The Effect of the COVID-19 Pandemic on Turnover Intentions among Field Technicians: A Case Study in Philippines

Abstract

The COVID-19 pandemic has caused several disruptions, necessitating adaptation to the current circumstances. The concept of the “New Normal” has been introduced to facilitate coexistence with the virus. Nevertheless, numerous industries saw significant impacts, both in terms of financial losses and personnel attrition. This development has a significant impact on the agriculture industry, particularly on field technicians (FTs). The present study seeks to understand the factors that influence the inclination to leave one’s job among field technicians. A purposive sampling strategy was used to choose fifty-three participants who were then requested to complete a survey-type questionnaire on various factors including perceived supervisor support, workload, perceived alternative jobs, perceived benefits, COVID-19, and job satisfaction. A SmartPLS structural equation modeling (SEM) analysis indicated that job satisfaction did not operate as a mediator in the relationship between turnover intention and its determinants, such as workload, supervisor support, benefits, and employment alternatives. Furthermore, this study verified that the restrictions imposed during the COVID-19 epidemic did not influence the connection between job satisfaction and turnover intention. This study represents one of the initial investigations conducted on workers in the Philippine farm sector during the ongoing COVID-19 pandemic. Ultimately, the discoveries could be utilized to assess the distinct circumstances arising from the current global COVID-19 pandemic.

1. Introduction

The COVID-19 epidemic and its associated containment measures, such as social isolation and travel restrictions, have led to a decrease in the labor force across all economic sectors and resulted in job losses [1]. The impact on the agricultural labor force was significant and posed a threat to food security, particularly in nations with a strong agricultural sector like the Philippines [2]. Due to its direct connection to food production in the Philippines, the agriculture industry was one of the few sectors that continued to operate during the COVID-19 pandemic. Small-scale farmers persisted in cultivating their customary crops. In addition, corporate groups maintained their involvement with smallholder farmers by providing them with crop inputs such as seeds, fertilizers, and crop protection goods, facilitated by agronomists or field technicians (FTs).

The agriculture sector is a vital industry that plays a significant role in producing most staple food products. It has a crucial role in supplying food for consumption to meet the rapidly growing population’s needs [3]. Furthermore, it is often regarded as the essential resource in the supply chain, as it is needed for every individual on the planet [4]. Moreover, the agriculture business encompasses a wide range of systems, which may involve professionals such as agronomists or field technicians. FTs possess knowledge in agriculture, is fluent in the local dialect, and carries out field promotional activities such as conducting crop demonstration trials and organizing farmer meetings. The agriculture sector is a vital industry that plays a significant role in producing the majority of staple food products. It has a crucial role in supplying food for consumption to meet the rapidly growing population’s need [3]. Furthermore, it is often regarded as the essential resource in the supply chain, as it is needed for every individual on the planet [4]. Moreover, the agriculture business encompasses a wide range of systems, which may involve professionals such as agronomists or field technicians. FTs possess knowledge in agriculture, are fluent in the local dialect, and carry out field promotional activities such as conducting crop demonstration trials and organizing farmer meetings.

After completing the official product training, field technicians (FTs) are allocated specific territories or areas to work in. Furthermore, they are regarded as essential personnel, and foreign Talent (FT) serve a crucial function in their employer’s operations. Therefore, it is crucial to ensure that FTs remain actively involved and motivated to sustain their exceptional provision of specialized services to farmers and their employers. Nevertheless, because of the COVID-19 epidemic, numerous industries have experienced a substantial surge in employee turnover. This trend has had an impact on the agriculture sector.

However, in 2021, an employee attrition rate of 20% was reported. Furthermore, the employment and training procedure was significantly more intricate because of the existing limitations imposed by the COVID-19 epidemic. Consequently, the task of recruiting and keeping highly skilled full-time employees poses a considerable difficulty for corporate firms such as XYZ. Crucially, firms must devise a method to retain any remaining high-performing individuals. One possible course of action for an organization is to conduct a study on the factors that influence the intention of its surviving employees to leave the company.

A study examining the turnover intention of foreign workers (FTs) in the agriculture business specifically concentrated on the differences in job satisfaction levels among agricultural technicians in Kenya across various sectors [5]. FTs, like agricultural extension workers in academia and the government sector, primarily serve smallholder farmers due to the nature of their duties. FTs can also be likened to salespeople since they endeavor to promote and market technology to both farmers and retail establishments. Due to the limited scope of job descriptions in FT jobs, it is challenging to directly compare them to equivalent positions in other countries. Furthermore, given the gravity of an organization’s high turnover rate, it is imperative to understand the personnel characteristics that contribute to turnover intention. The selection of field technicians from crop input suppliers in the Philippines during the COVID-19 epidemic is a unique aspect of this study due to these factors. Furthermore, a separate study conducted in the agriculture sector has also discovered a correlation between turnover intention and the level of education [6]. Addressing these knowledge deficiencies can be particularly difficult, particularly when considering behavioral aspects.

The aim of this study is to comprehend the factors that influence the intention of field technicians working for agricultural input providers in the Philippines to leave their jobs. This study specifically examines the factors that influence turnover intention and the interplay between variables such as perceived supervisor support, workload, and perceived alternative job opportunities. This study will additionally validate the presence of the COVID-19 pandemic as a moderating factor and the function of job satisfaction as a mediation factor in relation to turnover intentions. The findings may contribute to a comprehensive comprehension of the turnover intents of field technicians (FTs) within the COVID-19 pandemic, hence assisting firms such as XYZ in devising solutions or cures to retain these crucial personnel.

2. Methodology

2.1. Participants

The survey participants were selected from the remaining fifty-three (53) technicians in XYZ. The objective of this study was to have a minimum of 50 participants. When using a purposive sampling strategy in a study, a sample size of 41 respondents was used, which produced solid results [7]. In order to identify strong effects in basic mediation models, a relatively modest sample size of at least 30 cases is required, as is typically needed in experimental research [7,8]. Furthermore, a study conducted by Dworkin [9] highlighted that numerous studies indicate that a participant range of 5 to 50 is sufficient for qualitative studies. Nevertheless, as the present investigation employs a quantitative methodology, Kock and Hadaya [8] propose that a small sample size is adequate for PLS-SEM when it is established that the effect size is substantial. Therefore, a limited number of participants were chosen due to the significant path coefficients resulting in substantial effect sizes.

2.2. Questionnaire

Ahmad et al. [10] created and verified a job satisfaction survey for healthcare professionals, which included appropriate assessments for the constructs of perceived supervisor support, workload, and perceived benefits. A moderated–mediated model was used to generate inquiries for the constructs of perceived job alternatives, job satisfaction, and turnover intention [11]. New questions have been formulated based on the results of the exit interviews and the author’s understanding of the nature and circumstances of the FT’s work during the ongoing COVID-19 protocols. These questions pertain to the workload’s impact on quality time with the family, perceived benefits such as incentives and health benefits, and the effects of the COVID-19 pandemic. The questionnaire was created according to the conceptual framework, as illustrated in

Table 1. Questionnaire: constructs and measures.

Construct	Item	Measure	Reference
Perceived supervisor support	PS1	I am clear on what my supervisor expects of my job performance.	[10]
	PS2	My supervisor can address my questions or concerns.	[10]
	PS3	My supervisor clearly communicates his expectations.	[10]
Workload	WL1	I am satisfied with the number of my working hours.	[10]
	WL2	My workload is reasonable.	[10]
	WL3	My workload allows me to spend quality time with my family.
Perceived job alternatives	AJ1	If I resign, chances are low that I would be able to find a good or better job.	[11]
	AJ2	It would not be easy to find an acceptable alternative job.	[11]
	AJ3	If I resign, it will take me a long time to find a job as good as this one.	[11]
Perceived benefits	PB1	My pay is fair for my responsibilities.	[10]
	PB2	I received the right amount of incentives for my accomplishments.
	PB3	I am satisfied with the health benefits extended by the company.
The COVID-19 pandemic	CO1	The COVID-19 pandemic and its restrictive protocols worsened my work environment.
	CO2	The COVID-19 pandemic and its restrictive protocols increased my workload.
	CO3	The COVID-19 pandemic and its restrictive protocols made me look for job alternatives.
	CO4	The COVID-19 pandemic and its restrictive protocols made me less satisfied with my job.
Job satisfaction	JS1	I feel satisfied with my job.	[11]
	JS2	I enjoy doing my job.	[11]
	JS3	I am enthusiastic about my job.	[11]
Turnover intention	TO1	I do not think of quitting this job.	[11]
	TO2	I will not look for another job.	[11]
	TO3	I will stay in this job next year.	[11]

. The examination was conducted using a five-point Likert scale.

The respondents were surveyed offline between September and November 2021 within XYZ internal firm to collect data. The questionnaire was distributed via email along with the recommended timetable. A telephone interview was conducted with respondents who were assigned to various regions in the Philippines. A pilot test was conducted on 5 respondents to assess the survey questionnaire. Upon completion, the survey data were processed and analyzed with Microsoft Excel 365 and SmartPLS 3.2.9.

2.3. Structural Equation Modeling

This study employed partial least squares structural equation modeling (PLS-SEM), which is a type of variance-based structural equation modeling. The author employed the structural equation modeling (SEM) approach to validate the presence of job satisfaction as a mediator in the relationship between four drivers (work environment, workload, perceived benefits, and perceived job alternatives) and turnover intention. Additionally, the author examined the moderating role of the COVID-19 pandemic on these proposed relationships. When examining mediation models, methods like PLS-SEM, which is a form of structural equation modeling, are regarded as preferable approaches [12]. Partial least squares structural equation modeling (PLS-SEM) employs an iterative procedure that takes into account the complete structure of the model [13]. PLS-SEM is capable of accurately estimating parameters even when sample sizes are small, because of its reliance on partial regressions [12]. The research models were evaluated using Smart PLS version 3.2.9. The hypotheses that have been tested are presented:

H1. 

Supervisor support has an effect (direct and indirect) on job satisfaction and turnover intention.

H2. 

Workload has an effect (direct and indirect) on job satisfaction and turnover intention.

H3. 

Perceived job alternatives have an effect (direct and indirect) on job satisfaction and turnover intention.

H4. 

Perceived benefits have an effect (direct and indirect) on job satisfaction and turnover intention.

3. Results

3.1. Demographic Data

The data were collected from a sample of 53 male participants, who represented all of the active FTs as of 30 October 2021. The sampling method used was purposive. With the exception of one individual who held a degree in finance, all the participants possessed bachelor’s degrees in subjects linked to agriculture. The mean age was 31 years old, with tenure ranging from 0.3 to 10.2 years. Furthermore, the company’s average organizational tenure stood at 2.1 years. Approximately 41.5% (22 out of 53) of the individuals were in a state of matrimony. All the participants were employed directly by the company and received normal medical and hospitalization benefits.

3.2. Model Analysis

3.2.1. Measurement Model Analysis

Figure 1 depicts the initial model created by SmartPLS to investigate the mediation of job satisfaction and the moderating effect of COVID-19 pandemic limits. The initial step was assessing the measurement model to determine the reliability of the constructs and the extent to which they converge. Figure 1 displays the factor loadings and R2 values. Cronbach’s alpha (α), composite reliability (CR), and the average variance extracted (AVE) were calculated for each latent construct during the initial run. Jahmani et al. [14] suggested that factor loadings should be above the threshold of 0.7 in order to achieve statistical significance. Hair et al. [15] established that the optimal threshold value for CR is 0.70 and also recommended a cut-off criterion value of 0.50 for the AVE. Consequently, five out of the twenty-two indications, specifically WL3, AJ3, PB1, CO1, and CO2, were eliminated. A modified version was created, as depicted in Figure 2 below.

The improved model demonstrated satisfactory values for α, CR, and the AVE, as presented in

Table 2. Reliability and convergent validity tests.

Constructs	Code	Cronbach’s Alpha	Composite Reliability	AVE
Perceived job alternatives	AJ	0.807	0.897	0.814
COVID-19 restrictions	CO	0.512 *	0.799	0.667
Job satisfaction	JS	0.873	0.922	0.798
Perceived benefits	PB	0.554 *	0.813	0.686
Perceived supervisor support	PS	0.804	0.883	0.716
Turnover intention	TO	0.945	0.965	0.901
Workload	WL	0.617 *	0.835	0.718

* Low Cronbach’s alpha values (α < 0.7).

. The poor Cronbach’s alpha values (α < 0.7) reported for the Co, PB, and WL constructs were mitigated by acceptable CR values. Therefore, the establishment of construct dependability and convergent validity has been confirmed.

Subsequently, an evaluation was conducted to determine the extent to which the measurement model has divergent validity. Hair et al. [15] suggested using the Fornell–Larcker criterion (FLC) to assess discriminant validity. The FLC assesses the connection between the constructs and the square root of the AVE for each latent variable [16]. The FLC values derived for the model are presented in

Table 3. Divergent validity: Fornell–Larcker criterion (FLC).

Factors	AJ	CO	JS	PB	PS	TO	WL
AJ	0.902 *
CO	−0.109	0.817 *
JS	0.309	−0.368	0.893 *
PB	−0.008	0.333	−0.061	0.828 *
PS	−0.048	−0.101	0.527	0.049	0.846 *
TO	0.320	−0.573	0.565	−0.250	0.328	0.949 *
WL	−0.077	0.213	−0.467	−0.087	−0.501	−0.403	0.847 *

* FLC values for AVE square root.

below.

Referring to Table 3, it can be observed that the bolded values on the main diagonal (or the square root of the AVE) are all larger than the values located below them. The results validate that the highest values along the diagonal were more statistically significant in relation to the association of the construct with all other constructs. Therefore, the divergent validity of the FLC test was established. The cross-loading factors were calculated and presented in

Table 4. Divergent validity: cross-loadings.

Factors	AJ	CO	JS	PB	PS	TO	WL
AJ1	0.975	−0.162	0.325	−0.048	−0.064	0.381	−0.042
AJ2	0.824	0.051	0.195	0.096	0.004	0.087	−0.146
Co3	−0.154	0.883	−0.322	0.370	−0.032	−0.540	0.229
Co4	0.001	0.744	−0.280	0.140	−0.157	−0.379	0.101
JS1	0.289	−0.381	0.942	0.015	0.487	0.552	−0.465
JS2	0.425	−0.364	0.859	−0.057	0.366	0.371	−0.318
JS3	0.149	−0.251	0.877	−0.123	0.541	0.567	−0.449
PB2	0.121	0.339	−0.038	0.893	0.002	−0.245	−0.076
PB3	−0.190	0.192	−0.070	0.758	0.098	−0.158	−0.068
PS1	0.007	−0.137	0.385	0.015	0.821	0.244	−0.389
PS2	−0.198	−0.058	0.388	0.193	0.813	0.203	−0.472
PS3	0.032	−0.070	0.537	−0.041	0.902	0.357	−0.425
TO1	0.233	−0.522	0.545	−0.224	0.301	0.909	−0.386
TO2	0.348	−0.527	0.524	−0.242	0.308	0.962	−0.412
TO3	0.326	−0.582	0.542	−0.245	0.325	0.976	−0.350
WL1	−0.107	0.030	−0.346	0.044	−0.391	−0.247	0.789
WL2	−0.036	0.290	−0.436	−0.158	−0.455	−0.413	0.901

.

Table 4 demonstrates that the cross loadings did not reveal any problems with divergent validity. This is because the parent construct indicators had a larger correlation with each other compared to their correlation with other constructs. The assessment of divergent validity can also employ the heterotrait–monotrait (HTMT) ratio. Franke et al. [17] suggested that the heterotrait–monotrait ratio should be below 0.85. The HTMT ratio derived for the model is presented in

Table 5. Divergent validity: heterotrait–monotrait (HTMT) ratio.

Factors	AJ	CO	JS	PB	PS	TO	WL
AJ
Co	0.177
JS	0.354	0.552
PB	0.288	0.555	0.166
PS	0.097	0.195	0.606	0.172
TO	0.282	0.805	0.612	0.335	0.362
WL	0.159	0.332	0.618	0.199	0.710	0.510

.

Confirmation of divergent validity for the HTMT ratio test was obtained when all HTMT ratios were found to be below 0.85. Therefore, all three experiments have verified the distinctiveness of the measurement model.

3.2.2. Structural Model Analysis

An assessment of a structural model is conducted using the R2 values, Q2 values, and the relevance of the paths. Figure 3 depicts the structural model that was created using bootstrapping with 500 sub-samples. The model was evaluated using a bias-corrected and -accelerated method, and a one-tailed test was conducted at a 5% significant level.

The original mean values of the path coefficients are being compared to hypothetical mean values. A one-sample t-test compares the mean of a sample population to a fictitious mean [18]. The crucial t value for a one-sample or one-tailed test with a significance level of 5% and 52 degrees of freedom is 2.22. The correlations between AJ-JS (t = 2.684), CO-TO (t = 2.328), and PS-JS (t = 3.252) were statistically significant. Therefore, AJ, Co, and PS had substantial effects on JS, TO, and JS, respectively.

The R2 value for the dependent variable quantifies the magnitude of each structural path. The R2 value of 0.439 indicates that AJ and PS account for 43.9% of the variations observed in JS. The R2 value of 0.541 indicates that AJ and CO are responsible for 54.1% of the variations observed in TO.

The Q2 value quantifies the predictive significance of the endogenous constructs, with a value greater than 0 indicating predictive importance. The findings shown in

Table 6. Predictive relevance of endogenous constructs: Q2 values.

Construct	SSO	SSE	Q2
JS	159.00	113.60	0.286
TO	159.00	89.12	0.439

provide evidence of the statistical relevance in predicting the endogenous constructs. The Q2 figures for JS and TO are 0.286 and 0.439, respectively.

Table 6 demonstrates a statistically significant relationship between the job satisfaction and turnover dimensions. Therefore, the endogenous constructs JS and TO have predictive relevance.

3.3. Mediation Analysis

This study conducted a mediation analysis to examine if JS plays a mediating function in the individual associations between TO and PS, WL, AJ, and PB, as presented in

Table 7. Mediation analysis: total, direct and indirect effects.

Total Effect	β Coefficient	T Statistic	p-Value
AJ → TO	0.276	2.565	0.005
PB → TO	−0.149	1.387	0.083 *
PS → TO	0.203	1.481	0.070 *
WL → TO	−0.212	1.866	0.031
Direct Effect
AJ → TO	0.206	1.919	0.028
PB → TO	−0.126	1.143	0.127 *
PS → TO	0.107	0.719	0.236 *
WL → TO	−0.158	1.370	0.086 *
Co-Mod-JS → TO	0.026	0.123	0.451 *
Indirect Effect
AJ → JS → TO	0.070	1.218	0.112 *
PB → JS → TO	−0.023	0.618	0.268 *
PS → JS → TO	0.097	1.191	0.117 *
WL → JS → TO	−0.053	1.175	0.120 *

* Insignificance path.

.

Mediation is established when the indirect effect is statistically significant [19]. Table 7 demonstrates that the indirect effect, namely, the individual influence of AJ, PB, PS, and WL on TO through the mediator JS, is not statistically significant. Therefore, there is no intervention or interference in any of the four (4) relationships.

When considering paths related to workload (WL), the overall impact (WL-TO) was shown to be considerable, despite the fact that the individual paths for the direct (WL-TO) and indirect (WL-JS-TO) effects were found to be minor. It is important to observe that the sum of the beta coefficients for the direct effect (WL-TO, β = −0.158) and the indirect impact (WL-JS-TO, β = −0.053) accurately equals the beta coefficient for the total effect (WL-TO, β = −0.212). These findings indicate that the influence of WL on TO is substantial, even without the involvement of JS as a mediator. Nevertheless, the influence of WL on TO through JS (or the indirect effect of JS) and the influence of WL on TO in the presence of JS (or the direct effect of JS) are not significant.

Simple mediation models can reveal strong effects with a relatively modest sample size, often as low as 30 cases [7]. However, researchers are advised to exercise cautious when confirming complete mediation [7]. When the sample size is lower, moderation is more likely to be classified as full rather than partial [20]. Given that this study focuses primarily on the presence or absence of mediation rather than the degree of mediation, a sample size of 53 respondents should be sufficient to establish the lack of mediation in the model.

3.4. Moderation Analysis

This study aimed to validate and explore the moderating influence of COVID-19 limits (CO) on the relationship between job satisfaction (JS) and turnover (TO). This study proposed that the limitations imposed during the COVID-19 epidemic can impact the intensity or the orientation of the correlation between work satisfaction and turnover. The findings from Table 3 indicate that JS did not have a moderating effect on the JS-TO association (β = 0.026, t = 0.123, p = 0.451). Figure 4 depicts the correlation between work satisfaction and turnover under different degrees of influence from the COVID-19 constraints.

The link between JS and TO stays consistent, regardless of whether the impact is strong (as indicated by the green line) or low (as indicated by the red line). This is evident from the constant slope of all three lines in Figure 4. The statement affirms that the COVID-19 pandemic does not play a moderating effect in the connection between work satisfaction and turnover.

3.5. Hypothesis Testing

H1, H2, H3, and H4 hypothesized that job satisfaction would operate as a mediator between turnover intention and perceived supervisor support, workload, perceived job alternatives, and perceived benefits, respectively. The indirect effects hypothesized in H1, H2, H3, and H4 were shown to be statistically negligible, as seen by the results presented in Table 3. Therefore, there was no evidence to support the hypothesis that H1, H2, H3, and H4 had a mediating role in the association between work satisfaction and the four relationships.

H5 hypothesized that the correlation between job satisfaction and turnover intention would be attenuated in the presence of a significant impact from the COVID-19 pandemic. The findings from Figure 2 and Table 3 provide evidence that JS does not have a moderating effect on the JS-TO association (β = 0.026, t = 0.123, p = 0.451). Similarly, H5 did not obtain empirical evidence to show its position as a moderator in the context of the COVID-19 pandemic.

3.6. Alternative Models for Consideration

This study proposed that job satisfaction plays a mediating function and that the COVID-19 pandemic has a moderating role. The results did not corroborate the five (5) study hypotheses. These findings are notably applicable to the modified structural model depicted in Figure 4. This study investigated different model scenarios to gain further understanding of the connections between job satisfaction, turnover intention, and the factors that contribute to turnover intention during the ongoing COVID-19 pandemic. An investigation was conducted where JS acted as the moderator, as depicted in Figure 5.

Figure 5, which displays the results of the simple slope analysis, demonstrates that the different amounts of JS had no effect on the direction or steepness of the line that represents the link between CO and TO. A bootstrap route weighting scheme provides evidence that the moderating influence of JS on the CO-TO relationship is statistically insignificant (β = 0.020, t = 0.152, p = 0.440). Furthermore, the statistical analysis presented in

Table 8. Job satisfaction as possible moderator.

JS as Moderator on	β Coefficient	T Statistic	p-Value
CO → TO	0.020	0.152	0.440
PB → TO	−0.020	0.185	0.427
PS → TO	0.097	0.523	0.300
WL → TO	0.087	0.404	0.343
AJ → TO	0.018	0.181	0.428

does not provide evidence to support the idea that JS plays a moderating function in the other linkages within the model. Table 8 does not provide evidence of job satisfaction playing a moderating function in the specified model trajectories.

4. Discussion

In 2021, the Philippines implemented health and travel restrictions in response to the ongoing COVID-19 outbreak. XYZ, a business organization operating in the food sector, was one of the few permitted by the government to continue its operations without interruption during the period known as the “New Normal”. Consequently, numerous organizations encountered a substantial rate of employee turnover as a majority of employees preferred remote work than physically commuting to the workplace. Regrettably, the XYZ operation experienced a turnover rate of 24% within a condensed five-month timeframe in 2021. This study examined the factors that influence the intention of field technicians (FTs) to leave their job during the COVID-19 epidemic. This study proposed that job satisfaction would operate as a mediator in the association between turnover intention and four identified factors: perceived supervisor support, workload, perceived alternative jobs, and perceived perks.

Furthermore, the analysis indicated that the COVID-19 limits will have a moderating effect on the connection between job satisfaction and turnover. The results of partial least squares structural equation modeling (PLS-SEM) indicate that there is no empirical evidence to support the idea that job satisfaction (JS) plays a mediating function, or that COVID-19 has a moderating effect. Empirical evidence did not support the use of job satisfaction as a moderator in an alternative model. These findings are directly relevant to the XYZ scenario, derived from the responses of survey participants between September 2021 and November 2021.

4.1. Theoretical Contribution

4.1.1. Applicability of PLS-SEM

PLS-SEM has demonstrated its convenience as a technique for assessing the pertinent variables in the investigation. Indicators that passed the appropriate statistical tests for construct reliability and validity accurately represented latent variables such as turnover intention and its drivers, work satisfaction, and the effects of the COVID-19 pandemic. The SmartPLS software employed for conducting PLS-SEM analyses demonstrated efficacy in the simple mediation and moderation model of this investigation. Partial least squares structural equation modeling (PLS-SEM) accurately calculated parameters with a small sample size of 50 participants in this study. Additionally, it achieved adequate statistical power in evaluating the model, conducting mediation studies, and performing moderation analyses.

The outcomes of the PLS-SEM analysis have verified the reliability of the construct and the convergent validity of the measurement model. Confirmation of divergent validity was achieved by the utilization of tests that employed the Fornell–Larcker criterion, cross-loading factors, and heterotrait–monotrait ratio. Once the reliability and validity of the measurement model were established, the structural model was also developed and confirmed for collinearity. Additionally, the importance of the path, the strength of the structural path (R2), and the predictive relevance of the endogenous components (Q2) were also confirmed.

Regarding the mediation study, the results did not offer empirical evidence that validate the role of work satisfaction as a mediator in the individual link between turnover intention and each of the four identified determinants of turnover intention. Conversely, the research conducted by Richards and Kieffer [21] demonstrated that job satisfaction, specifically in terms of job support and workload, has a direct impact on employment retention among associate-level nurses. The data indicate that a higher workload and insufficient help and guidance from supervisors lead to higher rates of job turnover [21,22]. Salahat and Al-Hamdan [23] found that work-related factors, such as compensation, had a considerable impact on both job satisfaction and turnover intention. Contrary to expectations, the results did not offer any empirical evidence indicating that COVID-19 limits play a moderating influence in the link between job satisfaction and turnover intention. Weiss et al. [24] found that employees with higher status had a significant decrease in job satisfaction during the initial phase of the COVID-19 epidemic.

4.1.2. Support for Past Studies on Job Satisfaction and Turnover Intention

The results corroborated the findings of previous studies, namely, on perceived job choices, COVID-19 restrictions, and reported supervisor support, in relation to job satisfaction and turnover intention. The findings of the present study demonstrate that the perception of alternative work opportunities has a substantial influence on an individual’s level of contentment with their current job. Furthermore, the research conducted by Rahman [25] corroborated this study’s findings by establishing a connection between job satisfaction and loyalty and commitment. Based on the findings, the respondents who were full-time employees showed a high level of job satisfaction due to a limited number of other job options. The COVID-19 limits had a negative influence on turnover intention. These restrictions resulted in a decrease in the workforce and led to significant job losses [1]. Labor in the agriculture industry was negatively impacted by the same restrictions [2]. This study found that carbon monoxide (CO) has a notable effect on total output (TO). The FTs hypothesized that the COVID-19 limits had a detrimental impact on job opportunities and job satisfaction, leading to a decrease in their desire to leave their current job.

In addition, a previous study examined the influence of perceived supervisor support (PS) on job satisfaction (JS) and found that PS decreases turnover (TO) [26]. Another study determined that TO is significantly and negatively correlated with JS [27]. The study found that PS had a notable influence on JS. Indeed, the study found a strong positive correlation between the perception of assistance from supervisors and job satisfaction among the FTs. All the remaining pathways in the model had insignificant correlations. Therefore, the results of this study provided evidence that corroborates prior research on the influence of workload, perceived advantages, and work satisfaction on turnover intentions.

4.1.3. Foundation for Future Studies

The scarcity of literature on the application of macro-ergonomics in agriculture is exacerbated by the additional challenge posed by the ongoing COVID-19 pandemic. This study, which focuses on workers in the Philippines’ farm businesses during the COVID-19 pandemic, offers valuable insights for future macro-ergonomics studies in the field of agriculture.

4.2. Practical Implications

4.2.1. For Trading Company A

The current study’s conclusions could perhaps offer valuable insights into the predicament of XYZ. The COVID-19 phenomenon is an unpredictable factor that XYZ cannot control. Nevertheless, XYZ can prioritize the management issues that are fully under its control. Perceived supervisor support, perceived job alternatives, and workload are important factors to consider when initiating internal process improvement. Ensuring the appropriate hiring and training procedures for supervisors can lead to the establishment of robust supervisor support. The protocol governing the communication process between supervisors and FTs can be audited and enhanced as needed. XYZ can enhance the image of career alternatives by bolstering its competitive position within the industry. Efforts should be made to enhance and effectively convey the strong position of XYZ to potential job candidates, in order to establish a clear perception of superiority over other job options. The burden extends beyond the sheer quantity of hours spent working. FTs should possess a comprehensive comprehension of the routine tasks that need to be accomplished, which will serve as a solid basis for establishing a manageable workload in the perspective of FTs. XYZ should consistently adopt a robust procedure for monitoring the level of job satisfaction among FTs. Crucially, XYZ must demonstrate genuine commitment in tackling concerns and implementing measures outlined in these routine job satisfaction surveys to gain insight into the present state of their workforce.

4.2.2. For Other Business Organizations

The results of this study can be extrapolated to other businesses that are comparable to XYZ, especially those facing similar challenges. Organizations in related agricultural sectors can also gain advantages by understanding the factors that influence employees’ intention to leave and their level of job satisfaction. These organizations are involved in the marketing and distribution of agricultural inputs such as fertilizers, herbicides, animal feed, and animal health products. One significant similarity they have with XYZ is the utilization of frontline workers who perform similar tasks as FTs, such as interacting with farmers and channel partners to promote their organizations. It is reasonable to assume that the attitudes and objectives of these frontline workers during the COVID-19 epidemic are likely similar to those of the foreign workers in XYZ.

The findings of this study can be applied to non-agriculture business groups by making slight modifications to the underlying concepts and their accompanying measurements. Field-based representatives who are responsible for sales and marketing of products like fast-moving consumer goods (FMCG) and pharmaceuticals can be appropriate targets for interventions aimed at addressing turnover intention during the COVID-19 pandemic, based on similar analyses. One of the promising opportunities is government employees engaging in community-based services. Gaining insight into the tastes and perspectives of these government personnel is crucial for ensuring their continued commitment to their assigned areas. Implementing a relevant PLS-SEM analysis might reduce the likelihood of losing these employees due to superior job opportunities, excessive workload, inadequate supervisor support, or perceived inferior benefits.

4.2.3. For Government Policy Makers

Government authorities must formulate legislation to facilitate the flourishing of the Philippines’ economy in the midst of the ongoing COVID-19 pandemic. Staff turnover incurs fees for an organization that is already burdened by higher costs due to the COVID-19 constraints. Efforts to protect the physical and mental welfare of employees should be counterbalanced by measures aimed at revitalizing the economy. Business companies that employ frontliners such as the foreign talent (FT) of XYZ require government assistance. Similarly, government measures should refrain from exacerbating the responsibilities of frontline workers such as foreign talent (FT). A valuable starting point is to acknowledge the presence of turnover intention and the various reasons that contribute to this intention among frontline employees, such as full-time workers.

4.2.4. Post-Pandemic Implications

The findings of the current study highlight the impact of COVID-19 on the employees. Even after the COVID-19 pandemic, its effect on the whole world has been devastating in various aspects, including the psychological effect on employees. This may lead to significant shifts in the way employees work and interact with each other [28]. Employees would still be evaluating their decisions based on their experience with COVID-19. This was also revealed in several studies that the pandemic has led to changing employee attitudes regarding future job employment, well-being, remote and flexible work, and job satisfaction [29,30]. Companies may utilize these findings to re-evaluate their policies and working arrangements as the COVID-19 pandemic has since affected traditional work practices.

5. Conclusions

There is still a lot to learn about the dynamics that influence turnover intention. Further contributing to the intricacy is the ongoing COVID-19 epidemic. The exit interview findings from the resigned full-time employees indicated that workload, perceived job alternatives, and perceived supervisor support were identified as potential factors influencing their intention to leave the company. An additional factor that was considered is the potential impact of perceived rewards on an individual’s intention to leave their current position. The concept of job satisfaction acting as a mediator and the influence of COVID-19 as a moderator may have added complexity to what could have been a straightforward situation of remaining in one’s current job during times of uncertainty.

Supervisor support did not have a mediating effect on the association between job satisfaction and turnover. High supervisor support alone results in reduced employee turnover. The impact of supervisor assistance on turnover intention remained unchanged with the inclusion of job satisfaction as a potential mediator in the model. According to this study, the introduction of work satisfaction as a mediating variable did not influence their impact on turnover intention. The COVID-19 epidemic led to travel restrictions that impacted the activity of the field technicians. H5 proposed that the COVID-19 pandemic will impact the relationship between work satisfaction and turnover, acting as a moderating variable. The study findings indicated a different outcome. The COVID-19 travel limitations led to a decrease in job opportunities. Curiously, the majority of interviewees reported that their job satisfaction had improved, despite having to spend more time traveling. They developed a greater sense of gratitude towards their current employment in light of the uncertain circumstances. Therefore, the COVID-19 pandemic did not have a moderating effect on the relationship between work satisfaction and turnover. Given the results of this study, it would be intriguing to explore the role of perceived career alternatives as a potential moderator in the context of COVID-19.

This paper is subject to various constraints. Initially, the factors influencing employees’ intentions to leave were identified by analyzing the information gathered during departure interviews with former full-time employees. Consulting a comprehensive literature review from comparable situations such as XYZ might be beneficial. Furthermore, data were gathered within the timeframe of September to November 2021, which was impacted by the COVID-19 pandemic. This technique would limit the generalizability of the findings in this study to the specific conditions and context of XYZ. Furthermore, expanding the range of participants in the survey would result in a larger sample size, hence enhancing the reliability of the statistical analysis. SEM analyses impose restrictions on the methodology, hence creating the possibility of exploring alternative analysis approaches. In future studies, it is recommended that researchers not only include female participants but also perform a multigroup analysis. It is recommended that further research be carried out in various nations or through a longitudinal study.