The social exchange theory is among the most influential conceptual frameworks for understanding organizational behaviors. Social exchange theory is a central premise that exchanging social and material resources is a fundamental human interaction in the workplace ( Blau, 1964 ). Several studies have demonstrated that social exchange relationships can predict job satisfaction, organizational commitment, organizational citizenship behaviors, and employee intentions to leave ( Matthijs Bal et al., 2010 ). Social exchange theory (SET) is among the most influential conceptual paradigms for understanding workplace behavior ( Memon et al., 2017 ), which is determined by the rewards of interaction between people or activities minus the penalty/cost of that interaction may have ( Griffith et al., 2006 ). Similar to this theoretical perspective, when employers’ and employees’ efforts are recognized and fairly rewarded, they reciprocate by going above and beyond their roles ( Agut et al., 2009 ; Lichtenstein, 2011 ). The social exchange theory is thus crucial for our model regarding the interaction between organization and employee to achieve something that both parties feel is fair or transparent in return for their effort and commitment ( Coates, 2010 ). Thus, this study aims to explain the relationship between research variables using social exchange theory ( Alias et al., 2018 ), as proposed in Figure 1 .

Herzberg’s motivation theory (i.e., two-factor theory and motivation-hygiene theory) applies to investigate employee motivation in the workplace ( Herzberg et al., 1959 ). Herzberg’s motivator-hygiene theory provided a great stimulus to the investigators for advancing research on job satisfaction ( Malik & Naeem, 2013 ). Central to the theories of motivation that collectively attempt to understand the thought processes that individual employees go through in determining how to behave in the workplace ( Steers et al., 2004 ). Herzberg’s two-factor theory (i.e., motivation-hygiene theory)—intrinsic factors are related to employee job satisfaction, while extrinsic factors are associated with job dissatisfaction ( Robbins et al., 2020 ). Herzberg’s motivation theory is one of the content theories of motivation to understand employee motivation in the workplace ( Peramatzis & Galanakis, 2022 ).

The compensation system is associated with performance appraisal, determining an employee’s work results. The compensation system also looks at how to create good work conditions, find competent employees, and successfully develop staff competency, all of which enable the company to guarantee high motivation and work satisfaction ( Türk, 2008 ). These HRM practices contribute to attracting, motivating, and retaining human capital (the knowledge, skills, and abilities embedded in people), which can help a company gain a competitive advantage ( Raymond 2017 ). All forms of pay or rewards going to employees and arising from their employment ( Dessler, 2020a ). Variable compensation plans that pay employees based on some performance measure ( Robbins et al., 2018 ). The compensation system relates to all forms of beneficial pay or rewards going to employees and arising from their employment ( Dessler, 2020b ).

Based on social exchange theory, employees repay benefits and opportunities provided by their working organization through continued effective performance when they find that the organization supports them by engaging in various human resource practices ( Karatepe, 2012 ). Perceived organizational support is “the extent to which employees believe their organization values their contributions and genuinely cares about their wellbeing” ( Kinicki, 2021: p. 59 ). Perceived organizational support refers to the “degree to which employees believe an organization values their contribution and cares about their well-being” ( Robbins & Judge, 2022b: p. 113 ).

Paullay et al. (1994) point out that job involvement is a psychological state wherein an individual “is cognitively preoccupied with, engaged in, and concerned with one’s present job” (p. 224). As managers, we always need to do things that will improve productivity, performance, and employee involvement or engagement and reduce absenteeism and turnover ( Lussier & Hendon, 2019 ). Job involvement represents how much an individual is personally engaged in their work role ( Kinicki, 2021 ). Managers can foster satisfying work environments to fuel employees’ job involvement. Job involvement refers to the “degree to which a person identifies with a job, actively participates in it, and considers performance important to self-worth” ( Robbins & Judge, 2022a: p. 66 ). A particularly strong relationship has been found between employee job involvement or engagement and retention, that is, labor turnover rates; higher levels of job involvement are strongly associated with lower turnover or increased high rates of job retention ( Derek et al., 2020 ).

Locke (1976) defines job satisfaction as “a pleasurable or positive emotional state resulting from the appraisal of one’s job or job experiences” (p. 1300). Job satisfaction is individuals’ positive or negative judgments, emotional reactions, and attitudes toward their job ( Weiss, 2002 ). Job satisfaction is defined as an affective reaction to a job that compares perceived outcomes with desired expectations ( Kardam & Rangnekar, 2012 ). According to Herzberg’s theory, job satisfaction is an employee’s positive feelings towards her or his work ( Liu et al., 2016 ). Job satisfaction involves an employee’s general attitude toward his or her job ( Robbins et al., 2018 ). Job satisfaction reflects the extent to which an individual likes their job. Formally defined, job satisfaction is an affective or emotional response toward various facets of your job ( Kinicki, 2021 ). Job satisfaction is a “positive feeling about one’s job resulting from an evaluation of its characteristics” ( Robbins & Judge, 2022b: p. 111 ). Education research has extensively examined the factors affecting teacher work satisfaction. Teachers should deliver better education when they are happy ( Jerrim, 2024 ). Thus, this research defines job satisfaction as the efficiency or result of the work that exceeds the expectations according to the work plan the organization has set for an individual employee to perform their roles and responsibilities.

Retention is typically measured by labor turnover rates, which measure employee departures from an organization’s employment in their simplest form. Within this wide definition, however, turnover can take many forms. For example, employees may resign, retire, be dismissed, or be made redundant. It is important to understand the reasons behind turnover to reduce it: labor turnover is extremely costly, and most organizations work to reduce turnover levels to minimize these costs ( Derek et al., 2020 ). Job retention is defined as a continuously held position at the same organization earning a competitive wage for more than one year ( Thomas & Morgan, 2021 ). Employee or job retention is getting competent and qualified people into the venture, which is the first step in effectively managing human resources. An entrepreneur wants to keep the people they have hired and trained. A unique and important employee retention issue entrepreneurs must deal with is compensation. Whereas traditional organizations are more likely to view compensation from monetary rewards (base pay, benefits, and incentives), smaller entrepreneurial firms are more likely to view compensation from a total rewards perspective ( Robbins et al., 2020 ). For these firms, compensation system is highly associated with psychological rewards, learning opportunities, recognition, and economic rewards (i.e., base pay and incentives) ( Iqbal & Inayat, 2022 ). Thus, job retention refers to job retention as employment status rather than length on the same job or at the same employer.

This form of support can help individuals deal with work problems ( Lambert & Hogan, 2010 ). Perceived social support deals with employee’s perception related to others care and help them with work matters ( Davidson & Wang, 2011 ). Perceived social support is defined as the provision of beneficial help for at least two individuals that is intended to buffer psychological distress and enhance emotional well-being ( Lee et al., 2013: p. 380 ). Perceived social support is collegial relationships with coworkers or supervisors ( Robbins & Judge, 2022b ). Therefore, this study argues that perceived social support is valuable because it provides psychological support, assistance, feedback, and motivation for employees.

The compensation system represents an important research variable to organizations and the field of HRM to predict individual employee behavioral and attitudinal outcomes ( Kumar, 2016 ). This study highlights that key sub-dimensions of HRM practices, such as compensation, work-life balance, and employee engagement, are vital to employees and influence employee retention ( Ngwenya & Aigbavboa, 2017 ). Key contributors to job satisfaction include compensation, pay, trust between employees and senior management, and job security, which use their knowledge, skills, and abilities to handle work effectively ( Chong et al., 2020 ). Compensation is imperative for the employees as it offers a sense of security, autonomy, and improved self-worth that could lead to employee commitment and productivity ( Chong et al., 2020 ). The compensation system is one of the most important HRM practices to motivate employees in the workplace. Chong et al. (2020) investigated the impact of the compensation system, which is one of the most important sub-dimensions of human resource management practices (i.e., planning & recruitment, training programs, compensation, and performance appraisal) on employee satisfaction among 302 Jordanian telecommunication companies employees and found that compensation system significantly affects employee satisfaction. According to Herzberg’s theory of motivation, employee motivation is associated with a psychological process of an individual that arouses our interest in doing something and directs and guides behavior ( House & Wigdor, 1967 ), which positively correlates to job satisfaction ( Kinicki, 2021 ). Compensation system plays a significant role in determining employees’ job satisfaction. Based on the above research arguments, thus study proposes the following hypothesis:

Hypothesis 1: Compensation system has a positively significant impact on teacher job satisfaction.

Perceived organizational support is expected to be significantly positively related to employee career satisfaction ( Armstrong‐Stassen & Ursel, 2009 ). Organizational support (i.e., supervisors, managers, and leaders) can help increase job satisfaction by ensuring new employees understand their new job tasks and feel connected to others in the organization ( Kinicki, 2021 ). Perceived organizational support is associated with employees’ general belief in their organization, the value of their contribution, and care about their well-being. High levels of perceived organizational support increase job satisfaction and lower turnover ( Robbins et al., 2018 ). Therefore, this study assumes that older workers who perceived greater organizational support would be more satisfied with their career than those who experienced less organizational support. Then, the following hypothesis is proposed:

Hypothesis 2: Perceived organizational support has a positively significant impact on teacher job satisfaction.

Job involvement and job satisfaction have positively correlated ( Mathieu & Farr, 1991 ). Many years of research have demonstrated that job involvement is moderately related to job satisfaction ( Brown, 1996 ). Job involvement is crucial in improving employee satisfaction in the Turkish hospitality industry ( Kuruüzüm et al., 2009 ). They concluded that there is a significant relationship between job involvement and job satisfaction ( Zopiatis et al., 2014 ). High levels of employee job involvement are related to fewer absences ( Kumari, 2011 ), lower resignation rates, and higher employee engagement and job satisfaction ( Robbins et al., 2018 ). There is a positive link between employee job involvement and job satisfaction among millennial workers ( García et al., 2019 ). The influence of job involvement on job satisfaction is confirmed by Fernández-Salinero et al. (2020) . In recent studies, Gajić et al. (2021) are unequivocal in their claim that job involvement significantly influences employee’s job satisfaction. According to self-efficacy theory, researcher indicates that teacher self-efficacy is crucial for retaining satisfied and effective teachers and maximizing student achievement ( Perera et al., 2022 ). Therefore, the following research hypothesis is proposed:

Hypothesis 3: Job involvement has a positively significant impact on teacher job satisfaction.

Retention shortages may be caused by a range of push and pull factors (i.e., low wage level and low employee satisfaction) at the individual level, resulting in increased job turnover ( Rubin & Dierdorff, 2011 ). Holland et al. (2017) study demonstrated this, evidencing the relationship between effective job satisfaction mechanisms, strong engagement, and reduced turnover or increased levels of job retention. With push factors, the problem is dissatisfaction with the job or the organization, leading to an unwanted turnover. Job satisfaction may also indirectly affect labor productivity because higher satisfaction decreases turnover rates and absenteeism and increases company commitment ( García et al., 2019 ). The main reason so many fails is the need for mechanisms for picking up signs of job dissatisfaction. If there is no opportunity to voice concerns, unhappy employees will inevitably start looking elsewhere ( Derek et al., 2020 ). Job satisfaction has a moderately strong negative relationship with employees’ turnover intention ( Kinicki, 2021 ). Therefore, the key antecedent of job satisfaction has recently resurged due to relevant implications for organizations in designing management strategies and, thus, reducing turnover and absenteeism and improving employee performance. Then, the following hypothesis is proposed:

Hypothesis 4: Job satisfaction has a positively significant impact on teacher job retention.

In educational context, teachers who experienced satisfaction at their school and satisfaction with the teaching profession were more likely to remain ( Perrachione et al., 2008 ). In particular, previous studies have argued that perceived social support may play a prominent role in enhancing expatriate adjustment and performance in cross-cultural contexts ( Lee et al., 2013 ). Given its significance, it is important to explore how different forms of perceived social support may be associated with job satisfaction and job retention ( Lambert & Hogan, 2010 ). This study assumes that individuals’ ability to adapt to high perceived social support has been positively related to their better job satisfaction and reduced turnover. Mwiti et al. (2021) indicated that job satisfaction is strongly associated with leaving the profession. This study assumes that when a teacher has highly perceived social support, it will strengthen the relationship between job satisfaction and teacher job retention. This assumption explains that when teachers have highly perceived social support, they intend to have a lower perception of leaving even if they are dissatisfied with the job. Therefore, the following hypothesis is proposed:

Hypothesis 5: The link between job satisfaction and job retention is strong when the perceived social support is also high.

The conceptual framework of this study aims at exploring the antecedents and consequences of job satisfaction on teachers’ job retention in HEIs in Cambodia. We also applied the existing theoretical background to explain the critical research hypotheses proposed in Figure 1 . Indeed, the moderating role of perceived social support was explained to strengthen the relationship between job satisfaction and job retention.

This study consists of five key research variables: compensation system, perceived organizational support, job involvement, job satisfaction, and job retention. The compensation system has five items adopted by Shah (2019) . Perceived organizational support consists of 3 items adopted from Lee et al. (2013) . Job involvement has six items adopted from Kanungo (1982) . Perceived social support has four items adopted from Lee et al. (2013) . Job satisfaction has seven items adopted from Wanous and Lawler (1972) . Teacher’s retention has four items adopted from Ashton (2018) . All questionnaire items used a 5-point scale and double-back translation (i.e., English-Khmer-English) to validate the meaning of the questionnaire translation, as attached in the Appendix.

Non-probability sampling technique (i.e., purposive and snowball sampling) ( Cooper & Schindler, 2014 ) was adopted to invite senior lecturer, assistant professors, associated professors, and professor to participate in the survey. A self-administered and online survey was also adopted to collect information from respondents. An unknown population is used to determine the sample sizes for this study. As recommended by Bowerman et al. (2019) to collect data from HEIs, as shown

$n=p\left(1-p\right){\left(\frac{Z_{\alpha /2}}{B}\right)}^2$,

where: n = sample sizes; p = Probability is 0.5 (50%); $Z_{\alpha /2}$ = Significant level at 1.96 with confident interval of 0.05; B = Tolerance Error is 0.07 (7%).

$n=0.5\left(1-0.5\right){\left(\frac{1.96}{0.07}\right)}^2=196$

This study applies structural equation modeling (SEM) to test the proposed research hypotheses. Thus, the formal sample size should be at least 196 respondents. We distributed 450 questionnaires to 58 HEIs, allowing us to invite their employees to participate in our survey. We received returns from 389 participants, including a formal 36 HEIs. However, 60 questionnaires were withdrawn because they contained serious missing data. Therefore, we selected 329 participants (with 329/389 = 84.57%) from 36 HEIs as the valid samples for this study. We requested each HEI to select 5 - 7 senior lectures, assistant professors, associated professors, and professors who are currently teaching full-time or part-time in HEIs to participate in the survey. Saunders et al. (2019) suggested that since mail survey questionnaires typically have a response rate between 30 and 50 percent, self-administered questionnaires should undoubtedly have a response rate of more than 50 percent. Therefore, the study viewed the response rate of 84.57 percent as satisfactory.

We used frequency analysis to condense and structure data pertaining to respondents’ personal information. Researchers frequently use frequency distribution as a data analysis approach because it offers a convenient overview of the data. This method offers precise values and percentages of certain variables targeted by the researcher ( Bennett et al., 2022 ). Frequency analysis utilizes pie charts, bar charts, histograms, and cross-tabulations as effective methods for interpreting and analyzing respondents’ personal information, such as gender, age, employment, education, and income. The result of Table 1 indicates that among 329 participants, 279 are male (84.8%) and 50 are female (15.2%). The findings showed that lecturers (163; 49.5%) were the dominant participants in this study. Indeed, about 66% of the participants (217) hold a master’s degree. Thus, more than 84% of male respondents have participated in this study.

Descriptive statistics provide researchers with numerical and graphical methods for summarizing acquired data in quantitative terms. Furthermore, it helps the researcher streamline extensive data in a practical and logical manner ( Jaggi, 2016 ). We utilized descriptive statistics to present the data, displaying the means, standard deviation, the overall background of respondents, and the variables created.

Table 2 shows that “perceived social support” has a mean value ranging from 3.65 (i.e., PSS1—My supervisor supports me with necessary human resources in my work) to 3.82 (i.e., PSS2-My supervisor supports me with necessary finances during my work). Thus, these research variables approach the level of “agree”. Indeed, the standard deviation of this study is less than 1, indicating that respondents’ perceptions of rating scores are not significantly different from each other. This data raises concerns about outliers.

The purpose of performing factor analysis is to define the underlying structure among the research variables ( Hair Jr. et al., 2019 ). To ensure a high level of validation for the results, we employ the following guidelines to assess the factor analysis process: 1) KMO > 0.50 and Bartlett’s test is significant at p-value < 0.05; 2) Communalities of each variable should be higher than >0.5; 3) Factor Loading ≥ 0.60 or 0.70; 4) Eigenvalue ≥ 1; 5) Cumulative percentage > 60%. This report also used the “Varimax rotation” and “Principal Component Extraction” methods to validate the meaning of research variables. Therefore, we may delete some research items if they have a factor loading score below 0.60 or communality values below 0.50. We adopted the factor analysis procedure to conduct the reliability test, leaving the items intact. We commonly apply the estimate of the coefficient Alpha, an indicator of convergent validity measurement, to evaluate the reliability of questionnaire items ( Hair et al., 2010 ). According to Churchill Jr. ( Churchill Jr., 1979: p. 68 ), coefficient Alpha “absolutely should be the first measure one calculates to assess the quality of the instrument”. To assess its representation of structure, one should apply common factor analysis and reliability tests ( Hair et al., 2021 ). We use two criteria for the reliability test to check the construct validity: 1) the item-to-total correlation must be at least 0.5; 2) the coefficient alpha (α) must be at least 0.6 (for an exploratory study) or 0.70 (for a confirmatory study) (see Table 3 ).

Source: Hair Jr. et al. (2019) .

According to Table 3 ’s threshold values applied to evaluate the results of Table 4 , all questionnaire items are valid (i.e., all questionnaire items have an FL greater than 0.70) and reliability coefficient (α = 0.852, or 85.20% > 70%). Thus, these research variables are reliable and valid for this study. According to Table 4 , α = 0.85.

Note: FL = Factor loading (score), KMO = Kaiser-Meyer-Olkin, Eig. = Eigenvalue, CU% = Cumulative %, ITC = Item-total correlation, α = Cronbach Alpha.

According to Table 3 ’s threshold values and Table 5 ’s results, all questionnaire items are valid (i.e., all items have an FL greater than 0.70) and reliability coefficient (α = 0.91, or 91% > 70%). Thus, these research variables are reliable and valid for this study.

According to Table 3 ’s threshold values and Table 6 ’s results, all questionnaire items are valid (i.e., all items have an FL greater than 0.70) and reliability coefficient (α = 0.92, or 92% > 70%). Thus, these research variables are reliable and valid for this study.

According to Table 3 ’s threshold values and Table 7 ’s results, all questionnaire items are valid (i.e., all items have an FL greater than 0.70) and reliability coefficient (α = 0.89, or 89% > 70%). Thus, these research variables are reliable and valid for this study.

According to Table 3 ’s threshold values and Table 8 ’s results, all questionnaire items are valid (i.e., all items have an FL greater than 0.70) and reliability coefficient (α = 0.79, or 79% > 70%). Thus, these research variables are reliable and valid for this study.

According to Table 3 ’s threshold values and Table 9 ’s results, all questionnaire items are valid (i.e., all items have an FL greater than 0.70) and reliability coefficient (α = 0.91, or 91% > 70%). Thus, these research variables are reliable and valid for this study.

Correlation is the measure of the size and direction of the linear relationship between the two variables, while squared correlation is the measure of the strength of the association between them ( Tabachnick et al., 2018 ). The correlation matrix ( Table 10 ) illustrates the interrelationship among key research variables as proposed in the conceptual framework. All of the variables in this study have a strong positive relationship with each other, as shown in Table 10 . The Pearson correlation coefficient test with a two-tailed test gave this relationship with **p-value < 0.01. We evaluated the correlation between the variables using the correlation matrix ( Steiger, 1980 ). The study’s findings demonstrate that among these relationships, “perceived social support” exhibits a highly and strongly significant correlation with teachers’ job satisfaction, with a correlation coefficient of approximately r = 0.790, or 79.0%. Indeed, “compensation system” also has a strongly significant correlation with “teacher’s job satisfaction” with r = 0.716 or 71.6%, respectively. Thus, in HEI contexts, “compensation systems” and “perceived social support” play the most important roles in increasing teachers’ “job satisfaction” and “job retention”.

**. Correlation is significant at the 0.01 level (2-tailed). $\bar{X}$ = Mean; S = Standard Deviation; PSS = Perceived social support; COS=Compensation system; POS = Perceived organizational support; TRE = Teacher job retention; JOI = Job involvement.

Tabachnick et al. (2018) use regression to predict a score on one variable from a score on the other. Regression analysis describes the relationship between a dependent variable and one or more independent variables ( Daniels & Minot, 2019 ). Correlation and regression are associative techniques that help researchers determine if there is a consistent and systematic relationship between two or more variables ( Hair et al., 2021 ). We begin with the basic scenario of a linear relationship between a single dependent variable and a single independent variable, referred to as a simple regression analysis (Equation (1)). When more than one independent variable or multiple independent variables predict a single dependent variable, we refer to this as a multiple regression analysis (Equation (2)).

Simple regression equation:

$Y={\text{β}}_0+{\text{β}}_1{X}_1+\text{ε}$ (1)

Multiple regression equation:

$Y={\text{β}}_0+{\text{β}}_1{X}_1+{\text{β}}_2{X}_2+\cdots {\text{β}}_n{X}_n+\text{ε}$ (2)

where:

Y is the dependent variable.

X is the independent variable.

β₀ is the constant or y intercept.

β₁ is the slope or coefficient on X.

ε is the error term.

Indeed, Zikmund et al. (2013) define a correlation coefficient as a statistical measure of covariation or association between two variables, indicating the regression weight between the independent and dependent variables. The correlation coefficient represents the power of the significant level of the relationship by considering the significant level of the p-value (i.e., **p < 0.05, *p < 0.1) of the Pearson correlation method.

As suggested by Hair Jr. et al. (2019) (i.e., Table 11 ), the significant relationship between independent variables and dependent variables, R² ≥ 0.10, F-value ≥ 4, and significant level of t-value ≥ |1.96| at p-value < 0.05, respectively. The results of Table 12 of multiple regression analysis showed that “Compensation System” has a positive significant contribution to teacher’s job satisfaction i.e.,— Model1: β₁ = 0.483***, t-value = 10.552 > 1.96 and p-value = 0.000 < 0.001, which Hypothesis 1 is accepted and supported. “Perceived organizational support” also has a positive significant impact on teacher’s job satisfaction—i.e., Model2: β₂ = 0.159***, t-value = 4.189 > 1.96, and p-value = 0.000 < 0.001, which Hypothesis 2 is accepted and supported.

Source: Hair Jr. et al. (2019) .

***p < 0.001, **p < 0.05, *p < 0.1 and significant at t-value > 1.96. Note: COS = Compensation system; POS = Perceived organizational support; JOI = Job involvement. M = Model; β = Beta coefficient.

“Job Involvement” also has a positive significant impact on teacher’s job satisfaction—i.e., Model3: β₃ = 0.291***, t-value = 6.592 > 1.96, and p-value = 0.000 < 0.001, which Hypothesis 3 is accepted and supported. Thus, all three proposed research Hypotheses are well-supported for this study.

The result of Table 13 indicates that “Teacher’s job satisfaction” has a positive significant contribution to teacher’s job retention i.e.— Model4: β₄ = 0.682***, t-value = 16.859 > 1.96 and p-value = 0.000 < 0.001, which Hypothesis 4 is accepted and supported.

***p < 0.001, **p < 0.05, *p < 0.1 and significant level of t-value > 1.96.

We adopt the Baron and Kenny (1986) procedure to test the moderating effect of “perceived social support”. Hair et al. (2013) also recommended the rules of thumb of the hierarchical linear model, which indicated that R² > 0.10, Adjusted-R² > 0.10, F-value > 4, t-value > 1.96, and significant p-value <0.05. The result of Table 14 illustrates that the moderating effect of “perceived social support” has a significant effect on the relationship between job satisfaction and job retention, with β coefficients = 0.098** (or 9.8%), t-value = 2.283 > 1.96, and p-value = 0.023 < 0.05. Therefore, this study reasonably supports and confirms research hypothesis 5. This study concluded that when teachers perceive a high level of social support in their organization, the relationship between job satisfaction and job retention will be stronger.

Note: ***p < 0.001, **p < 0.05 and significant at t-value > 1.96.

$\hat{Y}=a+b_{X1}+c_{X2}+d_{X3}=3.759+0.521X_1\text{+}0.204X_2\text{+}0.098X_3$ (3)

a: Constant of β-value = 3.759.

$b_{X1}$: The independent variable of β-value coefficient = 0.521.

$c_{X2}$: The dependent variable of β-value coefficient = 0.204.

$d_{X3}$: The moderating variable of β-value coefficient = 0.098.