“I-just-wanna-get-by” hurts teachers and their work: Linking preschool teacher identity to work withdrawals in an emerging economy

The benefits of quality early childhood education (ECE) for individuals and societies in low- and middle-income countries (LMICs) have been documented in the literature. The benefit-to-cost ratio (that is, the benefits society gains compared to the amount spent) of increasing preschool enrolment to about 25% or 50% in LMICs is estimated to be between 6.4% to 17%, but school attendance is low in these countries, particularly those in sub-Saharan Africa (SSA) (Engle et al., 2011). In Ghana, with technical and financial support from development partners, the preschool enrolment rate is at a record high (75% net enrollment). However, situational reports show decreasing teaching quality and learning outcomes (Ghana Education Service, 2013; Ministry of Education Ghana, 2016). Intervention programs to address quality challenges in Ghana show mixed results (Wolf, 2019; Wolf et al., 2019a, b). And that persistent gains in child learning and development depend on teacher emotional support for children, and teacher burnout (Wolf, 2019).

Considering the criticality of emotional contagion in teaching and learning outcomes (Burić & Moè, 2020; Frenzel et al., 2009; Moè et al., 2021, 2022), teacher burnout, specifically emotional exhaustion/ overextension, is thus central to ECE outcomes. Yet, Ghanaian preschool teachers are emotionally overextended (Aboagye et al., 2018; Lee & Wolf, 2019; Ntim et al. 2023). Poor work conditions (Agezo, 2010; Opoku et al., 2019; Peele & Wolf, 2020), lack of professional development (Bennell & Akyeampong, 2007; Osei, 2006), and lower salary levels (Salifu, 2014, 2015; Sam et al., 2014) have been implicated in Ghanaian teachers’ anxiety, depression, burnout, and attrition. Going beyond the structural conditions of the workplace, this study contends that teacher professional identity offers a critical vantage point in understanding emotional labor, and the consequent emotional exhaustion and withdrawal.

Teacher identity (seen as "socially shared and coherent set of meanings that define the particular professional role of teachers" [Hanna et al., 2020, p. 2]) is socially constituted at both the societal and the organizational level, and therefore becomes an identifiable pre-existing entity (Burke & Stets, 2009; Hanna et al., 2020). Thus, teacher professional identity represents a role identity (i.e., an aspect of a person that assumes and safeguards the role of a teacher [Burke & Stets, 2009]). In effect, who a teacher is and what s/he does, including emotional labor are to a large extent specified by the society and schools. As a result, all that is expected of a teacher can be represented as a psychological construct such as a belief or an attitude, which can induce emotions, generate thoughts, and guide teachers' behaviors (Hanna et al., 2020).

We argue therefore that teachers whose overarching motivation is getting a paycheck in hope of transitioning into a better job (Harris, 2020) without the passion for teaching may struggle to perform the emotional labor of caring for students (Zapf et al., 2021). Emotional labor, defined by Hochschild (1983, p. 7) as “the management of feeling to create a publicly observable facial and bodily display” is a vital part of teaching. Teachers who do not identify with emotional demands on them may surface act (i.e., suppress emotions) to meet display rules. Suppressing emotion on the job would exhaust them over time, leading to work withdrawal. In this study, we consider important teacher work withdrawals (i.e., presenteeism and lateness) in SSA such as Ghana (Fobih et al., 1999). These behaviors produce children’s poor developmental and learning outcomes (Harding et al., 2019). When teachers have passion for teaching and nurturing children, they are more predisposed to use deep acting (i.e., modifying emotions) to deliver high-quality teaching and care, keeping emotional exhaustion, and work withdrawal in check.

In sum, given ECE's stressful work conditions in SSA, teachers' short-term orientation, low motivation, and professional development, we employ the job demands-resources (JD-R) theory to argue that Ghanaian preschool teachers experience ill-being – influencing their work withdrawals, and the consequent child developmental and learning outcomes. That is because, in addition to the structural resource constraints, many of them lack the requisite personal motivational resources (including teacher identity) that nurture the enactment of effective emotion management strategies appropriate for effortless attainment of ECE’s emotional requirements.

Ghana is a Lower-Middle Income West African Country with a population of 31.08 million people (Ghana Statistical Service, 2021). The Government of Ghana (GoG) incorporated ECE into the free Compulsory Universal Basic Education (fCUBE) in 2007. Before then, preschool was the sole responsibility of parents or families (The White Paper, 1995). The Ghanaian preschool was administered on a standard two-year duration, consisting of year 1 (i.e., KG1 for 4-year-olds) and year 2 (i.e., KG2 for 5-year-olds). Given the enthusiasm of the Ghanaian people for education, this development led to a surge in preschool enrolment rates (above 70% year on year since 2011 (UNESCO, 2019)); however, corresponding investments to ensure quality ECE were largely missing. As a result, education quality problems have been recorded. These include poor classroom quality and the consequent abysmal learning outcomes (Ministry of Education Ghana, 2018; UNESCO, 2015).

In 2018, the GoG extended the 2-year preschool duration to 3 years, perhaps in hopes that spending enough time at the pre-primary level could solve some of the challenges later down the line (Ministry of Education Ghana, 2018). But schooling is not learning, and LMICs like Ghana record some of the poorest learning outcomes despite increased enrolment and longer stay in school (Pritchett, 2013). To put in perspective, the author finds a whopping 83% of all 15-year-olds who sat for the 2007 Trends in International Mathematics and Science Study (TIMSS) fell below the minimum score of 400 (the failure rate rises to 98% if the learning standard is raised to 450 which is the minimum score in U.S.). This indicates that Ghana has “a very shallow learning profile” (Pritchett, 2013, p. 80), meaning students make little progress in learning per year of schooling.

The quality challenges pose a significant threat to the benefits of preschool education as a critical period for future academic success, later flourishing socially and occupationally in adult life, and the overall societal development (Angrist et al., 2021). The ECE curriculum developed in 2004 for preschool education captures all aspects of children’s developmental needs, but evaluation results indicate that teachers lack the requisite techniques and motivation to execute the goals of the curriculum (Ministry of Education Ghana, 2016). Though series of interventions have been implemented to improve ECE quality, particularly in the area of process quality (Wolf & Peele, 2019; Wolf et al., 2019a, b), assessment reports of these intervention programs have not been encouraging (Wolf, 2019). We note with concern that, some of these process outcomes are found to be dependent on resources and support systems that are fundamentally conditioned by structural deficits. For instance, learning outcomes were revealed, unsurprisingly, to be contingent on teachers' well-being and mobilization of emotional resources to support students. These outcomes are largely produced by job design that overburdens teachers (Bennell & Akyeampong, 2007; Osei, 2006). If Ghanaian teachers are emotionally exhausted (Lee & Wolf, 2019), any expectations of close teacher–child relationship (Aboagye et al., 2019, 2020a, 2020b) purely based on Ghanaian collectivism (Affum-Osei et al., 2019a, 2019b) may not be guaranteed. Teacher professional identity in the domain of motivation may explicate the health impairment process of teacher ill-being (strain) and the consequent work withdrawals, influencing students’ learning outcomes, by revealing elements that promote the enactment of healthy emotional labor strategy in executing emotional labor at ECE centers.

The JD-R theory (Demerouti et al., 2001) is the leading theoretical model for explicating organizational stress. The theory stipulates that employee well-being can be configured by job demands on employees and the job resources made available to the employees in meeting these demands. Demerouti et al., (2001, p. 501) refer to job demands as “those physical, social, or organizational aspects of the job that require sustained physical or mental effort and are therefore associated with certain physiological and psychological costs (e.g., exhaustion)”, and job resources as “those physical, psychological, social, or organizational aspects of the job that may do any of the following: (a) be functional in achieving work goals; (b) reduce job demands and the associated physiological and psychological costs; stimulate personal growth and development”. Job demands are seen as stressors that potentially produce strain (e.g., exhaustion), impairing employees’ health whilst job resources are considered health-protective, engendering motivational impetus (e.g., engagement). These two mechanisms are referred to as the energetic (strain) and motivational (engagement) effects of job demands and resources, respectively (Aboagye et al., 2020b; Affum-Osei et al., 2021). In addition to the main effects, the interaction effects of these two job characteristics have been observed – the buffer effect of job resources (Antwi et al., 2019; Boadi et al., 2022) and the amplifying effects of job demands (Bakker et al., 2007). Also, personal resources that enhance employees’ resilience and boost their capacity, and the appraisal of stressors as either a challenge or a hindrance have been incorporated (Belle et al., 2022; Schaufeli & Taris, 2014).

In this study, we assess the effect of personal resources (teacher identity: motivation domain) on the emotional labor strategies that are employed by teachers in meeting their emotional display rules set out by society and their schools. Examining teachers' professional identity in the domain of their motivational drive for the teaching work is important in LMICs. The difficult labor market conditions in LMICs (Affum-Osei et al., 2019a, 2019b) imply that there are teachers who join the profession not because of their innate drive and care for children but because they have to earn a living of some sort and teaching, both in the private and public sectors, appears readily available. Such a situation has the potential to produce teachers who have weak and unstable professional identity. With enormous job demands on teachers in LMICs coupled with meagre job resources (Bennell & Akyeampong, 2007; Osei, 2006), teachers with weak professional identity may be unlikely to meet emotional display rules using appropriate emotion management strategies. Such teachers are more likely to suppress their emotions just so they can get the next paycheck in hopes of finding a better job. Eventually, emotional suppression will become overbearing – causing strain (e.g., emotional exhaustion), and consequently work withdrawals. On the other hand, we argue that a strong teacher identity is a personal resource that affords teachers effortless reappraisal of their emotions, reducing the potential for emotional exhaustion and the attendant work withdrawals. Here, teacher identity is seen as teacher personal resource necessary for maneuvering stressful work contexts, where job demand-resource imbalance is high.

Ample literature evidence demonstrates the relevance of teachers’ emotional labor in education quality (see the meta-analytic study of Yin et al., 2019). However, the question of whether or not teacher professional identity (motivational component) influences emotional labor is unknown. So, our first contribution is to test this relationship (please see Fig. 1). Further, emotional labor on teachers’ experience of strain is examined; however, whether emotional labor induces teachers’ work withdrawal behaviors through strain (emotional exhaustion in our study) is also unknown. Our second contribution is to test the mediation role of teachers’ emotional exhaustion in emotional labor and work withdrawal strategies (please see Fig. 1).

The theoretical model

Full size image

Teacher identity and emotional labor strategies

Teaching can be an emotionally taxing activity (Yin et al., 2019). This is particularly true in ECE (Aboagye et al., 2020a, 2020b). Indeed, preschoolers can be emotionally dependent on their caregivers and/ or teachers. One important personal resource that can better facilitate teacher adaptation to the physical, cognitive, and emotional work of ECE in a resource-deficient environment in SSA is teacher identity. Zhang et al. (2016), for instance, project teacher identity as encompassing intrinsic, extrinsic, and volitional components. The author defines intrinsic value identity as reflecting teachers' "subjective feeling toward the teaching profession", extrinsic value identity as composing of teachers' "cognitions about external factors of the teaching profession" and volitional behavior identity as representing teachers' "behavioral engagement and willingness to join the profession" (Zhang et al., 2016, p. 4). Two of these three dimensions of teacher identity (intrinsic value identity and volitional behavior identity) have been amalgamated to form the motivation domain of teacher identity (Hanna et al., 2020). These two dimensions articulate how teachers identify with the teaching profession regardless of the perks of it. In this sense, teacher identity becomes an individual difference variable that identifies one teacher from the other regardless of work conditions.

As an enduring individual resource, teachers with strong identity are found to be enthusiastic about their roles and are highly emotionally involved in teaching and caring for students (Rots et al., 2010; Zembylas, 2003). This enthusiasm for teaching and caring energizes teachers to generate a sound support system for socializing students (Nias, 2002). It is in effect not unexpected that strong and stable teacher identity is protective of teacher emotional well-being (Zembylas, 2013), leading to quality teaching and learning interactions and activities in the classroom (Agee, 2004). In the fast-changing educational environment, teachers with strong identity are known to adapt and thrive in abrupt disruptions such as changing curriculum (Lasky, 2005). We, therefore, expect teacher identity to be the emotional anchor that grounds teachers when laboring emotionally. We contend that teachers who are grounded in their identity will be able to process their emotions and modify them in ways that align with the contextual emotional demands required of a teacher in any particular situation. However, preschool teachers with weak and unstable identity will fake their displayed emotions in order not to reveal their felt emotions for the sake of the exchange value of faking them. Therefore, we hypothesize that:

• Hypothesis 1: Teacher identity relates a) positively with deep acting, but b) negatively with surface acting.

For teachers to achieve their performance goals, they need a long-term collaborative synergy on the parts of students, their parents, colleagues, and management. As a result, the relations between teachers and these stakeholders tend to be closer than other service relationships (Yin et al., 2019). To cultivate these relationships toward a common goal (quality student developmental and learning outcomes), teachers are expected to be emotionally positive in their engagements and eschew negative emotions (Hochschild, 1983; Neuber et al., 2022). Nevertheless, teachers, even those with the best of intentions, may experience different shades of emotions on the job, some of which are unpleasant or downright ugly. In effect, emotion regulation becomes critical to ensure that teachers’ emotions conform to the broad societal and school-specific display rules to avoid teachers’ negative emotions thwarting the achievement of educational goals. As pointed out earlier, two main emotion regulation strategies have been put forward in the literature – deep acting and surface acting (Grandey, 2000; Hochschild, 1983). Expectedly, self-regulation of emotions in hopes of meeting socio-emotional expectations at school can be intensely resource-consuming (Bakker & Demerouti, 2007), producing various socio-emotional, psychological, physiological, and behavioral outcomes (Corcoran et al., 2018; Yin et al., 2019). On one hand, the balancing act required for teacher’s dual role of teaching and caring can be exciting and rewarding (Isenbarger & Zembylas, 2006). On the other, when wrong strategies are deployed in this act, both teachers and students become worse off (Yin et al., 2017).

The two emotion regulation strategies are either response- or antecedent-focused (Grandey, 2000). Surface acting (i.e., response-focused emotion regulation) is said to involve “changing the outward display of emotions so that they are in line with display rules but leaves the emotions felt mostly unchanged”, whilst deep acting (i.e., antecedent-focused emotion regulation) is seen as “the attempt to alter felt emotions to create a state of emotional consonance between displayed and felt emotions” (Zapf et al., 2021, p. 145). Teachers who feign unfelt emotion and suppress felt ones are noted to record poor organizational outcomes; however, those who adapt felt emotions to produce desirable ones via reappraisal or attention deployment fare better on the job (Burić & Mornar, 2022; Hochschild, 1983). These results have been witnessed to a considerable extent in the broad literature on emotional labor and employee outcomes. For example, the meta-analytic study of Kammeyer-Mueller et al. (2013) revealed that the display of positive emotions at work enhances employee job satisfaction and performance. These positive outcomes have a rippling effect consistent with the action-tendencies of positive emotions where employees who meet their organizational demands or performance goals including emotional goals feel good about their job and adopt approach behaviors (Grandey, 2000). Literature evidence demonstrates that teachers who deep act are less likely to withdraw from their teaching and care duties (Luo et al., 2019), but rather increase performance (Kim et al., 2017).

On the other hand, Kammeyer-Mueller et al. (2013) found surface acting or emotion suppression and faking of emotion to decrease job satisfaction and performance. The feeling of incompetence and disconnect may cause dissatisfied employees to engage in avoid behaviors (Cacioppo et al., 1999; Conley & You, 2021). Teachers who deploy surface acting have been fingered to frequently leave their working floor, play, or/and surf internet with available devices (e.g., phone, computer) during work hours (Bailey, 1996). Therefore, preschool teachers who surface act are more likely to withdraw. For example, Dudenhöffer et al. (2017) discovered that German teachers who reported being sick at school were unengaged. Three years later, a similar study by Howard and Howard (2020) found German public-school teachers who were stressed out to be frequently absent from school, and even those who showed up to school were also found unengaged. From a moral perspective, Shapira-Lishchinsky (2014) found that teachers’ perceptions of whether they are duly compensated for or not and organizational climate motivated lateness and absenteeism, respectively. So, the varied withdrawal behaviors can be thought of as maladaptive coping strategies. We contend therefore that ineffective emotion self-regulation mechanisms induce work withdrawal (Muraven & Baumeister, 2000). Accordingly, we expect teachers who surface act to withdraw from work whilst those who deep act to stay engaged at work. The following hypotheses have therefore been offered.

• Hypothesis 2a: Teacher deep acting relates negatively with teacher (i) presenteeism and (ii) lateness.

• Hypothesis 2b: Teacher surface acting relates positively with teacher (i) presenteeism and (ii) lateness.

Though the JD-R theory (Demerouti et al., 2001) broadly explicates employee and organizational outcomes of job demands and resources, it emphasizes that these outcomes are possible through the management of employee strain (Bakker et al., 2007). Therefore, we submit that the effects of preschool teachers’ emotional labor strategies on their work withdrawals will be transmitted through emotional exhaustion. Suppressing felt emotions and faking unfelt ones to meet emotional demands have health-impairing effects (Kammeyer-Mueller et al., 2013). The regulatory cost of efforts exerted in making sure suppressed negative emotions do not seep through displayed positive emotions (Zapf, 2002) produces emotional fatigue and exhaustion (Bakker & Demerouti, 2017). Not expressing felt emotions induces a sense of inauthenticity. Over time, feelings of inauthenticity impoverish sense of self, depleting psychological resources (Erickson & Wharton, 1997). Besides perceived inauthentic self-expression, the felt negative emotion is negative affect in and of itself which, in the long run, results in emotional exhaustion (Hülsheger & Schewe, 2011; Weiss & Cropanzano, 1996).

On the other hand, modifying felt emotions to express positive emotions that are in sync with display rules reduces experience of job strain (Kammeyer-Mueller et al., 2013). In fact, satisfying emotional demands are linked with feeling accomplished, and satisfied with one’s job (Bhave & Glomb, 2016; Diefendorff & Richard, 2003). The positive psychological outcome of meeting emotional demands at work has been attributed to the confirmation of one’s identity and dignity (Humphrey et al., 2015). Additionally, serving others in a way that meets their needs creates a bond that over time deepens. The bond teachers share with students, their parents, colleagues, and management may nourish their affiliation needs (Zapf, 2002). Naturally, satisfied customers, grateful colleagues, and proud employers show employees that they matter. The recognition and appreciation are such a morale booster for employees’ well-being, diminishing employees’ emotional exhaustion (Humphrey et al., 2015).

Emotional exhaustion – considered as a state of being overextended, and depleted of physical and emotional resources (Maslach & Leiter, 2017), has detrimental effect on teacher-student relationship, as well as quality classroom practices (Aboagye et al., 2020a). Emotionally drained teachers feel at risk. The teachers’ sense of danger reduces their work commitment, and other citizenship behaviors (Delegach et al., 2017; Raman et al., 2016). Additionally, the failure to replenish lost psychological resources produces angry, retaliatory, and aggressive behaviors during classroom interaction, and eventually withdrawing from work (Back et al., 2020; Ryu et al., 2020). It is conceivable that teachers’ emotional labor strategies may affect teacher work withdrawals directly and indirectly through emotional exhaustion. Specifically, surface acting may result in extensive tension causing resource loss, thereby increasing emotional exhaustion and work withdrawal behaviors (Grandey, 2003; Hülsheger et al., 2010). However, teachers who deploy deep acting during classroom interactions may protect personal resources, thereby reducing the risk of emotional exhaustion (Steinhardt et al., 2011), effectively preventing the likeliness of work withdrawal behaviors. The following hypothesis has therefore been proposed:

• Hypothesis 3a: Emotional exhaustion mediates the positive effect of teacher surface acting on teacher (i) presenteeism and (ii) lateness.

• Hypothesis 3b: Emotional exhaustion mediates the negative effect of teacher deep acting on teacher (i) presenteeism and (ii) lateness.

Participant

The present study adopts a cross-sectional dataset in examining preschool teachers’ work withdrawal, a serious educational challenge in developing countries (Bennell & Akyeampong, 2007; Opoku et al., 2019; Peele & Wolf, 2021). This paper forms a section of a bigger project on the role of teachers’ emotional labor in work withdrawal given the unique challenges teachers face in developing nations. Before conducting this survey, the first author obtained permission from Zhejiang Normal University Research Ethics Committee and Ghana Education Service. One hundred and seventy-seven preschools were recruited from two municipal Assemblies in Ashanti Region (i.e., Kwabre East and Atwima Nwabiagya) using stratification sampling technique: ownership rights (i.e., government/public and non-government/private schools). These districts host various ethnic, cultural, technological, economic, and administrative activities of the region, with schools enlisting pupils and teachers from different economic and socio-cultural backgrounds.

Schools were selected according to socio-economic and cultural structure. Most of the schools had two kindergarten classes (i.e., KG1 consisting of four years and KG2 five years old pupils) comprising of both students with or without special developmental needs, and one to five teachers. In schools with only one or two to four kindergarten teachers, all teachers participated in the study. Not more than four teachers (two each from K1 and KG2) were selected per school, and randomly sampled if there were more than four preschool teachers in the school. All data were collected through a survey administered, upon which souvenirs were given after completion as a sign of appreciation. A final sample of 574 preschool teachers comprised 377(65.7%) female and 197(34.3%) male teachers, with majority holding diploma certificates and above [461(80.4%)], teaching in public schools [357(62.2%)], and with no or little effect of COVID-19 [455(79.2%)]. Their ages, work experiences, and class sizes range from 18–59 years (Mean = 32.21, SD = 8.07), 0.5–37 years (Mean = 8.45, SD = 5.66), and 9–60 pupils (Mean = 36.30, SD = 13.18), respectively see Table 1.

Teacher Identity: With the preamble: “Why did you become a preschool teacher?”, teacher identity was measured with the motivation domain on (Hanna et al., 2020, p. 8) Teacher Identity Measurement Scale (TIMS). This domain consists of two subscales: intrinsic career value (ICV), comprising 5 items (e.g., Because I like being a preschool teacher); and work with children, consisting 3 items (e.g., Because I like to work with children), measuring on a 7-point Likert scale (1 = Strongly disagree, 7 = Strongly agree). One item from the ICV (Because I want to have my own class) was found inappropriate for this study and was therefore deleted, leaving seven (7) items in all. The subscales of the motivation domain of the TIMS have been found reliable and valid using sample two (Hanna et al., 2020).

Emotional labor: Teachers’ emotional labor (i.e., surface and deep acting) were measured using the 10 items from Teacher Emotional Labor Strategy Scale (TELSS) adapted by Yin (2012). Surface acting consists of six items (e.g., I put on a ‘show’ or ‘performance’ when interacting with students or their parents), and deep acting comprises four items (e.g., I make an effort to actually feel the emotions that I need to display towards students or their parents). Teachers were asked to rate emotional labor strategies used on a 5-point Likert scale from 1 (strongly disagree) to 5 (strongly agree). This scale has been found reliable and valid in previous studies. For instance, Yin et al., (2013) found surface acting to be reliable at α = 0.74 and deep acting at α = 0.78, and used to predict teaching satisfaction.

Emotional exhaustion: Teachers’ emotional exhaustion component of Maslach Burnout Inventory, Educator’s Survey (MBI-ES) (Maslach & Jackson, 1981) was used in this study. This subscale consists of 9 items measuring teacher’s state of overextension and drain (e.g., I feel emotionally drained from my work), rated on a 7-point Likert scale ranging from 1 = Never to 7 = Every day. The emotional exhaustion subcomponent of the MBI-ES has been found reliable (a = 0.81) and valid in a developing nation’s context such as Ghana, and predicted conflict teacher–child relationship acceptably (Aboagye et al., 2020a).

Lateness: Nine (9) items (e.g., I find it acceptable to be ten minutes late to work) measuring lateness with seven (7) point scale (1 = Strongly disagree, to 7 = Strongly agree) developed by Foust et al. (2006) was used. The item 1 to 6 were reverse coded. This scale has been found satisfactory reliable and valid by Foust et al., (2006; α = 0.84) and correlates with job satisfaction.

Presenteeism: Six (6) items (e.g., I delay starting on new projects at work because of stress) developed by Gilbreath and Frew (2008), rate on a five-point Likert-type scale ranging from 1 = Never to 5 = All the time, was adopted. Chia and Chu (2016) found 6-item presenteeism scale to be sufficiently reliable (α = 0.92) and correlated with job stress.

Covariate: Teachers’ personal and job-related factors were examined as control variables in the model due to their influence on emotional labor (Grandey, 2000; Hochschild, 1989; Olson et al., 2019; Ulufer & Soran, 2019), and behavioral formation (LeBlanc et al., 2014). These variables include: gender (1 = Male & 2 = Female), age (teachers were asked to state their age which was categorized into: 1 =  ≤ 29, 2 = 30–39, 3 = 40 and above), education level (1 = High school and lower, 2 = College/Diploma 3 = Bachelor’s degree and above), school type (1 = public school & 2 = private school), work experience (teachers were asked to state their work experience in years and then classified into: 1 =  ≤ 5, 2 = 6–10 & 3 = 11 years and above), COVID-19 effect (teachers were asked “How much has COVID-19 has influenced their life: 1 = Not much, 2 = Somewhat, 3 = Much & 4 = A great deal), and class size (teacher stated total number of pupils in his/her class and then classified into: 1 =  ≤ 15, 2 = 16–30 3 = 31–45 & 4 =  > 45). See Table 1 for detailed participants’ information.

To prepare our data, a series of preliminary procedures were performed. We examined missing data patterns with Little and Rubin's (2002) “missing completely at random (MCAR)” test. Also, using Herman’s single factor exploratory factor analysis (EFA), we assessed for the presence of common method variance (CMV).

Before testing our structural model, we first tested the measurement model, and examined descriptive statistics and inter-constructs correlation. The measurement and structural models’ fit were assessed following Hu and Bentler's (1999) Chi-Square statistics (χ²), the Root Mean Square Error of Approximation (RMSEA ≤ 0.08), Normed Fit Index (NFI ≥ 0.90), Comparative Fit Index (CFI ≥ 0.90), Tucker–Lewis Index (TLI ≥ 0.90) and Standardized Root Mean Square Residual (SRMR ≤ 0.08) for acceptable fit as applied in recent studies (Antwi et al., 2022a, b). Last, structural equation modeling (SEM) was employed to test the study hypotheses. SEM was utilized because it supports the complex structural direct and indirect relationships between variables in a single model and gives meaningful and valid estimates (Bagozzi & Yi, 1988). All analyses were conducted with IBM SPSS version 26 and IBM AMOS version 24 using maximum likelihood estimation.

Measurement model

We adopted a two-stage strategy in analyzing the research data using IBM SPSS version 26 and AMOS version 27. Before proceeding to test the measurement model of our study, the pattern of missing data was examined using Little’s “missing completely at random” test which showed statistically nonsignificant (x² = 232.386, df = 246, p = 0.740). Subsequently, the missing data were treated using the full information maximum likelihood recommended by (Enders & Bandalos, 2001). Further, Harman’s one-factor EFA was examined to check common method variance (CMV) (Podsakoff et al., 2003), where all the 41 items were loaded on one general factor. The factor explained 31.98% of the common variance, indicating that CMV is not an issue in the data.

Test of the measurement model revealed a good fit (X² = 1601.63, df = 975, X²/df = 1.64, NFI = 0.93, CFI/TLI = 0.97/0.97, RMSEA = 0.03, SRMR = 0.04), as well as CFA loadings after item-level modification. Constructs’ discriminant validity was also observed as average variance extracted (AVEs) (0.54 ≤ AVEs ≤ 0.83) were greater than maximum shared variance (MSVs) (0.06 ≤ MSVs ≤ 0.34), and the square root of AVEs (0.74 ≤ \(\surd \mathrm{AVEs}\)  ≤ 0.91) were also greater than inter-construct partial correlation (-0.46 ≤ r ≤ 0.58) (Fornell & Larcker, 1981). In addition, heterotrait-monotrait (HTMT) correlation ratio met the cut-off points (0.85 or 0.90) recommended by Henseler et al. (2015). Convergent validity was met with AVEs being greater than 0.50. Again, constructs’ internal reliabilities (0.87 ≤ a ≤ 0.98), composite reliabilities (0.80 ≤ CR ≤ 0.98), and maximal reliability (0.80 ≤ a ≤ 0.98) met the minimum threshold (0.70). The constructs’ normality statistics (mean ± SD, skewness, and kurtosis), validity and reliability (average variance extracted (AVE) and maximum shared variance (MSV), square root of AVEs, composite reliability, Cronbach’s alphas, maximum reliability, and inter-constructs’ correlation coefficient) have been presented in Table 2.

Second, structural equation model (SEM) was employed to examine a series of relationships between (i) teacher identity and surface acting and deep acting, (ii) surface acting and lateness and presenteeism, and (iii) deep acting and lateness and presenteeism. Gender, age, education, school type, COVID-19 effect, work experience, and class size were treated as control variables in the model, which gender (Male = 0 & Female = 1), school type (Public = 0 & Private = 1), education level (SHS and lower as a reference), and COVID-19 effect (Not much as a reference) were dummy coded, and age, work experience, and class size were treated as continuous variables. The model provided an acceptable fit index to the data (X² = 1102.97, df = 702, X²/df = 1.57, NFI = 0.95, CFI/TLI = 0.98/0.98, RMSEA = 0.03, SRMR = 0.06). The standardized regression estimates revealed that teacher identity had a significant positive association with deep acting (β = 0.49, P < 0.001), but significant negative association with surface acting (β = -0.23, P < 0.001). Further, deep acting had a significant negative relationship with lateness (β = -0.41, P < 0.001), and presenteeism (β = -0.20, P < 0.001). Surface acting revealed a significant positive association with lateness (β = 0.17, P < 0.001), and presenteeism (β = 0.14, P < 0.01) (see Fig. 2).

Hypotheses 1 & 2, structural model examining the direct effect of teacher identity on emotional labor, and emotional labor on work withdrawal. Note: Regression estimates for the direct paths are standardized values. †p < .100, *p < .050, **p < .010, and ***p < .001

Full size image

Further, the indirect effects of deep acting and surface acting on lateness and presenteeism via emotional exhaustion were tested using a biased-corrected bootstrapped method with 95% confidence interval. The model provided an acceptable fit index to the data; (X² = 1277.45, df = 777, X²/df = 1.64, NFI = 0.94, CFI/TLI = 0.98/0.97, RMSEA = 0.03, SRMR = 0.03). The results revealed a significant indirect association of deep acting with lateness, and presenteeism through emotional exhaustion. The indirect association of surface acting with lateness, and presenteeism were not significant, having zero in the confidence interval. Partial mediation was established; thus, emotional exhaustion partially mediates the relationship between deep acting and lateness, and presenteeism, as the direct effects remained significant after the mediation. Hence, there was a significant association between deep acting and lateness and presenteeism directly and indirectly (see Table 3).

Results’ summary

Prior research on emotional labor has documented its costs and benefits to employees and their organizations (Hülsheger & Schewe, 2011; Kammeyer-Mueller et al., 2013). In education, emotional labor is fingered to be a central component of what makes great teachers to achieve the socio-emotional and academic goals of education (Corcoran et al., 2018; O’Connor, 2008). The centrality of emotional labor in quality education and the absence of relevant research on it in LMIC occasioned this study. Using cross-sectional data from preschool teachers in a LMIC (i.e., Ghana), the present study examined i) the effect of motivational aspect of teacher identity on emotional labor, ii) the influence of emotional labor on work withdrawal, and iii) the mediating effect of emotional exhaustion in the emotional labor and work withdrawal relationship, utilizing JD-R theory (Demerouti et al., 2001). Generally, our findings reveal the relevance of preschool teacher personal resource (i.e., teacher identity) in emotion regulation where strong teacher identity promotes deep acting but diminishes surface acting, supporting hypotheses 1a & 1b. Emotion regulation strategies (deep acting and surface acting) were found to determine preschool teachers’ work withdrawals, giving support to hypotheses 2a & 2b. Lastly, teachers’ strain (i.e., emotional exhaustion) was discovered to mediate the role of deep acting on preschool teachers’ presenteeism and lateness.

From a theoretical point of view, the present study enriches theory and research in many ways. First, the influence of teacher identity on teachers' strategy for emotion regulation at the preschool environment emphasizes the salience of identity-related resources in exploration of employee strain and organizational performance. This finding strengthens the argument for harnessing employees' motivation as a personal resource for wellness and goals achievement (Schaufeli & Taris, 2014). Therefore, teacher identity proves to be a salient motivational and/ or buffer resource in the JD-R theory (Demerouti et al., 2001). Empirically, this result is consistent with those of prior studies. For example, teacher identity has been identified as a condition for teacher enthusiasm (Rots et al., 2010; Zembylas, 2003) and adaptation (Lasky, 2005). When teachers who have strong identity are enthused over their teaching profession, and can easily adapt to emergent issues in the classroom, it becomes theoretically tenable why teacher identity as a personal resource facilitates deep acting whilst inhibiting surface acting, expanding the empirical evidence on the role of personal resources in the JD-R theory.

Second, the effects of preschool teachers’ emotional labor strategies – deep acting and surface acting, on their lateness and presenteeism suggest that teachers’ healthy interpersonal relationships in their work environment are crucial to their flourishing on the job. This supports the broad literature on personal accomplishment and positive organizational behaviors (Grandey, 2000; Humphrey et al., 2015). That is, it confirms the research evidence that meeting emotional display rules supports employees (preschool teachers in our case) sense of self-efficacy and personal accomplishment as well as an improved sense of shared bond of friendship and belongingness in the relationships formed with relevant parties involved in ECE (Zapf, 2002). Teachers are therefore likely to report to school with enthusiasm and stay engaged whiles there in spite of the challenging work conditions. Therefore, studies on ways teachers in LMICs can meet their emotion display rules will go a long way in helping to improve quality ECE education in these countries.

Lastly, the mediation role of preschool teachers’ strain (i.e., emotional exhaustion) in deep acting and work withdrawal explicates the necessity of job demand-resource balance. That preschool teachers whose displayed emotions are in consonant with their felt emotion do not withdraw from work because having a self-consistent emotional expression inhibits the incidence of emotional exhaustion which is the catalyst for work withdrawal. This finding highlights the salience of preschool teachers’ authentic self-expression consistent with the emotion work of Erickson and Wharton (1997) which illuminates that being inauthentic in interactive service work produces strain. Also, the recognition that comes with meeting emotional expectations enhances well-being and improved work behaviors by preventing the strain (Holman, 2016; Humphrey et al., 2015).

First, as it stands, more teachers across levels of education in Ghana, and likely the case in other LMICs, particularly at the ECE centres have not been trained for the work. These educators are not professionals. Professionally, they are trained for myriad other jobs, including business management and policy development (Harris, 2020). The high unemployment rates in Ghana (Affum-Osei et al., 2019a, 2019b) drive these professionals into teaching jobs, mostly considered as short-term transitory career detour. Unfortunately, some of these new recruits have no passion for the profession nor the desire to nurture children. These ultimately result in poor-quality ECE. We recommend that recruitment of teachers with no professional training in education should be trained for a reasonable time period to hone their sense of professional identity. This is necessary to curb not only unprofessional behaviors concerning child abuse, but also to develop and improve their teaching philosophies and goals.

Second, education programs for pre-and in-service teachers should pay attention to preschool teachers’ professional identity and emotional competencies. The current module of teacher education has been critiqued for its exclusive focus on content knowledge and teaching methodologies at the neglect of interpersonal and socio-emotional skills (Wolf et al., 2021). It is important for these programs to address the emotionality of preschool education by equipping teachers with the emotional intelligence necessary for the attainment of emotional goals, particularly teachers who are not trained for the profession. Additionally, management and headteachers should organize periodic seminars to train teachers on emotional demands of the teaching profession, as well as appropriate skills or strategies needed to perform emotional labor for both novice and experienced teachers. Educational workshops for educators are deemed as effective remedial strategies for achieving educational goals (Kinman et al., 2011; Schutz & Zembylas, 2009). Also, open communication should be allowed at seminars for experienced teachers to share how they manage their emotions in challenging situations.

Finally, even after all the professional development workshops, from time to time, teachers may not be their best on the job. Supervisors and managers should create a work culture that allows preschool teachers who do not bring their A-game to school for various reasons – sometimes, unrelated to work conditions, to be able to seek permission to leave without costly consequences to their career development. In this way, teachers may not come to school whilst sick and engage in presenteeism or report to work late as an adaptive mechanism. Also, preschool teachers who are emotionally exhausted should be given the tools to make necessary work and life adjustment for their health benefits. Such a work climate would induce a sense of safety and security (Hong, 2012). Additionally, management should create awareness of different forms of emotional labor strategies, and its advantages and disadvantages, in order to reduce emotional vulnerabilities among teachers (Barry, 2020). Specifically, management should encourage teachers to use deep acting whenever managing emotions both in and outside classroom to foster interaction quality and minimize strain and work withdrawals (Mérida-López et al., 2020).

Though the present study sheds vital theoretical and practical insights on the relations among preschool teachers’ identity, their emotional labor strategies, burnout, and work withdrawals, it is worth discussing the limitations of the study to guide future research. First, the cross-sectional design of the study constrains causality deductions from the study findings. Therefore, a longitudinal study that explores these relations over time may offer an in-depth understanding into them. Also, an experimental study of the constructs in this study may eliminate any potential alternative explanations for the effects observed in this study to inform policy actions that streamline preschool teacher recruitment and training to ensure quality ECE in LMICs. Second, the data on all the variables in the study were self-reported. Therefore, common-rater bias could be could confound the study results (Podsakoff et al., 2003). However, a number of procedural and statistical processes were adopted to ensure data quality. First, we embedded attention check items in the questionnaire to check respondents’ attention and hence, data quality. Second, we applied Harman one-factor EFA to explore the potential of common method variance (CMV) and the results indicated that CMV is not an issue in the data. Nevertheless, we advise that the results of this study be interpreted within its limitations. Last, considering that the data for this study is from an emerging economy, the applicability of the results in countries with similar labor and reward conditions may be tenable. However, such application must be done within the limitations of the study. Ideally, it would be best to replicate this study in other LMICs.

Considering the demand-reward imbalance in the teaching profession in low- and middle-income countries (LMICs), particularly those in sub-Saharan Africa (SSA), motivation deficit has been fingered as a persistent problem, leading to teachers’ ill-being and negative behaviors that produce poor teaching and learning outcomes. In this study, we conceptualize teacher professional identity as a motivational resource that relieves teachers of negative work attitudes and behaviors. Specifically, the results of this study demonstrate that preschool teachers’ professional identity promotes the enactment of positive emotional labor strategy (deep acting), whilst diminishing the inclination toward negative emotional labor strategy (surface acting). The results also show that deep acting reduces preschool teachers’ emotional exhaustion but surface acting reinforces emotional exhaustion. Further, preschool teachers’ emotional exhaustion was found to evoke unfavorable teacher work behaviors – lateness and presenteeism. Last, teachers’ emotional exhaustion mediated the effects of teachers’ emotional labor strategies on their work withdrawals—lateness and presenteeism. These findings demonstrate the critical role of teacher professional identity in education quality assurance in SSA in view of its downstream implications for emotional labor, emotional exhaustion, and work withdrawals.