Antecedents of happiness at work: The moderating role of gender

Abstract Happiness at the workplace (HAW) has received growing scrutiny in the last two decades. However, little is known about the factors that determine this central organizational phenomenon. This study investigates the relationship between organizational (high commitment human resources practices (HCHRPs); perceived organizational support (POS); and organizational respect for non-work (ORN)) and personal factors (passion for work (PFW), employee psychological capital (PsyCap), and gender), and HAW as a holistic and multidimensional construct. Data obtained from 359 university professors in the United Arab Emirates was analyzed using a (PLS/SEM) to test the research hypothesis. Findings indicate that POS, HCHRPs, PFW, and PsyCap have a significant positive relationship with HAW. However, no significant relationship between ORN and HAW was found. A PLS multi-group analysis indicated that gender was not found to ameliorate these relationships. Implications of the findings, limitations and future research are discussed.


PUBLIC INTEREST STATEMENT
Organizations around the world increasingly want to monitor their employees' happiness at work.Happier employees are more productive, more engaged, and more committed to their organization.In order to effectively manage this important issue, managers are striving to identify the factors that contribute to employees' happiness.For decades, happiness in the workplace has been studied through different aspects such as job satisfaction, engagement, affective commitment, flourishing levels, and wellbeing among others.More recently, a holistic approach to studying happiness in the workplace has been introduced by Fisher (2010).This study attempted to investigate the impact of a number of organizational and personal factors on employee happiness using the emergent holistic approach.Results show that organizational support, along with high commitment HRM practices, significantly impacted employee happiness.The findings also revealed that passionate, optimistic, hopeful, resilient and employees with selfbelief were happier than their counterparts, which was not affected by gender differences.

Introduction
In today's dynamic, competitive, and information-and knowledge-driven business environment, people have become of strategic importance due to a shift in business organization concentration from scarce financial capital to scarce human capital (Williams et al., 2016).Organizations are recognizing the vital role of employees in creating sustainable values through their innovative behavior.Consequently, employee's happiness at work has become one of the main concerns of policy makers and is considered an important tool for the retention of scarce and talented employees.Happiness at work refers to factors that reflect pleasant judgments, pleasant experiences or positive affective experiences in the workplace (Fisher, 2010).
Several reasons are behind the growing interest in HAW.First, most of us are spending a considerable portion of our lives at work (De Neve & Ward, 2017).Erdogan et al. (2012) have estimated that on average a person who starts working at the age of 21, and retires at 65, spends more than 90,000 hours working.In that regard, Gavin and Mason (2004) have considered HAW as among the major factors that shape people's life and happiness today.Second, it has been argued that HAW can lead employees to career success and improve organizational outcomes such as productivity/profitability, employee retention, and customer satisfaction (Boehm & Lyubomirsky, 2008;Fisher, 2010).Furthermore, HAW has become an indication of the quality of a working life; therefore, organizations are increasingly keen to achieve a good position in prestigious HAWrelated rankings such as the World's Best Workplaces, Bloomberg's Best Place to Work, Fortune 100's Best Companies to Work for, and so forth.Fisher (2010) was among the first scholars who identified the problem of studying HAW.Through an extensive review of the literature, Fisher (2010) concluded that HAW literature could be described as fragmented and was lacking a holistic measure.For decades, HAW has been used interchangeably, in the majority of the empirical literature, with various constructs such as job satisfaction, engagement, job involvement, affective organizational commitment, thriving, vigor, flow and intrinsic motivation, flourishing, wellbeing, quality of working life, and work affect.Although they all belong to the family of happiness constructs in some way, each of them is only able to measure a single aspect of HAW rather than capturing the whole phenomenon.In that regard, Fisher (2010, p. 391) has pointed out: If happiness at this level (the stable person level) is viewed as the proverbial elephant being examined by blind men, we can conclude that we have developed a good if isolated understanding of its parts, such as the trunk (e.g., job satisfaction) and the tail (e.g., typical mood at work).It may be that we have decomposed the beast into almost meaninglessly small pieces (e.g., the right ear of vigor, the left ear of thriving).Perhaps what is missing is a more holistic appreciation of the entire animal in the form of happiness at work.
With the absence of a holistic and distinct measure for HAW at the person level, Fisher (2010) took suggested a multidimensional (higher order) measure that consists of three first order measures: (1) employee's job satisfaction, (2) employee's affective organizational commitment, and (3) employee's engagement.These can work together to capture the majority of the variance in individual-level happiness at work.Since then, the number of studies conducted based on Fisher (2010) conceptualization has grown, yet they are still scarce and mostly done within western countries.The related literature suggests that there is a certain antecedent factor affecting HAW when measured based on Fisher (2010) conceptualization.For example, some studies have investigated the influence of different leadership styles on HAW, such as inspirational leadership (Salas-Vallina & Fernandez, 2017;Salas-Vallina et al., 2020), altruistic leadership (Salas-Vallina & Alegre, 2018), servant leadership (Salas-Vallina & Guerrero, 2018), and transformational leadership (Salas-Vallina et al., 2017).Still others have focused on the relationship between some organizational and personal factors and HAW, for example constructive dissensus (Salas-Vallina, 2020), gender diversity, diversity management and organizational inclusion (Mousa, 2020), gender differences (Mousa et al., 2020), PsyCap and perceptions of organizational virtues (Williams et al., 2015(Williams et al., , 2016)), employee recognition and work-life balance (El-Sharkawy et al., 2023); and leader-member exchange (LMX) (Qamar & Soomro, 2023).The scarcity of research that has adopted Fisher (2010) conceptualization for HAW, and the limited antecedents tested, have revealed a gap in the existing literature.This study attempts to address this scarcity by investigating the antecedents of HAW in the United Arab Emirates' (UAE) higher education sector, based on Fisher's (2010) holistic conceptualization of HAW.
Studying HAW in the UAE is worthwhile as the UAE ranked first in the Arab World in the World Happiness Report 2020 for the sixth consecutive year and was named among the world's best places to work in HSBC's Expat Explorer Survey.The UAE is also the first country in the world that created a Ministry of Happiness and Wellbeing to oversee its plans, programs and policies to achieve a happier society.
It is worth mentioning here that The Middle Eastern and North African perspective in general, and the UAE perspective in particular, about happiness has been influenced by this region's culture, which can be best described, according to Hofstede (1980), as a collectivist culture, and according to Hall (1976) as a high-context culture, which appreciates group membership, communication and interaction.AlNuaimi (2018) has pointed out that the people in the UAE might prioritize money as a major marker of their success and happiness due to social comparison, which is intense because of continuous engagement in active social life and family gatherings and occasions, which are considered a permanent component of the collectivist culture.This research seeks to find answers to the following two questions: How do organizational factors (high commitment human resources practices, perceived organizational support, and organizational respect for non-work), and personal factors (passion for work, employee psychological capital) contribute to employees' happiness at work? Do gender differences have any interaction effect on the relationship between these organizational and personal factors and employee happiness at work?So far, and to the best of the author's knowledge, there is no empirical study that has investigated the impact of those variables simultaneously within the moderating effect of gender.The rationale behind selecting these factors included their necessity in today's globalized and knowledge-based economy, which is relying more and more on employee innovation to attain competitive advantages, and their potential contribution in retaining talented employees.
The current study contributes to HAW literature in several ways.First and foremost, and contrary to the vast majority of previous literature that has merely focused on one aspect of HAW (e.g., wellbeing, job satisfaction, engagement), the present study addresses the missing holistic appreciation of this phenomenon, which has remained neglected for decades, by adopting Fisher's holistic framework in measuring and operating this construct.Therefore, more meaningful, beneficial, and realistic results were expected from the current study for both academic and practitioners than previous results that focused on one discrete or overlapping aspect of work-related happiness.The study is further expected to add to the overall body of research into positive organizational behavior/positive psychology, as it addresses a gap in the literature by testing the relationship between a set of central organizational and personal factors and HAW, and it identifies the gender differences in that regard.In addition, studying HAW within Middle-Eastern culture (i.e, UAE) is expected to add more to our understanding about this phenomenon, which is dominantly viewed from a western perspective.Fisher (2010) asserted that HAW can be predicted through employees' perceptions of organizational practices and qualities.The present study addresses the impact of organizational factors on HAW using the Job Demands-Resources (JD-R) model.The JD-R model suggests that workplace conditions can be categorized into two categories, job demands and job resources, and that each can lead to specific outcomes (Demerouti et al., 2001;Schaufeli & Taris, 2014).Considering the JD-R model, it is argued that the existence of specific job resources, such as POS, high commitment HRPs and the ORN, can promote positive attitudes/behaviors/emotions.Fisher (2010) argued that HAW can also be influenced by an employee's attributes, such as their personality, expectations, needs and preferences.It therefore draws upon Person-Environment Fit theory (P-E fit) (French et al., 1974), which suggests that specific personal characteristics, such as personality, skills, and knowledge, can promote or demote positive attitudes/behaviors/emotions.According to P-E fit, the fit between an employee and their organization can take two forms: supplies-values fit (S-V fit), which occurs when an organization fulfils employees' needs, and the demands-abilities fit (D-A fit), which occurs when an employee has the competencies/characteristics required to meet organizational demands (Kristof, 1993).Building upon that, the S-V fit was used to address organizational factors, since POS, HRPs and ORN can be adopted by organizations to make their employees happy, and the D-A fit was used to address the personal factors investigated in this study, since both PsyCap and PFW were required personal attributes from employees to meet organizational demands, but could also contribute to the HAW.Due to the scarcity of studies that have studied HAW with a holistic approach, this study used the following types of literature as indicators in building a set of hypotheses: (1) studies that used psychological, subjective, or affective wellbeing measures within the context of work; (2) studies that used subjective the happiness measure developed by Lyubomirsky and Lepper (1999); and (3) studies that used one single statement to measure HAW.

Organizational factors
This group included factors that were anticipated to impact employee HAW and were under the control of the organization.This study focused on the following factors.
2. 1.1.1. Perceived organizational support (POS).. Eisenberger et al. (1986) defined POS as "an employee's general perception concerning the extent to which the organization values their contributions and cares about their well-being" (P.500).Rhoades and Eisenberger (2002) argued that perceived fairness, supervisor support, and organizational rewards and job conditions can play an integral role in increasing POS.These three favorable treatments can, according to the JD-R model, be considered as a job resource (supplies in S-V fit), and it is anticipated that their existence in any organization could lead to high levels of HAW.An extensive review of the literature identified no study that directly investigated the relationship between POS and HAW within a holistic view.However, a review of the literature of other domains related to happiness, such as life happiness and wellbeing, exposed several studies that have examined the association between POS and life happiness/wellbeing.For example, Akgunduz et al. (2023), Gillet et al. (2012), Rhoades and Eisenberger (2002), Kosasih and Basit (2019), Joo and Lee (2017), Ni and Wang (2015), and Aggarwal-Gupta et al. ( 2010) have all found a positive relationship between POS and employee life happiness/wellbeing.This indicates that the more support the employee obtains from their organization, the happier they are at work.Hence, we posit the following hypothesis: H1: The perceived organizational support (POS) is positively related to an employee's happiness in the workplace (HAW).

Human resources practices (HRPs).. As far as
HRPs are concerned, this study embraced those HRM practices that promote high employee commitment at work.HRPs refer to sophisticated human resource management that implies a new kind of psychological contract based on trust, fairness of treatment, and delivery of promises (Guest, 2002).These types of practices change the relationship between employees and their employers to a more social relationship rather than a contractual one.Trust, fairness, and honesty, promoted by the high commitment HRPs, will be viewed by employees as favorable resources that encourage them to go the extra mile for their organizations and make them happier at work.A significant positive relationship has been reported between HRPs and life happiness/ wellbeing (e.g.Alfes et al., 2013;Baptiste, 2008;De Koeijer et al., 2014;Huang et al., 2016;Kim, 2019;Macky & Boxall, 2008).Therefore, we propose that: H2: High commitment human resources practices (HRPs) are positively related to employees' happiness in the workplace (HAW) Kirchmeyer (1995), ORN refers to employers acknowledging and valuing the non-work participation of workers and committing to supporting it.In that sense, organizations are required to provide their employees with personal resources and adopt practices such as flex-time, telecommuting, job sharing, compressed work week, or any other alternative or flexible work arrangements that help their employees achieve a level of work-life balance.It is anticipated that employees working in an organization that respects their participation in non-work activities will consider as favorable, in line with the JD-R model, job resources and extra job supplies, within the S-V fit, which make them happier at work.Our thorough literature review revealed no study that has tested the relationship between ORN and employees' HAW, nor even life happiness/wellbeing.Rather, we found many studies that investigated the relationship between employees' perceptions of work-life balance and their life happiness/wellbeing.For example, El-Sharkawy et al. ( 2023), Roy et al. (2022), Khan et al. (2020), Gröpel and Kuhl (2009), Otken and Erben (2013), Lunau et al. (2014), Allis and O'Driscol (2008), Yang et al. (2018), and Shams and Kadow (2019) have all confirmed that work-life balance and happiness/wellbeing are positively related.We therefore posit the following hypothesis: H3: The organization's respect for non-work (ORN) is positively related to employees' happiness at work (HAW).

Personal factors
This includes merits or attributes that employees might have gained/inherited and that are anticipated to impact their HAW.Some might argue that various personal factors (i.e., PsyCap and PFW) may predict happiness, yet the opposite causal direction may also apply.These reverse correlations are out of scope for this particular study.As mentioned earlier, the P-E fit model will be utilized to assess this relationship.

Psychological capital (PsyCap).
. PsyCap has been defined as an individual's positive psychological state of development, as characterized by high levels of four components: (1) selfefficacy, (2) optimism, (3) hope, and (4) resilience (Luthans et al., 2015).Employees at any organization are expected to have varied levels of positive PsyCap.Drawing upon the D-A fit part of the P-E fit model, employees with high levels of PsyCap will bring to the table a set of favorable qualities, such as self-efficacy, optimism, resilience, and hope, which facilitate them meeting organizational demands, and they subsequently foster high levels of HAW.Literature has suggested that PsyCap is positively associated with employees' life happiness/wellbeing at work (Avey et al., 2010;Culbertson et al., 2010;Joo & Lee, 2017;Kawalya et al., 2019;Kun & Gadanecz, 2019, Li et al., 2014;Williams et al., 2015Williams et al., , 2016)).Hence, we propose the following hypothesis: H4: Employees' psychological capital (PsyCap) is positively related to happiness at work (HAW).

Passion for work (PFW).
. PFW is defined as the extent to which people "love" to work and derive joy and pleasure from their investment in work-related activities (Baum & Locke, 2004).Vallerand et al. (2003;2003b) dualistic model of passion has identified two types of passion: obsessive passion, which is characterized by obligation or self-coercion; and harmonious passion, which is characterized by enthusiasm, engagement, and free acceptance of an activity.The present study focused on harmonious passion as it focused on people who were willingly and intrinsically passionate about their job (Ho et al., 2011).It was expected that this type of passion would contribute to an employee's HAW.Findings of a range of empirical studies have supported the positive relationship between PFW and life happiness/wellbeing (e.g., Curran et al., 2015;Moè, 2016;Rousseau & Vallerand, 2008;Vallerand et al., 2008;Yukhymenko-Lescroart & Sharma, 2019).We therefore posit the following hypothesis: H5: Passion for work (PFW) is positively related to happiness at work (HAW).

Moderating effect of gender
When it comes to gender differences, the literature tends to use social role theory (also called gender role or societal stereotype).This theory suggests that people form perceptions or beliefs about gender based on its attributes and observed behavior, and they deduce that each gender possess corresponding dispositions (Eagly & Wood, 2012).The stereotype that women are more sociable, more emotionally expressive, and less aggressive than men is quite widely held (Brody, 1997;Shields, 2002).Indeed, people believe that women express anger less and happiness more than men in the workplace.In other words, men have been seen as more expressive of anger while women have been seen as more expressive of HAW (Konrad et al., 2000;Sloan, 2012).Knowing the distinction between male and female employees in terms of the HAW is crucial from both theoretical and practical points of view.From a theoretical perspective, investigating this issue could contribute to addressing a debate between scholars.Conversely, from a practical perspective, knowing these differences might help businesses in designing a retention strategy that fits each gender.Furthermore, Konrad et al. (2000) have pointed out that gender roles might differ by culture and since most of the happiness/wellbeing has been studied in western culture, studying gender differences within a Middle Eastern cultural context is justified and worthwhile.
Empirical studies concerning gender differences in relation to life happiness or well-being levels have revealed high levels of inconsistency (Batz & Tay, 2018) and have so far been inconclusive (Wilks & Neto, 2013).For example, Booker et al. (2018), Burke et al. (2009), Inglehart (2002) and Mousa et al. (2020) have all failed to find a significant difference between male and female wellbeing/happiness.Yet others, such as Wilks and Neto (2013), Stevenson and Wolfers (2009), and Sloan (2012), have reported significant differences between genders.Hence, we propose the following hypothesis: H6: Gender moderates the relationship between (POS, HRPs, ORN, PsyCap, and PFW) and happiness at work (HAW), with a stronger relationship among female employees than male employees.

Participants
The participants were 359 university professors from different higher education institutions in the United Arab Emirates (UAE).Participants self-identified as male (67%), or female (33%).With regards to marital status, 85% of the participants were married and 15% were not married.The sample in this study comprised lecturers (23%), assistant professors (39%), associate professors (20%), and full professors (18%).Of them, 98% were expats and 2% were locals.As far as the location of the participants was concerned, 21% were from Abu Dhabi, 26% from Dubai, 40% from Sharjah, 3% from Ajman, 2% from Al Fujairah, 5% from Umm Al-Quwain, and 3% from Ras Al Khaima.

Happiness at work (HAW)
Following Williams et al. (2015) and Salas-Vallina et al. (2017), and based on Fisher (2010), the conceptualization of the HAW construct was a second-order variable that consisted of: (1) engagement with the work itself; (2) satisfaction with the job; and (3) feelings of affective commitment to the organization as a whole.This study combined three previously developed and validated scales that shaped employees' HAW in one second-order construct with 23 items.A six items scale, developed by Tsui et al. (1992), was used to measure job satisfaction.The following sample item is an example: "How satisfied are you with the nature of the work you perform?"A nine-item shortened organizational commitment questionnaire, developed by Mowday et al. (1979), was used to measure an organization's commitment.The following is a sample item: "I am willing to put in a great deal of effort beyond that normally expected in order to help this university/institute be successful".Finally, a UWES-9, developed by Schaufeli et al. (2006), was used to measure work engagement, with the following sample item: "At my work, I feel bursting with energy".The Cronbach's Alpha (α) for this construct was found to be 0.94 in this study.

Demographic variables
Participants were asked to report their gender, age, marital status, academic rank, nationality, and location of the higher education institutions they worked for.These variables were deliberately placed between the measurement of the dependent variable (HAW) and the measurement of all independent variables, in line with one of the procedural remedies suggested by Podsakoff et al. (2012), which provides a psychological separation between the measures of the dependent and independent variables to control for common method bias.

Passion for work
PFW was measured using Baum and Locke's (2004) five items.The sample item assessed the level of agreement with statements such as "I love my work".The Cronbach's Alpha (α) for this construct was found to be 0.84 in this study.

Psychological capital
PsyCap was measured using the 12-item self-rated short form PCQ-12 Questionnaire, developed by Luthans et al. (2007).PsyCap is a multidimensional construct comprising items that measure four sub-constructs: Hope, Self-Efficacy, Resilience, and Optimism.Sample items include: "I feel confident presenting information to a group of colleagues" (efficacy); "If I should find myself in a jam at work, I could think of many ways to get out of it" (hope); "I can get through difficult times at work because I've experienced difficulty before" (Resilience); and "I'm optimistic about what will happen to me in the future as it pertains to work" (Optimism).The Cronbach's Alpha (α) for this construct was found to be 0.87 in this study.

Perceived organizational support
POS was measured with an eight-item version of the original 36 items developed by Eisenberger et al. (1986) and used by Eisenberger et al. (1997).These eight items were loaded highly on the main factor, which seemed applicable to a wide range of organizations.The following is a sample item: "My organization really cares about my well-being".The Cronbach's Alpha (α) for this construct was found to be 0.89 in this study.

Human resource practices
HRPs were measured using the eight items developed by Gould- Williams and Davies (2005).The scale was based on those HRM practices fostering "high commitment" for employees toward their organization (Guest, 1997;Marchington & Grugulis, 2000;Wood & Albanese, 1995).The following is a sample item: "I am provided with sufficient opportunities for training or development".The Cronbach's Alpha (α) for this construct was 0.86 in this study.

Organizational respect for non-work
Four items, developed by Kirchmeyer (1995), were selected to measure ORN.These items reflect employee perceptions of the organization's respect-type response to non-work.Sample items included: "My organization accommodates employees' special non-work needs", and "My organization is flexible about employees' work schedules."The Cronbach's Alpha (α) for this construct was found to be 0.89 in this study.
The items of the continuous scales above were measures using a six-point Likert scale to avoid neutral responses.The negative-worded items in all measures were reversely scored.

Procedures
Data were collected based on a stratified convenience sample design and through an online large-scale survey that gathered responses from university professors from different higher education institutions in the United Arab Emirates (UAE).The online survey was conducted by a third party, which was a wellknown, specialist data service and marketing research organization: Dun and Bradstreet.Dun and Bradstreet's extensive database included a list of 18,600 emails for academic staff members working in higher education providers across all emirates of the UAE.This list served as a sampling frame.Leavy (2017) has recommended using some sample size calculators available online to calculate the ideal sample size for a particular study.The targeted sample size was calculated by the sample size calculator provided by Qualtrics (https://www.qualtrics.com/blog/calculating-sample-size/).The ideal sample size was found to be 370 responses, based on a population size of 18,600 and a confidence level of 95%.To reach the targeted sample size that seemed statistically representative, and to ensure a large enough sample size for the use of structural equation modelling, as recommended by Hair et al. (2017), an invitation email, along with a URL link, was sent to all of the targeted populations during the last quarter of 2019.A total of 396 questionnaires were received; however, 37 questionnaires were incomplete.The final realized sample, after eliminating the incomplete ones, was 359 usable questionnaires, which was very close to the ideal sample size calculated by Qualtrics.The total number of usable questionnaires also exceeded the rule of thumb suggested by Hair et al. (2017) with regards to the usability of PLS/SEM.In our model, there were five paths directed at the dependent variables (HAW); therefore, the minimum sample size should be fifty (5 × 10).Therefore, the sample size of this study is comfortably above this minimum.

Data analysis
Data analysis was run using Partial Least Square Structural Equation Modelling (PLS-SEM).SmartPLS 3.2.9software (Ringle et al., 2015) was used to analyse the data of this study.PLS-SEM was selected because it was suitable for testing theoretical models from a prediction perspective (Hair et al., 2019).Guidelines provided by Hair et al. (2016) and Chin (2010) were followed to analyse the data over the following two stages: testing the measurement modem, and testing the structural model.

Preliminary analysis
A preliminary analysis for the data was conducted to correct possible errors, detect and overcome outliers, examine variables' normality and multicollinearity, following the guidelines provided by Tabachnick and Fidell (2007).Armstrong and Overton (1977) recommended procedures to check for possible non-response.By performing the Mann-Whitney U test, we found no significant differences between the first third and the last third of the respondents' data, so we deduced that non-response bias did not exist in the current study.We also checked for common method bias (CMB) for all principal variables used in this study.A full collinearity assessment was conducted, based on the protocol provided by Kock (2015, p. 7) using SmartPLS.All variance inflation factors (VIFs) for the latent variables were found to be lower than the threshold of 3.3; thus, we could infer that common methods bias did not exist in this study.Both of the control variables in this study were categorical in nature.Therefore, an ANOVA analysis was conducted using the categorical control variables as independent variables (i.e., marital status, and age) and HAW as a dependent variable.None of the categorical control variables were significant and were therefore excluded from further analysis.

Measurement model test
For the test of the reflective measurement model, as the one used in this study, Hair et al. (2016) recommended examining four psychometric properties: convergent validity (average variance extracted), discriminant validity, internal consistency (composite reliability), and indicator reliability.In PLS-SEM, assessment of measurement model is equivalent to CFA is CB-SEM.
Convergent validity was checked through factor loadings, Composite Reliability (CR), and Average Variance Extracted (AVE).Reliability at the indicator level can be checked by examining the item loadings on their respective constructs.Hair et al. (2016) proposed 0.7 as an absolute standardised outer loading to ensure that the indicator has captured at least half of the variance; however, items with 0.5 or 0.6 loadings are still acceptable if additional indicators in the block as the basis of comparison exist (Chin, 1998, p. 325).A decision was made to keep all items with a loading over 0.5 as long as the composite reliability and the AVE for their respective constructs was still over the satisfactory level of 0.7 and 0.50 respectively (Henseler et al., 2009).This process resulted in omitting 5 items that did not meet the above condition (i.e.HAW23, HRP3, POS6, POS7, and PsyCap 9).The composite reliability (CR) values for all constructs (HAW = 0.91; HRPs = 0.89; POS = 0.91; PASH = 0.88; PsyCap = 0.91; Org.Resp = 0.91) exceeded the recommended value of 0.7, indicating a high level of reliability.The Average Variance Extracted (AVE) for all constructs (HAW = 0.52; HRPs = 0.55; POS = 0.61; PASH = 0.60; PsyCap = 0.52; Org.Resp = 0.68) also exceeded the threshold of 0.50, which indicates an adequate convergent validity as it explains more than half of the variance of its items.Table 1 dipcted the items loadings and the AVE and CR for the research constructs.
The correlation matrix exhibited in Table 2 illustrates that the square root of the AVE (diagonal values) of each variable was larger than its corresponding correlation coefficients, confirming a sufficient discriminant validity (Fornell & Larcker, 1981).Henseler et al. (2015) argued that the Fornell and Larcker criterion did not perform well, particularly when the items' loadings on a construct differed only slightly; therefore, they proposed the Heterotrait-monotrait Ratio (HTMT) of the correlation as a better replacement.Once our item loadings on their respective constructs differed slightly, the discriminant validity was tested again using the new technique.The results of the Heterotraitmonotrait Ratio (HTMT) test, as shown in Table 2, confirm an adequate discriminant validity as all values in the table are below the threshold value of 0.85 (Kline, 2011).

Structural model test
Having established confidence in the measurement model, the structural model needs to be tested to examine the hypothesized relationships depicted in the research model.Hair et al. (2016) have suggested examining five indicators in order to test the structural model: coefficients of determination (R 2 ); predictive relevance (Q 2 ); size and significance of path coefficients (Beta-β); f 2 effect sizes; and q 2 effect sizes.First, we examined the direct relationship between the variables.
Results indicated that all factors investigated in this study had positive significant relationships with HAW except for ORN.These findings provided support for research hypotheses H1, H2, H3 and H4 but not for H5.Table 3 summarizes these results.Moreover, the five variables tested in this study (PFW, PsyCap, HRPs, POS, and ORN) explained 73% of the variance in HAWM (R 2 = 0.734).The R 2 values of 0.734 is higher than the 0.60 value that Cohen (1988) suggests would indicate a substantial predictive power.To test the effect size, guidelines provided by Cohen (1988) were used, which are 0.02 for a small effect, 0.15 for a medium effect, and 0.35 for a large effect.Table 4 shows that relationships had an effect size that ranged between small for (POS ➔ HAW) and (HRPs ➔ HAW), medium for (PFW ➔ HAW) and large (PsyCap ➔ HAW).Along with R 2 , and f 2 , the PLS path model's predictive accuracy was checked by calculating the Q 2 using the blindfolding procedures.Results show that the endogenous variables (HAW) in the model had an acceptable predictive relevance, as its Q 2 was greater than zero, as shown in Figure 1  To test gender differences, a multi-group analysis (PLS-MGA) was conducted after two groups were created: male (1) and female (2).Table 5 below displays the bootstrapping results for the two groups.To check whether the differences between the two groups are significant, we checked the parametric test results obtained from the PLS/MGA results report, as depicted in Table 6 below.As there no significant differences were captured over all constructs investigated in this study, we refute research hypothesis H6.

Conclusion
In today's dynamic, highly competitive and complex business environment, employees' HAW is considered a central strategy for the retention of scarce and talented employees.This study has tested the influence of three organizational factors (POS, HRPs, and ORN) along with two personal factors (PFW, and employees' PsyCap) on HAW, with the moderating effects of gender.
Findings related to the impact of the organizational factors revealed that POS positively and significantly affected the employees' HAW.This result is in congruence with a number of wellbeing at work studies such as those by Gillet et al. (2012), Rhoades and Eisenberger (2002), Kosasih and Basit (2019), Joo and Lee (2017), Ni and Wang (2015), and Aggarwal-Gupta et al. (2010).
One unanticipated finding was that no significant relationship was found between ORN and HAW.This finding is inconsistent with results found in previous wellbeing at work literature (i.e.Gröpel and Kuhl (2009); Otken and Erben (2013); Lunau et al. (2014); Yang et al. (2018); Shams and Kadow (2019).A possible explanation for these results may be found in the special nature of the higher education sector.Rosa (2022) has argued that the steady deployment of a neoliberal agenda amongst higher education providers has come to interfere with the work-life balance of their employees.Academics clearly knew since day one in their institutions that there would be no balance between work and life, as they would be expected to do many educational duties (i.e., preparing lectures, marking, and researching) at home.With regards to the impact of personal factors, the current study found a strong association between employees' PFW and happiness.This finding is in agreement with some wellbeing at work studies such as those by Yukhymenko-Lescroart and Sharma (2019), Curran et al. (2015), Vallerand et al. (2008), andVallerand (2012).Another important finding was that employees' PsyCap was positively and significantly related to HAW.These results match those obtained in earlier happiness and wellbeing at work studies such as those by Williams et al. (2016), Avey et al. (2010), Culbertson et al. (2010), Kawalya et al. (2019), Kun andGadanecz (2019), andJoo andLee (2017).
As far as the moderating effect of gender on the relationships between organizational and personal factors above and employee HAW, the results of this study did not provide support for any significant differences between males and females.Although these results differ from some published studies in the area of life happiness or wellbeing (Gerdtham & Johannesson, 2001;Sloan, 2012;Stevenson & Wolfers, 2009;Wilks & Neto, 2013), they are consistent with those of Burke et al. (2009), Inglehart (2002), Mousa et al. (2020), and Okun and George (1984).

Theoretical implications
This study has contributed to our understanding about the antecedents of happiness in the workplace.It has also offered another attempt to operationalize a holistic approach for happiness at work, as developed by Fisher (2010), and it has provided support to it.Reporting happiness at work results from different parts of the globe (i.e., the Middle East) rather than the dominant part (i.e., western countries) might add more insight into this important organizational phenomenon and allow for cultural differences in that regard.

Managerial implications
The findings of this study present several valuable implications.HRPs have been identified as an important organizational driver for HAW, which shows the importance for organizations to adopt a high commitment to HRPs, since this makes a positive contribution to employees' HAW.This indicates that higher levels of employee involvement, more socializing activities, more participation, personal development opportunities, and most importantly, higher financial compensation, since money in the UAE culture is considered a marker of happiness, are all required to maintain high levels of happiness in the workplace.
As far as organizational support is concerned, Eisenberger et al. (1986) asserted that three general forms of perceived favorable treatment received from the organization (i.e., fairness, supervisor support, and organizational rewards and job conditions) represent the antecedents of POS.This emphasis on treating employees fairly and encouraging academic department chairs to provide continuous support for them, along with a fair and motivational organizational reward system and improving the quality of working lives, are highly recommended to maintain high levels of HAW.
With regards to the personal antecedents for HAW, results showed that employees' PFW and their PsyCap were highly regarded causes for their HAW.In spite of the scarcity of research on the nature of the mechanisms and interventions that could improve employees' harmonious PFW (Yukhymenko & Sharma, 2019), Zigarmi et al. (2009) pointed out that both organizational characteristics and job characteristics can play a vital role in improving employees' PFW.Therefore, improving the quality of working lives for the organization and designing jobs and work activities in ways that help faculties meet their basic needs for autonomy, feedback, excitement, and development, by providing required support and job resources, can contribute positively to employees' PFW.In that regard, Mehmood et al. (2023), Slemp et al. (2021), andTeng (2019) have all found that passion for work was significantly and positively corelated to job crafting.
The PsyCap constituent elements (hope, optimism; self-efficacy, and resilience) are considered to be "state-like", and as such, they can change and be developed (Luthans et al., 2007).Accordingly, organizations need to pay attention to enhancing employees' hope, resiliency, efficacy, and optimism if they want to develop employee HAW.Goswami and Goswami (2022) have pointed out that so far, limited work has been done to develop various psychological capital interventions.In that regard, Luthans et al. (2006) suggested using a one-hour micro intervention.Luthans et al. (2008) further highlighted an internet-based intervention, and Luthans et al. (2010) later stressed short training interventions in developing employees' PsyCap.Culbertson et al. (2010) have argued that using Luthans and his associated interventions is more effective than traditional interventions such as job redesign and organizational behavior modification.Kun and Gadanecz (2019) have also provided some suggestions to improve employees' PsyCap.For example, the hope dimension can be developed through helping employees set realistic goals, remove all barriers and obstacles confronting them, and using solution-focused brief coaching.Optimism can be enhanced through various types of interventions such as gratitude interventions and adopting optimistic thinking.Optimistic thinking and the use of strengths-based feedback, instead of weakness-focused feedback, on performance can enhance employee self-efficacy.This study also suggests that it is crucial for practitioners to select people with a strong PFW for the job they apply for and those who have high levels of PsyCap.This would be possible when organizations incorporate a PFW and a PsyCap test as part of the package of tests used as part of the selection process.

Limitations and future directions
The main limitation of this study lies in its design.Data were cross-sectional, not time-lagged, and the causal relationships were tested with a single study rather than in different time periods.However, the common method bias test, along with some procedural remedies, were conducted to avoid the effect of this design issue on the study results.The sample size of this study was acceptable (359), although a bigger sample would have allowed for a more powerful analysis.Another potential shortcoming in this study was that it was conducted in one single context (the higher education sector), which might reduce the generalizability of its results to other contexts.Job-related variables, along with potential reversal relationships between HAW and PFW/PsyCap, were also not considered in this study.These limitations may open the door for future research to address these gaps.For example, more studies are required to study the influence of some other organizational, job-related, leadership-based, and personal factors on HAW, which were not mentioned in this study, such as organizational culture, autonomy, the five big personality traits, abusive leadership, emotional intelligence, bullying and so forth.More research is further required to investigate gender differences since the results related to HAW are still inconclusive so far.

Figure 1 .
Figure 1.Research model and results.