Emotional Health in the Workplace

Emotional health is associated with lifestyle habits that promote a balanced psychological state, enabling individuals to manage emotions, maintain well-being, and prevent emotional disorders (^{Diener et al., 2019}; ^{Fredrickson, 2001}). However, the absence of emotional disturbances does not necessarily equate to complete emotional health. Factors such as a lack of defined goals, emotional emptiness, and unchecked anxiety may create imbalances (^{Suh & Punnet, 2022}).

In the workplace, emotions significantly influence organizational relationships (^{Dubreuil et al., 2021}). Positive emotions can improve performance, engagement, and resilience, enhancing motivation and promoting employee social connections (^{Diener et al., 2019}). Conversely, negative emotions might result in counterproductive work behaviors, decreased job satisfaction, and heightened emotional distress (^{Bauer & Spector, 2015}; ^{Nikolaev et al., 2019}).

An established framework for emotional health within organizations can strengthen employee resilience, psychological well-being, and work-life balance. Conversely, negative workplace factors such as excessive workload, time pressure, poor interpersonal relationships, and unsupportive organizational culture can impair emotional health (^{Koinis et al., 2015}; ^{Marcucci & Santos, 2021}).

Additionally, promoting positive emotions in professional settings encourages cooperative behaviors and mitigates workplace stress. Research has shown that recognition, appreciation, and gratitude significantly enhance employees psychological functioning and well-being (^{Di Fabio et al., 2017}; ^{Merino & Privado, 2015}). ^{Paakkanem et al. (2021)} emphasized that acknowledging positive emotions among co-workers strengthens interpersonal relationships and empathy, contributing to a more supportive work environment.

Given these considerations, this study aims to investigate the role of emotional health in the workplace, evaluating its relationship with positive and negative effects through the following hypotheses:

Positive and Negative Emotions in the Workplace

Positive emotions are associated with work performance, job satisfaction, and overall wellbeing, reinforcing employees' commitment and engagement (^{Fredrickson, 2001}; ^{Lyubomirsky et al., 2005}). These emotions expand cognitive and behavioral flexibility, enabling individuals to navigate workplace challenges more effectively (^{Diener & Seligman, 2002}; ^{Fredrickson, 1998}). ^{Sansone and Sansone (2010)} emphasized that gratitude enhances well-being, and in professional settings, promotes intrinsic motivation, job performance, and emotional stability (^{Armenta et al., 2017}; ^{Di Fabio et al., 2017}).

In addition, workplace recognition and appreciation have a positive impact on employee engagement, reinforcing motivation and job satisfaction (^{Fagley, 2018}; ^{Merino & Privado, 2015}). From a neurophysiological perspective, positive emotions contribute to psychological resilience and emotional health, improving employees' ability to manage stress and adapt to organizational changes (^{Alexander et al., 2021}).

In contrast, negative emotions in the workplace are associated with stress, burnout, and reduced job satisfaction (^{Nikolaev et al., 2019}). Such emotions, including anger and anxiety, can lead to dissatisfaction and negatively impact the organizational climate (^{Gross, 2015}; ^{Katana et al., 2019}; ^{Madrid et al., 2015}). ^{Bauer and Spector (2015}) showed that experiencing negative emotions in the workplace increases counterproductive behaviors, while ^{Leung and Lee (2014)} found that guilt and shame contribute to emotional distress and reduced motivation.

Furthermore, neuroscience studies suggest that emotions are processed through two mechanisms: one linked to reward and positive experiences (e.g., satisfaction, pleasure, and joy) and another associated with punishment and negative emotional responses (e.g., sadness, frustration, and stress) (^{Balestrin & De Marco, 2019}). ^{Fisher (2019)} argued that negative emotions have a greater impact in the workplace, as they tend to trigger stressful and conflict-prone situations more often than positive emotions.

However, researchers have suggested that positive emotions counteract the effects of negative emotions, promoting emotional regulation and resilience (^{Gupta, 2019}). ^{Silvestre and Vandenberghe (2013)} highlighted that positive emotions reduce stress and enhance interpersonal relationships, cognitive flexibility, and coping strategies. Employees who consistently experience positive emotions report higher well-being, job satisfaction, and organizational commitment (Alexandre et al., 2021; ^{Diener & Seligman, 2002}).

^{Colombo et al. (2021)} also that mindfulness and gratitude-based strategies significantly improve emotional regulation. These findings underscore the importance of developing and implementing workplace interventions that promote positive emotions and emotional intelligence to enhance well-being and productivity.

This study establishes a clear relationship between emotional health and workplace affect, aiming to provide organizations with a structured assessment tool to evaluate and foster employee well-being. The hypotheses, based on psychological and organizational behavior research, provide a foundation for future workplace interventions aimed at enhancing emotional health and professional performance.

Data Analysis

The data were imported into a Microsoft Excel spreadsheet (version 365) and analyzed by the SPPS® Software (version 26) and SmartPLS® version 4.1.0.9. The data were presented in tables and graphs expressed in absolute values and percentages and the instrument validation statistics.

First, construct validation was applied using the exploratory factor analysis method for positive and negative emotions to verify the relationship between indicators and dimensions (factor loadings). Subsequently, the partial least squares structural equational modeling (PLS-SEM) method was applied to estimate the relationships between the dimensions of the EHW Inventory and the PANAS by assessing the convergent validity of the proposed scale. To rank the dimensions of the instruments, the score standardization technique (Equation 1) was applied to make the dimensions of the scales comparable (^{Lopes, 2018}). A score of 0.00 to 33.33 is classified as low, 33.34 to 66.67 as moderate, and 66.68 to 100 as high.

Where:

Ss_i = standardized score for dimension I.

Sum = sum of valid scores for dimension I.

Minimum = lowest possible score for dimension I.

Maximum = highest possible score for dimension i.

Demographic Characteristics

The demographic and social data, including the Brazilian region where they work (work or study), gender, age group, and marital status, are listed in Table 1. Of the 483 respondents, 367 (67.09 %) live in southern Brazil, 275 (56.94 %) are male, 214 (44.31 %) are adults up to 30 years old, and 263 (54.45 %) are not married (single, separated, or widowed).

Consistency of the Instrument and Model Fit

The path model was fitted for 300 iterations, but stabilized after 7 iterations. The bootstraping technique was used for 5.000 subsamples, and the predictive relevance criterion of the model was determined by the blindfolding technique. The model stabilized after seven iterations, and the model fit was determined by standardized root mean square residuals (SRMR), square euclidean distance (SED), geodesic distance (Gd), and the normed fit index (NFI). The results confirmed that the proposed structural model fit the data with acceptable indices (SRMR = 0.073, SED = 7.135, Gd = 1.718, NFI = 0.812). The SRMR value was below 0.08 and the NFI value was above 0.8 (^{Henseler et al., 2016}), indicating a satisfactory and adequate structural model.

Table 1 Statistical details of respondents' demographic and social characteristics 

Note: n = number of individuals; % = percentage.

Correlation, Multicollinearity, and Descriptive Assessment of Dimensions

The values of the correlations between the dimensions and their significance are shown in Figure 1. The highest correlation was between negative emotions (NE) and negative affect (NA) (r = 0.693) and positive emotion (PE) and positive affect (PA) (r = 0.654), while the lowest correlation was between PA and NA (r = -0.174). We observed that all correlations were significant and, according to ^{Hair et al. (2019)}, were acceptable and did not compromise the structural model. Another relevant aspect concerns the dispersions: the positive emotions with positive affect and the negative emotions with negative affect. For the descriptive evaluation (Table 2), the mean and the standard deviation of the dimensions were calculated, and the mean values ranged from 2.01 to 2.78. Standard deviation values ranged from 0.903 to 0.995, showing that the means were representative for eachNof the dimensions.

Table 2 Factorial loadings, VIF, reliability, and average extracted variance 

Measurement Model Assessment

The factorial loadings, variance inflation factor (VIF) values for assessing the interrelationships between observed variables and their respective latent variables, as well as Cronbach's alpha, average variance extracted (AVE), and composite reliability (CR) are listed in Table 2. The VIF values ranged from 1.290 to 3.860, meeting the criterion of non-collinearity (i.e., VIF < 5.0) (^{Joseph et al., 2010}), while the alpha and CR values met the assumptions (0.7 < 0 < 0.95). Lastly, the AVE proposed by ^{Hair et al. (2022)} was also met (AVE > 0.5). The alpha of the PA was higher than the original scale 0.926 > 0.840, and the NA was lower than the original scale 0.894 < 0.90.

Figure 1 Pearson correlation matrix and descriptive statistics 

As for the discriminant validity of the model, the model was tested by the Heterotrait Monotrait Ratio (HTMT), as it provides the true estimate of the correlations between the dimensions and, therefore, better control of the relevance and accuracy of the proposed model (^{Henseler, 2011}; ^{Henseler et al., 2015}). The HTMT values were excellent, as the upper limit of the HTMT for the 95 % confidence interval, calculated using the bootstrapping method, was below 1. The values found were between 0.310 and 0.952, which are considered acceptable for the discriminant validity of the model (^{Hair et al., 2022}) (Table 3).

Table 3 Discriminant validity of the model by HTMT criterion* 

Note. * 95% confidence interval.

Structural Model Assessment

To evaluate the coefficients of explanation, a structural model analysis was used with PLS-SEM using bootstrapping techniques (5.000 subsamples) (^{Hair et al., 2022}). The blindfolding technique was used to evaluate the predictive relevance of the model (^{Geisser, 1975}; ^{Stone, 1974}). The path coefficients/hypotheses were analyzed by the Student t-test (Hair et al., 2022) (Table 4).

By analyzing the coefficients of explanation, a statistical significance was observed in the three predictive dimensions, highlighting that PE explained 72.9 % of PA and NE explained 62.9 % of NA. The same occurred with the predictive relevance PE - PA (Q² = 0.434) and NE - NA (Q² = 0.317.). Positive relationships were observed between PE - PA (p = 0.854; t = 6.024; p = 0.000) and NE - NA (p = 0.793; t = 3.246; p = 0.000) and negative relationship between PE - NE (p = -0.290; t = 5.889; p = 0.000) (Table 4).

Table 4 The relationship analysis through PLS-SEM and evaluation of model coefficients 

Figure 2 shows the descriptive measures of the indicators for each dimension. The highest positive emotion of the respondents was Gratitude (mean = 3.31; SD = 1.193), which is a reflection of having survived a pandemic, having gone through difficult times, and having expectations for better days, while the lowest was Kindness (2.37; 1.053). This reflects the lack of affection, little affection, and lack of care due to social confinement. The highest negative emotion was Stress (2.69; 1.247), which is a reflection of the moment of concern we are living due to physical and emotional overload. The lowest was Bitterness (1.63; 0.773), reflecting the extent to which people relied on each other during the pandemic to offer help and foster a sense of hospitality. As for the affections, we observed that the highest positive affect was Active (3.08; 1.255), which arises from the principle of people wanting to be effective, diligent, and resilient. The lowest positive affect was Excited (2.27; 1.029), that is, less excited and stimulated to perform tasks. The highest negative affect was Nervous (2.60; 1.181), which means feeling upset, distressed, and impatient with the situation we live. The lowest affect was Remorse (1.62; 0.717), less guilty and distressed.

Figure 2 Descriptive measure of emotion and affect indicators (mean) 

Figure 3 shows that both positive (60.04 %) and negative (47.41 %) emotions predominated in a moderate index in the individuals surveyed. In contrast, 66.29 % of the individuals surveyed exhibited a high level of negative affect, while 53.83 % demonstrated moderate positive affect.

Figure 3 Classification of dimensions 

Hypothesis Testing and Workplace Implications

H1: Positive emotions are positively correlated to positive affect. This hypothesis was confirmed, aligning with ^{Silton et al. (2020)}, who found that positive emotions at work are closely related to well-being and motivation. ^{Alexander et al. (2021)} further highlighted that experiencing positive emotions promotes happiness and engagement in professional settings.

H2: Negative emotions are positively correlated to negative affect. This hypothesis was also supported. Research indicates that negative emotions can contribute to stress and burnout, impacting employees' emotional health (^{Kang & Nan, 2022}). ^{Fadardi and Azadi (2017)} and ^{Almaraz et al. (2022)} discovered that workplace emotional health could be influenced by personal beliefs, including religious faith, which may mitigate some negative effects but not eliminate their impact on well-being. Thus, mindfulness-based training and emotional intelligence programs are recommended to improve emotional self-regulation and minimize negative emotions in professional environments (^{Kang & Nan, 2022}).

H3: Positive emotions are negatively correlated to negative emotions. This hypothesis was supported, albeit with a weak correlation. Studies suggest that positive emotions can counteract negative emotions by broadening individuals' cognitive and behavioral capabilities (^{Diener & Seligman, 2002}; ^{Gupta, 2019}; ^{Pressman et al., 2019}). In work settings, emotions such as joy and gratitude can enhance collaboration, creativity, and engagement, while reducing the effects of stress and anxiety (^{Gupta, 2019}; ^{Johnson et al., 2020}). ^{Fredrickson (2001)} posited that positive emotions widen cognitive flexibility and enhance adaptive behaviors, enabling employees to navigate challenges more effectively. ^{Fredrickson (2004)} showed that experiencing positive emotions in the workplace can reduce the aftereffects of stress and frustration, although the underlying mechanisms remain to be clarified.

Implications for Workplace Emotional Health

In work environments, positive and negative emotions influence decision-making, teamwork, and overall job satisfaction (^{Johnson et al., 2020}; ^{MacIntyre & Vincze, 2017}). While negative emotions can serve adaptive purposes, such as alerting employees to potential threats or conflicts, excessive negativity in the workplace can lead to disengagement and burnout. Consequently, fostering a culture of psychological safety, recognition, and emotional intelligence training can help organizations enhance employee well-being and productivity.' Positive emotions have the potential to neutralize negative emotional experiences by expanding employees' thought-action repertoire, reducing emotional exhaustion and fostering resilience in the workplace (^{Fredrickson, 2001}; Johnson et al., 2020). This highlights the significance of developing emotional health inventories tailored to professional contexts, enabling organizations to effectively assess, monitor, and improve employees' emotional well-being.