Advancements in digital technology have enabled organizations to efficiently handle various work tasks and processes through modern technological solutions. Bringing about organizational change is associated with many challenges. Thus, this study examines how employees engage in creative deviance to face the stress of digital tension. This study also examines the intermediate role of the paradox mindset. Paradox theory was used to examine all of the relationships in the study. The study model was analyzed using structural equation modeling (SEM) in Amos. Results of an online survey with 241 employees in the communication and information technology sector in Saudi Arabia revealed that employees engage in creative deviance as a response to digitalization represented in digital leadership. Additionally, paradox mindset functions as an intermediary variable in the relationship between digital leadership and creative deviance. These findings provide valuable insights into tension management within the organizational context. They also broaden the understanding of the paradox theory through its application to employees’ behaviors in times of tension.
Since the emergence of artificial intelligence (AI), the internet of things (IoT), cloud computing, and other technologies, firms in the business world have embraced many technologies that have led to digital transformation. Digital transformation is the prevalence of digital technologies that foster organizational change (Hanelt et al., 2021). Incorporating these digital technologies in firms aims to increase their efficacy in regard to internal operations and decision-making capabilities (Sharabati et al., 2024). Most previous research highlighted the negative effects of organizational change caused by digital transformation, such as strain Scholze and Hecker (2024); Zhang et al. (2024), lower job autonomy and job satisfaction (Abdulkareem et al., 2024), and lower mental health, social trust, and happiness (Trittin-Ulbrich et al., 2021). Thus, in order to overcome these negative effects, firms in the current digital transformation era need to have digital leaders to facilitate the acceptance of this organizational change. These digital leaders should be conscious and committed to guiding and directing their employees about how important these technologies are in their day-to-day work tasks and how important they are in achieving organizational objectives. In this context, digital leadership can be defined as leaders who facilitate digital technology adoption by enhancing employees’ awareness and providing guidance to develop their performance. There are a number of studies that illustrate how digital leadership has a positive effect on employees’ attitudes and behaviors, such as employee empowerment and work engagement (Li et al., 2024; C. Yang et al., 2024), affective commitment (Li et al., 2024), job satisfaction and employee performance (Hidayat et al., 2023), and job crafting and employee creativity (Zhu et al., 2022). These studies indicate that the tension toward digitalization through digital leadership can bring not only the acceptance of new technologies but also other desirable consequences. However, the presence of digital leadership alone is not enough to facilitate digital transformation, stimulating organizational scholars to seek new approaches to navigate such a transition.
Employees navigating digital transformation often face the challenge of reconciling traditional work methods with technological advancements. This includes adapting from routine-based decision-making to technology-driven processes and transitioning manual tasks to digital platforms while managing the complexities of new software and systems. Therefore, in order to handle these situations, employees need to manage competing demands, which requires them to have a paradox mindset. According to Klein et al. (2024), a paradox mindset is considered a positive attitude when employees encounter organizational change, specifically digital transformation. According to Miron-Spektor et al. (2018), a paradox mindset refers to an individual's capacity to recognize and embrace tensions, drawing motivation from them and leveraging the opportunities embedded within. In the paradoxical literature, there is a common misconception about the origin of the paradox mindset. Many studies argue that individuals with a paradox mindset are born, not made, claiming that paradox mindset is an inherent personal trait. However, it should be noted that the paradox mindset is not a personal characteristic that remains constant, but individuals can acquire it through compelling sense (Dweck et al., 1995). Individuals who embrace paradoxes tend to view contradictions as new opportunities when they successfully balance two opposites. Thus, individuals who embrace paradoxes are more innovative, flexible, and energized than others (Miron-Spektor et al., 2018). Several studies highlight the positive effects of having a paradox mindset, such as creativity (Ngo et al., 2024), thriving at work and innovative work behavior (Liu et al., 2020), and work engagement (Yin, 2023). Nevertheless, the paradox mindset is one of the most important elements that can facilitate digitization and bring many other positive outcomes.
One of the major disputes in workplace environments is employees’ deviant behaviors. It is widely recognized that workplace deviant behavior has a substantial adverse impact on organizations and their employees (Robinson & Bennett, 1995). Recent evolutions in workplace environments have influenced different aspects, and employees’ deviant behavior has also evolved to practice creative deviance. Creative deviance refers to instances where employees persist with new ideas even after being told by managers to discontinue them. (Mainemelis, 2010). C. Yang et al. (2024) emphasize that creative deviance originates from employees rather than management, and they gave three characteristics of creative deviance: (a) it includes self-initiated employee action; (b) it presupposes the disregard of managerial requests; and (c) its purpose is positive. According to Lin et al. (2016), creative deviance typically arises during the idea elaboration phase of the creative process, leading the employee to proceed to the stage of implementing the idea without managerial support, resulting in either creative outcome or failure. This failure will benefit the employee in terms of the trial-and-error learning process, contributing to their future creative efforts (Mainemelis, 2010). On the one hand, the benefit of creative deviance for employees is, to some extent, a beneficial behavior regardless of whether the final result is a success or failure. On the other hand, the benefit of creative deviance for organizations depends on the managerial reaction to it. Lin et al. (2016) emphasized that managers react to creative deviance by either forgiving, rewarding, punishing, ignoring, or manipulating, influencing the employee's choice to engage in creative deviance or creative performance. An important element of creative deviance is that it is a paradoxical behavior because, in striving for positive outcomes, employees may cross procedural boundaries (Liu & Zhou, 2021; Mainemelis, 2010). Previous studies have illustrated many antecedents of creative deviance, such as ethical leadership (Liu et al., 2021), empowering leadership (Liu et al., 2021), and servant leadership (Kumar et al., 2024). Although those and other studies have broadened our knowledge about the antecedents of creative deviance, none of them discuss the impact of how creative deviance occurs in digitalization due to practicing the paradox mindset.
To address the previously mentioned aspects of digital transformation, this study aims to investigate the impact of digital leadership on creative deviance and the role of paradox mindset as an intermediate variable through the lens of paradox theory. According to this theory, tensions are expected in systems that are susceptible to complexity and continuous changes, and organizations can gain benefits from these tensions; specifically, when two opposite forces occur at the same time, they challenge individuals to think differently and to be flexible, creative, and analytical (Smith & Lewis, 2011). Therefore, experiencing digital transformation is a paradoxical situation where employees need to balance competing demands to ensure effective transformation.
This study offers multiple contributions to the existing body of literature on workplace ethics and digital tensions. First, this study has identified digital leadership as an antecedent that can be used to enhance the paradox mindset. Additionally, this study has identified digital leadership as a novel antecedent of creative deviance. Previous studies (e.g., Kumar et al., 2024; Li et al., 2025; Liu et al., 2021) illustrated different leadership styles as antecedents of creative deviance; however, this study has specified digital leadership as an antecedent of creative deviance. Second, another novel antecedent of creative deviance, this study illustrated that a paradox mindset can increase creative deviance. Previous studies (e.g., Liu & Xu, 2019; van Assen & Caniëls, 2022) illustrated that a paradox mindset can increase innovative work behaviors. Third, this study identified digital transformation as a boundary condition, filling the gap in determining a boundary condition where digital leadership impacts organizational context (Li et al., 2024). Fourth, this study is the first study to base all the relationships on the paradox theory to explain the underlying mechanisms.
Literature review and hypothesesDigital leadership and creative devianceDigital leadership is a style of leadership that is specifically important in the context of digital transformation tension. The main role of this style of leadership is to enhance employees' awareness of the importance of using technologies that make organizational operations more effective. Thus, handling daily tasks that require employees to incorporate new technologies, with the presence of a digital leader, should encourage employees to pursue creative deviance. This is an expected behavior because dealing with contradictions will lead employees to, in some cases, try new ways or ideas that their leaders have rejected. Several studies found that the persistence of digital leadership increases employees’ creative deviance, such as Lu (2023), Sun et al. (2024), and Shie et al. (2025). Nonetheless, it is well established that leadership styles have been considered as an essential antecedent of creative deviance (Lu, 2023). For example, ethical leadership (Liu et al., 2021), empowering leadership (Li et al., 2025), and servant leadership (Kumar et al., 2024). Moreover, bottom-up leadership styles have shown a positive association with creative deviance, such as humble leadership (Zhang, 2023), platform leadership (Shie et al., 2025), and coaching leadership (Li et al., 2022). According to Shie et al. (2025), a resource-empowering leadership style can push employees beyond traditional constraints and motivate them to overcome challenges they face through higher confidence and capabilities throughout the innovation process. Digital leadership is a leadership style that provides resources to employees, such as awareness through knowledge and experience. Thus, digital leaders encourage engagement in creative deviance through provided resources and requirements. As mentioned previously, employees are required to balance competing demands, leading them to approach rejected ways and ideas to achieve positive outcomes.
Through the lens of paradox theory, contradictions between work tasks resulting from digital transformation are seen as opportunities, not difficulties, for development and creativity. Thus, the presence of a digital leader promotes an atmosphere of creative deviance that is considered constructive or positive behaviors rather than destructive or negative behavior. Consequently, the following hypothesis to explore the association between digital leadership and creative deviance is proposed:
H1 Digital leadership is positively related to creative deviance.
While avoiding contradictions in situations, ideas, or even discussions is expected behavior for employees under circumstances of tensions such as digital transformation, employees with a paradox mindset tend to perceive those contradictions as opportunities to implement or accept both opposite positions simultaneously. Furthermore, tensions are considered challenging because if employees perceive them as dilemmas, they will have unproductive behaviors, leading them to seek new approaches to consistently handle them (Miron-Spektor et al., 2018). According to Liu et al. (2020), the enhancement of an individual's propensity to face contradictions instead of averting them is associated with their paradox mindset. Moreover, digital transformation tension with the presence of digital leaders leads to the occurrence of multiple competing demands. For example, employees are required to acquire knowledge and skills regarding new technologies, and, at the same time, they are required to achieve specific productivity levels. Therefore, through the role of digital leaders, they can enhance employees’ adaptability by encouraging them to perceive those contradictions as opportunities that can be achieved simultaneously. Several studies highlighted that paradox mindset is an acquired skill (Batool et al., 2023; Boemelburg et al., 2020; Miron-Spektor et al., 2018). A recent study emphasized that managing paradoxes is a set of abilities that are acquired through ongoing experience and development (Nayebpour & Sehhat, 2023). Digital leaders are characterized by openness to change, new possibilities, innovation, and a future-focused approach (Fisk, 2002). Furthermore, the context of digital transformation itself is characterized by a shifting mindset “‘digital’ is not only interpreted as a technological phenomenon but also as a shift in mindset to create business value in new digital-related ways” (Brunner et al., 2023, p. 4); therefore, digital leaders have the ability to facilitate the creation of an environment that encourages paradoxical thinking. This is emphasized by Klein et al. (2024), who found that the paradox mindset is positively associated with acceptance of digital transformation. In the organizational literature, several studies have found that in general, the positive styles of leadership, such as servant Leadership (Bouzari & Safavi, 2021), ethical leadership (Xu & Yu, 2019), inclusive leadership (Shabeer et al., 2023), and transformational leadership (Żywiołek et al., 2022) are positively associated with employees’ adaptability, while the negative styles of leadership such as abusive supervision (Rasheed et al., 2021) has negative association with employees’ adaptability. Therefore, digital leadership, a positive leadership style, can stimulate employees’ paradox mindset by changing their way of thinking when facing contradictions, causing them to perceive these as opportunities rather than dilemmas.
Based on paradox theory, digital transformation is a source of major tension within organizations and provides competing demands. Thus, a successful digital leader can facilitate embracing these contradictions by enhancing the employees’ attitudes toward them. Therefore, the following hypothesis to explore the association between digital leadership and paradox mindset is proposed:
H2 Digital leadership is positively related to paradox mindset.
Developing innovative solutions in challenging situations is an expected result of employees with a paradox mindset, specifically in organizations that are experiencing digital transformation. Those solutions are not always acceptable to managers. Thus, employees engage in creative deviance, working on their ideas in order to make them more acceptable. According to Miron-Spektor et al. (2018), employees with a paradox mindset are proactive in balancing competing demands, which enables them to handle both routine and creative work. Likewise, Liu and Zhang (2022) emphasized that the paradox mindset of employees enhances their role by expanding and developing their innovative performance. Studies, such as van Assen and Caniëls (2022) and Liu and Xu (2019), have demonstrated that employees who embrace paradoxes tend to demonstrate more innovative behavior at work. Therefore, they are expected to engage in creative deviance as a part of innovative work behaviors. According to Yin (2023)), the paradox mindset is a personal resource that provides personal improvement and has a stress-buffering impact. Yin (2023)) also found that employees who embrace paradoxes tend to demonstrate more challenges, tend to let go of outdated attitudes and behaviors, and accept new ones. This kind of mindset motivates individuals to pursue their ideas with the aim of positive outcomes. Several studies found that paradox mindset has a positive impact on work engagement (Yin, 2023), confidence in decision making (Schädeli, 2025), and organizational citizenship behavior (Pan, 2021). Practicing creative deviance is a positive behavior due to the desired positive outcomes, at least from the point of view of the employee who engages in creative deviance. Therefore, employees who possess a paradox mindset and who perceive challenges as opportunities can stimulate their engagement in creative deviance to make their ideas more acceptable to their managers. Nonetheless, there are no studies that explore this relationship specifically in organizations that are function within digital transformation.
Based on paradox theory, digital transformation is a source of major tension within organizations and provides competing demands. Thus, employees who possess a paradox mindset are expected to practice creative deviance in order to develop new ideas and approaches to be implemented in their organizations. Therefore, the following hypothesis to explore the association between digital mindset and creative deviance is proposed:
H3 Paradox mindset is positively related to creative deviance.
Furthermore, the presence of a leader who facilitates digital transformation by enhancing employees’ awareness about the different aspects of new technologies will increase employees’ levels of paradox mindset because employees are required to manage the achievement of routine and new tasks. Likewise, the higher levels of employees’ paradox mindset will lead them to practice creative deviance to work on unacceptable managerial ideas to reach positive outcomes. Therefore, the following hypothesis to explore the intermediate role of the paradox mindset in the association between digital leadership and creative deviance is proposed:
H4 Paradox mindset mediates the relationship between digital leadership and creative deviance.
This study was conducted in Saudi Arabia. The study community was the communication and information technology sector. This sector was chosen for this study because it is one of the major sectors undergoing digital transformation in order to achieve national goals, specifically Saudi Vision 2030 (Kingdom of Saudi Arabia, 2016). The data was collected using non-probability sampling, specifically convenience and snowballing sampling, and an online questionnaire was used to collect the data. In order to ensure participants' anonymity, on the survey’s cover page, participants were reminded that their answers would be anonymous and would only be used for this study, following the ethical consideration proposed by Imam Mohammad Ibn Saud Islamic University (IMSIU).
The number of participants was established in line with the rule of thumb of Bentler and Chou (1987). This rule suggests that the ratio of the number of participants to items in the questionnaire should be at a minimum of 5 to 1 and at a maximum of 10 to 1. As a result, the aim was to reach a sample size between a minimum of 120 and a maximum of 240. Therefore, the sample size for this study was 241 participants. Table 1 shows the demographic characteristics of the participants.
Demographic characteristics (N = 241).
The measures used in this study are all established measures (Table A1) that have proven their validity and reliability. In addition, the measures were translated into Arabic using forward and backward translation methods Brislin (1970, 1986). In addition, the survey was pretested using a pilot study, Q-sorting, and think-aloud techniques. The data collection took place over three months, from December 2024 to February 2025, through an online survey, and the analysis was made using SPSS and Amos.
Digital leadershipDigital leadership was measured using a 6-item scale developed by Ulutaş and Arslan (2018). This scale demonstrated good validity and reliability, and the overall reported Cronbach’s alpha is 0.91 (Shin et al., 2023). The measure asks the respondents to indicate their agreement or disagreement with each statement. An example item is “Supervisor/leader raises awareness of the technologies that can be used to improve organizational processes.” The measure has a 5-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree.
Paradox mindsetParadox mindset was measured using a 9-item scale developed by Miron-Spektor et al. (2018). This scale demonstrated good validity and reliability, and the overall reported Cronbach’s alpha is 0.80 (Miron-Spektor et al., 2018). The measure asks the respondents to indicate their agreement or disagreement with each statement. An example item is “When I consider conflicting perspectives, I gain a better understanding of an issue.” The measure has a 5-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree.
Creative devianceCreative deviance was measured using a 9-item scale developed by Lin and Chen (2012). This scale demonstrated good validity and reliability, and the overall reported Cronbach’s alpha is 0.81 (Lin & Chen, 2012). The measure asks the respondents to indicate their agreement or disagreement with each statement. An example item is “I continued to improve some of the new ideas, although they did not receive my supervisor's approval.” The measure has a 5-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree.
Control variablesIt has been noted that the demographics of participants are crucial determinants that may impact paradox mindset and creative deviance. For example, previous studies found that paradox mindset is positively correlated with tenure (Liu & Zhang, 2022), creative deviance is positively correlated with education (Xu et al., 2022), and creative deviance is positively correlated with education and tenure (Lin et al., 2016). Therefore, gender, age, education, and tenure were controlled in this study.
Analytic strategyAll the study variables were latent variables, unobserved variables, except for the control variables. First, the assumptions of confirmatory factorial analysis (CFA) and structural equation modeling (SEM) were examined. These assumptions include missing data, outliers, sample size, multivariate normality, and multicollinearity. Even though the use of structural equation modeling has many disadvantages, such as the requirement of a larger sample size and the need for a strong theoretical foundation (Byrne, 2013; Kline, 2016), it has several advantages over other analysis methods, such as multiple regression. One advantage is that the use of SEM helps in decreasing measurement error because it allows measuring latent variables, such as the variables in this study (Kline, 2016), which also helps in enhancing the validity and reliability of the estimated relationships (MacCallum & Austin, 2000). Likewise, the use of SEM allows for testing all the study’s relationships simultaneously, which provides the study with a comprehensive view of the test theoretical model instead of other methods that test each relationship in a single approach (Bollen, 1989). Another unique characteristic of SEM is that it shows the extent to which the data fit the expected theoretical model by providing the fit indices (Hu & Bentler, 1999). Moreover, the SEM approach can handle complex theoretical models, such as incorporating the test of mediating and moderating variables (Byrne, 2013).
Another important aspect of the analysis is checking reliability and validity. Reliability reflects the extent to which the measurement is precise, solid, and able to be replicated, which can be estimated using Cronbach’s alpha (Dimitrov, 2012). Cronbach’s alpha is widely used due to its ease of computation and interpretation (Tavakol & Dennick, 2011). Moreover, it measures the variance of all items and their covariances (Cronbach, 1951); thus, it only requires a one-step test, unlike other methods that require two-step tests (Streiner, 2003). In addition, this method measures the average inter-item correlations and considers the possibility of item pairing within a scale (Cronbach, 1951). Furthermore, while the Kuder-Richardson method can handle only binary responses, Cronbach’s alpha can handle different types of scales, such as ordinal, continuous, Likert-type items (Cortina, 1993), therefore, justifying the spread of Cronbach’s alpha in various fields such as the social sciences, psychology, and education.
Additionally, testing for the construct validity, specifically convergent and discriminant validity. The criteria for evaluating convergent validity include three conditions that need to be met: (a) all CFA items should be higher than 0.5 (Hair et al., 2006); (b) the composite reliability (CR) should be higher than 0.7 (Hair et al., 2010); and (c) the average variance extracted of constructs should be higher than 0.5 (Hair et al., 2010). Those criteria have several advantages over traditional methods of evaluating convergent validity. First, the use of composite reliability is very accurate because it takes into consideration the actual factor loading of the items (Fornell & Larcker, 1981). Likewise, while other methods, such as inter-item correlation, might be misleading due to error, AVE is a rigorous method because a high AVE of a construct, 0.5 or higher, means that more than half of the variance on the construct is explained by its items (Bagozzi & Yi, 1988). The use of these methods collectively can provide the study with robust results because using other methods, such as exploratory factor analysis and evaluating factor loading individually, depends on subjective interpretations because there are no standardized cutoffs (Fornell & Larcker, 1981).
Furthermore, for evaluating the achievement of discriminant validity, Fornell and Larcker’s (1981) criteria was used, where the squared root for the AVE of each construct is higher than its correlation with the other constructs. The use of this criteria is beneficial specifically in the context of CFA and SEM because the output of these methods provides both variable-level variance, the AVE, and the inter-variable correlations. Furthermore, it incorporates measurement error in the AVE, which makes it more accurate (Bagozzi & Yi, 1988). According to Hair et al. (2010), this tool is a direct indication of discriminant validity because it provides numbers with a straightforward rule. Thus, it reduces ambiguity, and it is easy to apply in research. Additionally, Hair et al. (2010) emphasized that Fornell and Larcker’s method can be easily applied, especially after the specification of AVE and CR in the previous step of evaluating convergent validity. This provides consistency and comprehensiveness in evaluating the measurement model of the study. Likewise, while HTMT is a recent method of inspecting discriminant validity, there is some subjectivity regarding the threshold (either 0.85 or 0.90), which makes the specification of the threshold different across various studies, while the Fornell-Larcker criterion has one clear threshold where the square root of AVE is higher than the inter-variable correlations. Furthermore, Hair et al. (2017) recommend that researchers use the Fornell-Larcker criterion because it is universally applied, which makes the study easy to compare with previous studies.
Lastly, it is important to test for the common method bias (CMB) after achieving the construct validity and reliability. This study used Harman’s single-factor test to detect CMB. This technique does not require the incorporation of other variables, such as the social desirability scale, and it leads to the same conclusion, which is the detection of CMB. According to Kock (2020), Harman's test is simple due to its easy implementation compared with other techniques. Likewise, this method does not require specifying the source of the bias (Podsakoff et al., 2003); this is a straightforward tool. Additionally, this tool controls for the measurement error in the measurement model, unlike other simple tools for detecting CMB (Podsakoff et al., 2003). Lastly, it is a widely used tool due to its minimal resource requirements (Tehseen et al., 2017), which makes it applicable to compare the CMB levels in different studies.
The following step is to conduct a pooled CFA in order to confirm the expected structure of the used factors (Dimitrov, 2012). There are several criteria for fit indices to evaluate the study’s model fit; when the model fits the expected structure, the next step is to conduct SEM to test the relationships in the hypothesized model. Additionally, to examine the mediation in the model, the bootstrapping method was utilized. This method has several benefits over the traditional Sobel test and Baron and Kenny three-step procedure. It offers better statistical power and greater control of type 1 error (Ng & Lin, 2016). Likewise, it does not require a large sample size (Biesanz et al., 2010). The bootstrapping method relies on the fundamental principle that the mediation impact is confirmed only if the confidence interval does not include zero.
ResultsDescriptive statistics and correlationsTable 2 highlights the mean, standard deviations, and correlations among the study’s variables. As expected, digital leadership is positively correlated with paradox mindset (r = 0.33, p < .001), paradox mindset is positively correlated with creative deviance (r = 0.34, p < .001), and digital leadership is positively correlated with creative deviance (r = 0.25, p < .001). These correlations provide initial support for the study hypothesis.
Mean, standard deviations, and correlations among the study variables.
| Variable | Mean | SD | 1 | 2 | 3 | 4 | 5 | 6 | 7 | CR | AVE |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1. DL | 3.62 | 0.85 | (0.89) | 0.89 | 0.58 | ||||||
| 2. PM | 3.45 | 0.61 | 0.33⁎⁎⁎ | (0.83) | 0.83 | 0.36 | |||||
| 3. CD | 3.41 | 0.76 | 0.25⁎⁎⁎ | 0.34⁎⁎⁎ | (0.91) | 0.91 | 0.53 | ||||
| 4. Gender | 0.52 | 0.50 | 0.00 | −0.16* | −0.18⁎⁎ | 1 | – | – | |||
| 5. Age | 2.71 | 0.77 | 0.03 | 0.10 | 0.18⁎⁎ | −0.44⁎⁎⁎ | 1 | – | – | ||
| 6. MS | 0.58 | 0.49 | 0.02 | −0.04 | −0.11 | 0.30⁎⁎⁎ | −0.53⁎⁎⁎ | 1 | – | – | |
| 7. Education | 3.25 | 0.56 | 0.07 | 0.05 | −0.01 | −0.15* | 0.34⁎⁎⁎ | −0.20⁎⁎ | 1 | – | – |
| 8. Tenure | 2.86 | 1.15 | 0.07 | 0.17⁎⁎ | 0.27⁎⁎⁎ | −0.55⁎⁎⁎ | 0.82⁎⁎⁎ | −0.56⁎⁎⁎ | 0.32⁎⁎⁎ | – | – |
Note. DL = digital leadership, PM = paradox mindset, CD = creative deviance. Between parentheses are Cronbach's alpha, CR = composite reliability, and AVE = average variance extracted.
Reliability can be estimated using Cronbach’s alpha. According to George and Mallery (2003), Cronbach’s alpha values equal to or higher than 0.90 are excellent, and equal to or higher than 0.80 are good. As can be seen in Table 2, Cronbach’s alphas for digital leadership, paradox mindset, and creative deviance are 0.89, 0.83, and 0.91, respectively, indicating good reliability levels.
There are three criteria for evaluating convergent validity: (a) all the CFA items should be higher than 0.40 (Hair et al., 2006); (b) composite reliability (CR) should be higher than 0.70 (Hair et al., 2010); and (c) the average variance extracted (AVE) of constructs should be higher than 0.50 (Hair et al., 2010). Fig. 1 displays that all the study items were 0.40 or higher for the item's loadings. For the CR and AVE, as shown in Table 2, the CR for all constructs is above 0.80, but one construct, paradox mindset, has an AVE lower than 0.50. According to Fornell and Larcker (1981), if the AVE is lower than 0.50 but the CR is higher than 0.60, that indicates adequate convergent validity. Therefore, convergent validity is achieved.
For the discriminant validity, Table 3 displays Fornell and Larcker’s (1981) criteria, and it demonstrates a discriminant validity for the study constructs as the squared root for the AVE of each construct is higher than its correlation with the other constructs.
Common method biasIn this study, Harman’s single-factor test is used to detect CMB. According to Podsakoff et al. (2003), factor analysis is used to determine the value of the variation explained by all the items by incorporating them into one factor. Put differently, the study has CMB if this single factor accounts for >50 % of the variance. Based on this test, the items of this study explain 30.07 % of the variance, which indicates that there is no CMB in this study.
The measurement modelThe measurement model of the study is evaluated through confirmatory factorial analysis in order to confirm the expected structure of the used factors (Dimitrov, 2012). As can be seen in Table 4, all of the used fit indices have met the recommended criteria.
The fit of the measurement.
| Fit Index | Recommended Criteria | Measurement model |
|---|---|---|
| X2 /df | ≤ 5 | 1.71 |
| CFI | ≥ 0.90 | 0.93 |
| TLI | ≥ 0.90 | 0.92 |
| RMSEA | ≤ 0.08 | 0.05 (0.04–0.06) |
| SRMR | ≤ 0.08 | 0.06 |
Notes: X2 /df: the ratio of Chi-square to degrees of freedom, CFI: comparative fit index TLI: Tucker-Lewis index, RMSEA: root mean squared error of approximation, SRMR: Standardized Root Mean Squared Residual.
As can be seen in Table 5, all of the used fit indices have met the recommended criteria. Therefore, we can proceed to hypothesis testing.
The fit of the measurement.
| Fit Index | Recommended Criteria | Structural model |
|---|---|---|
| X2 /df | ≤ 5 | 1.61 |
| CFI | ≥ 0.90 | 0.93 |
| TLI | ≥ 0.90 | 0.92 |
| RMSEA | ≤ 0.08 | 0.05 (0.04–0.05) |
| SRMR | ≤ 0.08 | 0.05 |
Notes: X2 /df: the ratio of Chi-square to degrees of freedom, CFI: comparative fit index TLI: Tucker-Lewis index, RMSEA: root mean squared error of approximation, SRMR: Standardized Root Mean Squared Residual.
Fig. 2 highlights the structural model of the study. It includes the dependent variable, creative deviance; the independent variable, digital leadership; and the mediator, paradox mindset. Additionally, the analysis includes the control variables: gender, age, education, and tenure. Notably, paths are all standardized regression coefficients.
Table 6 highlights the results of the hypothesized model. Hypothesis 1 proposes that digital leadership is positively related to paradox mindset. Fig. 2 shows that this hypothesis is significant and positive (β = 0.35; p ≤ 0.001). Likewise, Hypothesis 2 proposes that paradox mindset is positively related to creative deviance. This hypothesis is significant and positive (β = 0.29; p ≤ 0.01). In addition, Hypothesis 3 proposes that digital leadership is positively related to creative deviance. This hypothesis is significant and positive (β = 0.16; p ≤ 0.05). Lastly, Hypothesis 4 proposes that paradox mindset mediates the relationship between digital leadership and creative deviance (Fig. 2). Thus, bootstrapping of a 2000 sample size revealed that the CI is [0.02, 0.17], showing that the CI does not contain zero, concluding that paradox mindset mediates the positive relationship between digital leadership and creative deviance, which also suggests partial mediation.
Testing the Direct and Indirect Relationships.
| Hypothesis | Direct Effect | Estimate | S.E. | Result | ||
|---|---|---|---|---|---|---|
| H1 | DL → PM | 0.35⁎⁎⁎ | 0.04 | Supported | ||
| H2 | PM → CD | 0.29⁎⁎ | 0.13 | Supported | ||
| H3 | DL → CD | 0.16* | 0.06 | Supported | ||
| Hypothesis | Indirect Effect | Estimate | S.E. | Percentile method (95%) | Result | |
| Lower bound | Upper bound | |||||
| H4 | DL → PM → CD | 0.10** | 0.03 | 0.02 | 0.17 | Supported – Partial Mediation |
Note. S.E. = standard error. DL = digital leadership, PM = paradox mindset, CD = creative deviance.
The control variables were all insignificant, except for tenure; it was positively related to creative deviance (β = 0.27; p ≤ 0.05). As shown in Fig. 3, the trend shows that the engagement in creative deviance is associated with the employees’ tenure; this implies that employees with higher tenures tend to engage in creative deviance.
Discussion
Creative deviance results in organizations being ambiguous, and research is still trying to uncover the consequences of practicing creative deviance, specifically under the condition of digital transformation. Therefore, this study facilitates the determination of antecedents of creative deviance. This study examined creative deviance as a consequence of having a paradox mindset in organizations that are digitally transforming. Based on the paradox theory, this study revealed that digital leaders, who enhance their employees’ awareness regarding new technologies, can increase their paradox mindset level because this kind of leadership communicates to their employees with messages that require them to handle both routine and new tasks. On the one hand, digital leadership can increase the employees’ levels of paradox mindset, which means considering challenges as opportunities. On the other hand, the absence of digital leadership can lower the employees’ paradox mindset, which can lead them to perceive digital transformation as a challenge that is hard to grasp. Accordingly, engaging in creative deviance is associated with having a higher level of paradox mindset. This means that employees with a positive view of the situation can come up with new ideas and behaviors they want to implement in their daily work tasks, but what prevents them is their manager's non-supportive position. Thus, they engage in creative deviance to spend more effort on their ideas, aiming to make the non-supported ideas supported by their managers. In other words, without a paradox mindset, as a facilitating method, employees will not easily engage in creative deviance because they lack the ability to be optimistic and maintain a positive perspective regarding the tension they are experiencing. It is noteworthy in this study that digital transformation as an organizational tension serves as a boundary condition for a paradox mindset to increase engagement in creative deviance. It should be noted that paradox mindset partially mediates the relationship between digital leadership and creative deviance, which implies that digital leadership has a direct impact on creative deviance in the absence of paradox mindset. Moreover, it implies that paradox mindset is one of many other factors that may have an intermediate impact on this relationship. Likewise, it highlights that digital leadership has a direct and indirect impact on creative deviance. The direct impact is through enhancing employees’ innovativeness, and the indirect impact is through stimulating their paradox mindset.
Furthermore, the findings of this study propose that tenure is associated with creative deviance, meaning that longer job tenure can enhance the chances of employees exhibiting creative deviance. This may occur for several reasons. For instance, the longer an employee spends in their job, the more they know about management's reaction to their engagement in creative deviance. As a result, there will either be no punishment or the employee will be rewarded for their behavior. Likewise, the employee may have high confidence in the outcome of their engagement in creative deviance, and they have already established credibility and trust within the organization. Furthermore, they may have a deeper understanding of the organizational structure, where routine is implemented, and where innovative ideas can be implemented. Based on previous findings, theoretical and practical implications are highlighted.
Theoretical implicationsThis study has several theoretical implications by answering different questions, which will enrich the literature regarding the findings of the relationships in this study. First, the positive association that exists between digital leadership and employees’ engagement in creative deviance exposes an individual-level consequence of digital leadership. Likewise, this study exposes that digital leadership is an antecedent of creative deviance. The text explaining this relationship also provides insight into new explanations built on paradox theory, confirming that digital leadership is a paradigm shift in leadership styles (Lu, 2023) and expanding the impact of digital leadership on creative deviance through the mediation of the paradox mindset.
In addition, this study identifies digital leadership as a groundbreaking antecedent of the paradox mindset. Recently, paradoxes in their different forms have emerged as a key area of interest for organizational scholars, such as paradoxical leadership (Xu et al., 2025), paradoxical tensions (Guo et al., 2025), and paradoxical mindset (Yoo & Roh, 2024). Even though there is a large interest in the paradoxical phenomena, limited studies have highlighted the antecedents of the paradox mindset (N. Yang et al., 2024). Drawing on the paradox theory (Smith & Lewis, 2011), this study emphasized that digital leaders who encourage managing competing demands also enhance the employees' paradox mindset. Therefore, this study contributes to the growing body of work on paradox mindset, emphasizing the role of leaders’ aspirations in managing it.
Furthermore, digital transformation was identified as a boundary condition in this study. In other words, the main motivation for employees’ engagement in creative deviance is this digital tension, which creates challenges and pressures. Without this boundary condition, there will not be new ideas for handling new situations. Limited studies have embraced boundary conditions regarding the relationship between paradox mindset and creative deviance, and this gap is encouraged to be uncovered due to the expected cultural differences in various studies (Liu et al., 2020). Consequently, the effect of a boundary condition in a culture might not have the same effect in another culture.
All the study relationships were built based on the paradox theory. Therefore, this is the first study that has embraced those relationships through the lens of paradox theory, which expands this theory in terms of how it relates to those constructs in times of tension. This study emphasized that the more employees have higher levels of paradox mindset in digital tensions, the higher their propensity to exhibit creative deviance, which is a positive tool for facing challenging situations. They perceive those situations as opportunities to work on rejected ideas and find new ways to facilitate the handling of different work requirements.
Practical implicationsThe results of the study suggest different practical implications. First, previous studies emphasized the importance of appointing transformative leadership (Schiuma et al., 2024) and servant leadership (Hamyeme et al., 2024) in times of tension in organizations. However, this study also suggests a new type of leadership: the digital leader. When there is internal tension in the organization, specifically digital tensions, organizations should appoint a leader who has the appropriate capabilities in terms of supporting employees to adapt the change and balance between current situation and the requirements to reach the targeted situation. The presence of a digital leader has the potential to enhance employees’ awareness, encourage them to anticipate different risks, and identify strategies to face those risks. Consequently, an important aspect of organizational strategic planning is to anticipate risks and identify strategies to face them, which will minimize expenses and promote rapid recovery from the consequences (Ahmad, 2024). This study sheds light on an essential aspect of the whole organization, which is to appoint a leader with experience in digital leadership in order to manage such tension with the appropriate instruments.
Second, the presence of digital leadership was found to be a successful driver of digital transformation, but the significant positive relationship between digital leadership and paradox mindset was not strong enough to diminish the impact of other factors. Therefore, employees differ in their adaptability levels regarding digital transformation. While it is true that this kind of organizational change requires high awareness, technical skills, and increased employees’ competencies, leading them to mental health workload (Körner et al., 2019; Sony & Mekoth, 2022). It is important to note that other factors may also affect employees’ adaptability, such as their personalities. For example, traits like low openness to experience might make change more stressful due to fear associated with failure or judgment. Likewise, some employees may have mindset rigidity, which means they refuse new opportunities for development. Another reason is the level of complexity of the needed skills, especially when there is not enough training. Moreover, the absence of incentives and motivation may lead the employees to resist spending efforts to shift their mindset or change their personality traits. Therefore, organizations should provide ongoing support, training, and mentoring programs. They also need to understand that employees have different personalities, which require them to provide different approaches to support, and to normalize failure as part of the growth process. Furthermore, organizations must provide sufficient motivation and incentives.
Third, another important aspect of managing tensions is managing stress associated with those tensions; managers should let employees know that the transformation from the current situation to the targeted situation is possible by following the managerial directions. Managers need to encourage their employees to handle those situations by enhancing their paradox mindset. Practicing the paradox mindset not only helps to handle current situations but also works as a buffering tool for stress in times of tension (Yin, 2023). Thus, organizations may offer training courses for their employees to facilitate their practice of the paradox mindset, reward employees who work on contradictory tasks, and praise their efforts in front of all employees. Furthermore, managers should act as role models for their employees by practically managing paradoxes, such as the employees' work demands and supporting their health (Glaser et al., 2019).
Fourth, even though some creative deviant behaviors are considered a waste of resources, if this behavior results in a positive outcome, it will bring various benefits to the organization. Thus, managers shouldn’t focus on preventing their employees from working on their rejected ideas. Specifically, in times of tension, employees suggest numerous ideas on how to handle new situations, and the appearance of a leader who is able to manage this tension is important. Thus, organizations should employ a leader with the required capabilities for such a situation. Likewise, managers should act as role models for their employees by practically showing them how a rejected idea can work if it has minor or major changes.
Fifth, this study illustrated that employees with extended tenure tend to engage more frequently in creative deviance. Thus, organizations must manage the hiring criteria and manage this behavior wisely because practicing creative deviance is not always about wasting resources; it still has ambiguous aspects in regard to the results of engaging in this behavior. Accordingly, managers shouldn’t strictly prevent their employees from working on their rejected ideas; however, they should be clearer about the reasons behind their rejections and whether it is possible to work on these rejected ideas in order to improve them and re-suggest them. Ultimately, an organization's success and a source of its competitive advantage may come from the application of creative ideas. These creative ideas are usually not spur of the moment, but they need to be studied, consulted, and refined to bring the desired benefit to the organization.
Conclusion, limitations, and directions for future researchIn current organizations, managing tensions has become harder than ever. How leaders manage those tensions, specifically digital tension and how they enhance employees' creative deviance and paradox mindset, is crucial to facilitating the required change. Drawing on paradox theory, this study investigates the association between digital leadership and creative deviance with the paradox mindset as a mediating mechanism. This study aimed to shed light on important aspects that will inform future research and expand our knowledge of these relationships.
Like any research, this study has certain limitations that offer opportunities for future investigation. First, even though this study fills the gap of determining a boundary condition where digital leadership has an impact on organizational context (Li et al., 2024), this is also a limitation that restricts the generalizability of the study’s findings to the organizations in other sectors that are not going through digital transformation, and non-probability sampling was used to collect the data. Thus, future research may conduct comparative studies within other sectors and countries, specifically to uncover the same relationships between organizations going through digital transformation and those that are not.
Second, this research highlighted a specific type of leadership and its impact on the paradox mindset, a very important variable beneficial in today’s organizational contexts. Therefore, there is a notable literature gap regarding paradox mindset organizational antecedents. As a result, future research is encouraged to find which variables can increase or decrease the employees' engagement in the paradox mindset. Likewise, while digital leaders have the ability to facilitate the creation of an environment that encourages paradoxical thinking, it is not inevitable that all employees will develop paradox mindset because several factors may impact this relationship; thus, future research is encouraged to investigate those factors, such as individual differences, organizational culture, organizational support, and learning opportunities.
Third, the outcome of this study is creative deviance, which is not yet considered a positive or negative outcome. Future research needs to consider what variables may facilitate this variable to make it more likely to receive positive outcomes. Likewise, future research needs to highlight what variables may facilitate this variable to avoid a negative outcome. Another crucial future advancement regarding the creative deviance construct is to develop two different constructs: constructive creative deviance and destructive creative deviance. These two constructs will open new paths in the creative deviance literature that are highly important for both research and real-world workplace implications.
Finally, although this study has no CMB, all the data were collected from one source: employees. Thus, future research is encouraged to collect data from multiple sources, including different levels of managers. Moreover, even though this study employed an advanced analytic method, making it a replicable data-driven study, future research is encouraged to conduct more innovative methods, such as longitudinal studies, comparative studies, and mixed methods studies.
FundingThis work was supported and funded by the Deanship of Scientific Research at Imam Mohammad Ibn Saud Islamic University (IMSIU) (grant number IMSIU-DDRSP2504).
Ethical approvalThe study was approved by Imam Mohammad Ibn Saud Islamic University (IMSIU) committee for the ethics of research.
Informed consentInformed consent was obtained from all individual participants included in the study.
CRediT authorship contribution statementBasmah Saad Alzamil: Writing – review & editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization.
The author declares no conflicts of interest.
The questionnaire.
| Variable | Reference |
|---|---|
| Digital leadership | (Ulutaş and Arslan (2018) |
| DL1 - Supervisor/leader raises the awareness of the employees of the institution about the risks of information technologies. | |
| DL2 - Supervisor/leader raises awareness of the technologies that can be used to improve organizational processes. | |
| DL3 - Supervisors/leaders determine the ethical behaviors required for informatics practices together with all stakeholders. | |
| DL4 - The supervisor plays an informative role in reducing resistance to innovations brought by information technologies. | |
| DL5 - Leaders share his/her own experiences about technological possibilities that will increase the contribution of their colleagues to the learning of organizational structure. | |
| DL6 - In order to increase participation in the corporate vision, a digital leader guides the employees of the institution about the technological tools that can be used. | |
| Paradox mindset | (Miron-Spektor et al. (2018). |
| PM1 - When I consider conflicting perspectives, I gain a better understanding of an issue. | |
| PM2 - I am comfortable dealing with conflicting demands at the same time. | |
| PM3 - Accepting contradictions is essential for my success. | |
| PM4 - Tension between ideas energizes me. | |
| PM5 - I enjoy it when I manage to pursue contradictory goals. | |
| PM6 - I often experience myself as simultaneously embracing conflicting demands. | |
| PM7 - I am comfortable working on tasks that contradict each other. | |
| PM8 - I feel uplifted when I realize that two opposites can be true. | |
| PM9 - I feel energized when I manage to address contradictory issues. | |
| Creative Deviance | (Lin & Chen, 2012) |
| When my immediate supervisor rejected some of my new ideas:CD1 - I continued to improve some of the new ideas, although they did not receive my supervisor's approval. | |
| CD2 - In my work time, I often thought about how to make the rejected ideas better. | |
| CD3 - Although my supervisor asked me to stop developing some new ideas, I still worked on these ideas. | |
| CD4 - Besides working on ideas that were approved by my supervisor, I also exerted effort in improving the rejected ideas by collecting information and trying again. | |
| CD5 - I spent some of my work time developing the ideas rejected by my supervisor. | |
| CD6 - Up to this point, I still have not given up on some of the rejected ideas. | |
| CD7 - I have improved some rejected ideas in my working hours. | |
| CD8 - Although some ideas were stopped by the supervisor, I worked on the improved versions of these ideas. | |
| CD9 - Using some of my work time or resources, I kept on working on the rejected ideas. |