression on "accommodation" and analyzed the responses to the second open-ended question asking respondents to identify ways in which their employer could improve the work-family balance issues of female employees.
Accommodation by the participants' employer (the university) was negatively correlated with the number of children in preschool or primary school (r = −.47, p < .01), negatively correlated with the age of the domestic worker (r = .35, p < .01), and positively correlated with living arrangement of the domestic worker (r = .29, p < .05). This could be because women with children in preschool and primary had the perception that the institution was less accommodating, and the living arrangement of domestic worker influenced the perceptions of accommodation. Possible explanations are: first, if a female employer had a young domestic worker she would still need to supervise the worker more closely than if she had a more mature worker. Second, a live-in domestic worker is more reliable than a day worker, who may not show up in the middle of the week if she has her own emergencies to attend to. Third, female employees with young children are more likely to miss work than those with older independent children.
A hierarchical regression was carried out to determine whether, after accounting for age of the participant and marital status, "living arrangement of domestic worker," "number of children in preschool and primary," and "preferred domestic worker age" accounted for significant variance in accommodation. Day workers living arrangement was the reference group coded 0 and the live-in in housing arrangement coded as 1. The regression of the total score for accommodation on age of the participant, and marital status was not significant, R 2 = .03, F(2, 50) = 0.81, p < .45. Inclusion of living arrangement of the domestic worker, and number of children in preschool and primary school accounted for a significant variance in accommodation, ΔR 2 = .29, F(3, 47) = 6.69, p < .001. The overall regression model was significant, R 2 = .32, F(5, 47) = 4.45, p = .002. The mean of the reference group (day workers, M = 1.82) and live-in in domestic worker (M = 2.45) was significant (β = .26, p = .04). Apart from living arrangements of the domestic worker, the number of children in preschool and primary (β = −.35, p < .02) was a predictor of accommodation, whereas domestic worker's age (β = .23, p = .09) and marital status (β = .04, p = .74) were not. In sum, the degree to which women perceived their employer (university) as accommodating depended on three things: age of the domestic worker, living arrangement of the domestic worker and the age of the children.
Women with live-in domestic workers felt the university was more accommodating compared with those with day domestic workers. Possible reasons are: first, a live-in domestic worker makes the participants worry less about the rigid work schedule or policies at workplace. Second, a day domestic worker schedule may not fit well with participants' schedules for picking up, and dropping off children from school. Third, in case the participant is delayed at work or in traffic jams, which are prevalent in big cities, she may inconvenience the day worker, who is eager to leave for her home.
Women with children in preschool and primary school perceived the workplace as less accommodating. These participants (university employees) may prefer schedules that allow them to leave early or come late to work because they have to prepare children in the morning for school, drop children off at school, and later pick them up and attend to their children's homework.
The second open-ended question asked for participants' opinions on ways the university can improve work-family balance issues for its staff (see Table 3). Forty-five women responded to this question, and the results are presented in Table 3. The participants identified three ways the university could help them better deal with work-life balance: (a) better work organization (58%), (b) providing child care facilities (34%), and (c) a miscellaneous category (8%).
The primary suggestion for how universities can improve female work-family issues had to do with work organization. In this category, 45% of the responses had to do with flexible work schedules. That is, participants suggested that flexible work arrangements would accommodate women with young children-especially reporting and leaving time, and some women proposed a schedule that allows someone to work half days for a year. Another 15.8% of participants saw the need of better understanding between the employer and employee when a need arose. For example, if there was an emergency, the employer might give the employee some time off. Some (21%) participants felt the need of more work leave days, such as increased maternity days and/or unpaid leave to take care of children; 10.5% saw the need of increased pay to be able to afford the minimum pay of domestic workers. Finally, 7.9% had the opinion that they need a lighter workload; that is, employees with young children should be given less work, or have work schedules that do not interfere with weekends off. In case they have to work on weekends, they should be compensated with leave days.
The second largest category had to do with establishment of daycare or baby care facilities at workplace-which 33.8% of participants recommended. As one participant noted, "The employer should have an option of daycare facilities for students and employees who do not have house helps." The third major category was a miscellaneous one comprising various suggestions, including the need to educate female staff about family planning, the need for respecting the female employees, and incentive for good leadership. However, some participants were of the opinion that the responsibility for coping with work-family balance issues lay primarily with the women employees, suggesting that they should be more organized. One participant said, "There is nothing the employer can do about it. Female employees have to work as much as their counterparts. It is for the female employee to get organized and focused." To summarize, university women view their employer as less accommodating if they have younger children than older children. The women prefer to have a more flexible work arrangement, empathetic supervisors, more leave days, lighter load, more pay, daycare centers, and no interference with weekend time off. As noted earlier, Simkin and Hillage (1992) identified these same policies family-friendly work policies. However, some respondents said that women should be more efficacious and organize themselves. I did conceptual mapping (see Figure 1) and combined categories of responses to Research Questions 1 and 2 (presented in Tables 2 and 3) to see how the different categories are related or influence each other, and to identify the core categories (Charmaz, 2006;Strauss & Corbin, 1998). My core category was work-family balance strategy. This strategy was mainly influenced by flexibility and affordability. The ability to afford to pay the increased wages stipulatedby the changes in domestic workers legislation determined whether the woman will employ a domestic worker as a lifework balance strategy or not.
Affordability further influenced the nature of the employment whether on need basis or monthly. A domestic worker opted to work for those women who could pay her more, and thus kept on shifting from employer to employer. The affordability of the domestic worker increased the flexibility of these working university women. If daycare centers are cheaper than domestic workers, the female employer may prefer to use them rather than employ domestic workers.
However, the university can increase the flexibility of their female employees' schedules through better work organization and provision of daycare facilities (for those with very young children). If the employer (university) daycare centers were subsidized, many women could afford them. This will bring a shift from the use of domestic workers to daycare centers.
The results of this study seem to show that the time is ripe for work institutions in Kenya and other Sub-Saharan African countries to rethink ways to make their workplaces more family-friendly. With a backdrop of increased legislation to control minimum wage of domestic workers, and declining support from extended family (Mokomane, 2014), it is time for institutions to rethink national work-family policies including community-based child care practices.
One of the reasons why companies enact work-familyfriendly policies is to increase the employees' flexibility at work and address time-based conflict; if an employee is allowed to reduce their working hours, it will also reduce strain-based conflict (Greenhaus & Beutell, 1985). Epie's and Ituma's (2014) study stressed the need of family-friendly policies in Nigeria among professionals. Similar to Kenya, Annor (2014) observed that Ghana also lacks governmental and employer policies on work-family balance issues.
Thus, beside statutory leaves such as maternity and annual leaves, there are no leaves for parental care. While some universities provide crèches, and schools, the schools are only for school-age children; therefore, both Ghanaian and Kenyan universities need to address this issue. Annor's study found that inadequate pay was a stressor, as families had to invest more hours to supplement their income with other sources. Annor's findings support the current study where women were of the opinion that a higher salary would enable them to afford the legislated domestic worker salary. Finally, there is need for governments to have comprehensive analysis of the effects of some policies. This study shows if the cost of hiring domestic workers becomes prohibitive, it will lead to a shift from the use of domestic workers to daycare centers, assuming the cost of daycare centers remains unchanged. However, if the government fails to enact family-friendly policies legislation, the work-family balance needs of the employees may never be addressed by organizations, in spite of the evidence that organizations can gain competitive advantage if they are rated by the employees as family-friendly. For example, audit companies in the Western countries have used family-friendly and femalefriendly policies to retain their talented women employees in the organization and reduce their turnover (Cohen & Single, 2001;Wooten, 2001).

Conclusion
The use of domestic workers is largely influenced by reliability and affordability. As the cost of domestic workers increases, women shift their work-family practices to using them on an as-needed or weekend basis, and taking their older children to daycare centers. The university employers in these two institutions still lack policies that promote job flexibility or daycare centers which women need to improve their work-family balance practices.
Policies both from the government and work institutions can shape the work-family strategies open to women in the formal sector. Organizations cannot assume that families have adequate resources in the form of domestic workers and extended family to help their workers resolve work-family balance issues. This study is meant to initiate a conversation about family-friendly organizations within the university. Women in Kenya are employed by different sectors, so we need more sector-specific studies to identify the professionspecific stressors and how these sectors can enact familyfriendly policies.
Acknowledgments I thank Dr. Judith Martin of Arizona State University for her valuable guidance in writing this article.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research and/or authorship of this article. Note 1. I was unable to establish the population of women employees in the public university in this study.