Readiness of Digital Transformation among Malaysian Digital Talents

,


Introduction
Throughout the period of fourth industrial revolution and Covid-19 epidemic, most of the corporate sectors are undergoing digital transformation (Dhesi, 2021) with the purpose to increase productivity, reduce energy and material consumptions as well as improve working conditions in the organisations (Machado, Wintroth, Almstrom, Oberg, Kurdve & AlMashalah, 2021;Gilch & Sieweke, 2020). According to the Malaysian Digital Economy Corporation (MDEC), the epidemic has increased the speed of change, with the number of digital job openings tripling year over year (The Star, 2021a). However, Malaysia still has a long way to go in terms of digitisation and digitalisation, with 54% of Malaysian businesses just begun their digital journey in 2020 (The Star, 2021b). The major difficulty encountered by Malaysian businesses that undergoing digital transformation is the shortage of human resources with digital capabilities who are completely prepared for digital transformation (digital expert) (Yapp, 2020). This is expected to worsen the employee readiness of digital transformation, while a digital talent process. Thus, readiness for digital transformation, alternatively also being referred to as "digital transformation readiness" or "digital readiness" or "digital change" (Nasution et al., 2018). It begins with the adoption and use of digital technology and progresses to an implied fundamental transformation of the organisation, or the purposeful pursuit of value creation. It refers to an individual's capacity to adapt to a digital environment and its associated technologies such as by developing new or changing existing business model and customer experiences to meet changing business and market needs (Dolganova & Deeva, 2019).

Digital Organisational Culture
Culture is crucial in settings the digital workplace transformation, especially during the period of digital transformation in which requires people to have an open mindset, be adaptable and ready to change (Alkhamery, Zainol & Al-Nashmi, 2021). According to Schein (1984), culture is transferred to employees through the establishment of certain ideals in their thoughts and the daily procedures in which they engage. Harshak et al. (2013) argue that organisations cannot transform their culture just by convincing individuals of the benefits of digitisation. According to Samal, Patra and Chatterjee (2019), establishing a digital environment and attitude is essential for shaping employees' readiness for digital transformation.

Digital Literacy
Historically, digital literacy was confined as the ability to read, write, and utilise textual resources in a variety of contexts (Centindamar, Abedin & Shirahada, 2021). Whereas, Gilster (1997) defines digital literacy as the ability to comprehend and use information presented in a variety of formats and from a variety of sources when presented via computers, has since evolved into an all-encompassing term with numerous applications in the ICT literature. While Eshet-Alkalai (2004) refers digital literacy as the technical and operational skills that required for optimal computer use; other researchers have expanded the definition to include the high cognitive ability that required to access, analyse, and create information by using digital resources and technological gadgets (Tapscott, 1998;Van Laar, Van Deursen, Van Dijk & De Haan, 2017). In overall, digital literacy can be defined as the ability, knowledge, and capability of workers to utilise digital technology in work-related tasks. As a result, it has the ability to significantly boost the use of digital technology in the organisation (Centindamar, Abedin & Shirahada, 2021). Furthermore, digital literacy is a competency that denotes an individual's familiarity with and ability to utilise digital technology in a range of scenarios (Messic 1984;Cetindamar, Abedin & Shirahada, 2021).

Attitudes Towards Digital Transformation
Attitudes towards behaviour is a component of the TBP. Attitudes reflect one's beliefs about the repercussions of one's actions and the perceived value of these consequences determines one's willingness to take action (Hardin-Fanning & Ricks, 2017). Attitude can be described as a person's tendency to feel, think, or behave favourably or adversely toward the object (Meske 2019). Hardin-Fanning and Ricks (2017), Eby et al. (2000), Kotter (1996), and Martin (1998) primarily focus on positive attitudes as a technique of fostering desired behavioural intention changes. Apart from that, positive attitudes assist individuals in developing physical, intellectual, social, and psychological resources, therefore increasing their resilience and reducing their resistance (Hardin-Fanning & Ricks, 2017). Hence, attitude is a significant determinant of success and failure in any endeavor of digital transformation (Meske, 2019).

Self-Efficacy
Self-efficacy and perceived behavioural control are classified as the same concepts (Bandura 1997). Bandura (1997) defines self-efficacy as the ability to exert influence on one's choice of activities and environmental circumstances, which involves effort and perseverance. When a person's self-efficacy is poor, they tend to postpone completing a task. On the contrary, individuals who believe they are capable of completing the work are more likely to do so. For this research, self-efficacy is referred to behavioural predictor of making a digitally changed or transforming choice when individuals think they are capable of making the shift to digital (Kahveci, 2021;Venkatesh, Morris, Davis & Davis, 2003;Yunus, Ang & Hashim, 2021). Basically, it refers to how individuals perceive their confidence in using digital technology and how they control the usage based on available skills, knowledge, infrastructure and resources (Vimalkumar, Singh & Gouda, 2021). Individuals with self-efficacy in digital transformation show characteristics such as self-confidence, accurate self-evaluation, willingness to take risks and a sense of accomplishment (Vimalkumar, Singh & Gouda, 2021).

Autocratic Leadership
Leadership is thought to be a significant source of influence in a group context inside an organisation (De Cremer, 2006). Autocratic leaders are often seen as restricting group members' influence and voice over collective decisionmaking processes and exhibiting a controlling and aggressive leadership style that shows little regard for followers' ijbm.ccsenet.org International Journal of Business and Management Vol. 18, No. 4; ideas (Caillier, 2020). Thus, autocratic leadership is defined in this research as the leader's dominance and control over the process of debating views and ideas that affecting the group's real choice especially in the innovation and digital transformation issues (Okecha, Joureih & Oluwatobi Okeniyi, 2020).

Self-Efficacy and Readiness of Digital Transformation
According to the SCCT and TBP theories, self-efficacy has an effect on the amount of effort expended to improve behaviour and the perseverance with which one persists in the face of obstacles and failures that may diminish motivation (Ajzen, 2002). When a digital talent has self-efficacy, he or she can confidently do a certain digital activity while using digital technology to accomplish the goal (Deja, Rak & Bell, 2021) because self-efficacy is a subjective assessment of an individual's confidence in the ease with which digital technology can be employed (Oh, Kho, Choi & Lee, 2022). Moreover, Nasution, Arnita and Azzahra (2021) also discover that those who have a high self-efficacy for technology and skills are more confident in engaging with any technology and thus leading them to increase their readiness of digital transformation. In other words, an individual's readiness to adapt to the digital transformation is strongly influenced by their confidence in their ability to carry out future actions (Madden, Ellen & Ajzen, 1992;Tommasetti et al., 2018;Deja, Rak & Bell, 2021). Hence, it can be hypothesised as: H1: Self-efficacy has positive relationship with readiness of digital transformation

Attitudes Towards Digital Transformation and Self-Efficacy
According to Kahveci (2021), the challenge encountered by many organisational employers is cultivating positive employee attitudes. If their attitudes are negative or contemptuous, employees are unlikely to participate confidently in any digital transformation process (Erdem, 2015). When employees embark negative attitudes towards digital transformation, they will experience a loss of self-esteem, which will demotivate them to learn, develop, and embrace changes (Garavan, McCarthy, Lai, Murphy, Sheehan & Carbery, 2020). Oppositely, Udo et al. (2010) and Olugbola (2017) state that employees who have positive attitudes towards digital transformation are more likely to be confident about their capabilities to produce effects than those who have low and negative attitudes towards engaging in the digital transformation. Thus, it can hypothesise as: H2: Attitudes towards digital transformation has positive relationship with self-efficacy

Attitudes Towards Digital Transformation, Self-Efficacy and Readiness of Digital Transformation
Olugbola (2017) states that employees who have positive attitudes are more likely to set up their mindset to be ready for new workplace environment than those who have low and negative attitudes to engage in digital transformation. Positive attitudes towards digital transformation enhances the people a broader range of potential thoughts and behaviour, as people with a positive outlook expect and receive positive outcomes more often. The positivity hereby enables the employees to quickly understand the benefits that they will attain from changing into digital transformation (Oh et al. 2022), thus, give rise to the confident of digital talents to adopt the readiness of digital transformation (Garavan et al., 2020). Once the organisational members have the positive attitudes towards digital transformation, it is easy for the employees to prepare for digital transformation, as them more confident to their knowledge, skills, behavioural control and emotions (Garavan et al., 2020). Thus, it can hypothesise as: H3: Self-efficacy mediates the relationship between attitudes towards digital transformation and readiness of digital transformation

Digital Literacy and Self-Efficacy
Prior research indicates that digital literacy improves employees' self-efficacy (Deja, Rak & Bell, 2021). According to Hamidi et al. (2018), digital skills and knowledge of employees are critical for the adoption of digital technologies in order to perform the digital transformation. Within the context of digital transformation, digital talents' confident behaviour is governed by their capacity to engage with digital technology (Singh & Hess, 2017). When digital talents are able to learn and work in an environment where communication and access to information are increasingly facilitated by digital technologies, they will understand how technology can benefit their work and assist them in performing the tasks (Wang et al., 2014;Trenerry et al., 2021). In other words, employees that having digital literacy on digital technologies will increase their self-confidence in using digital technologies and able to achieve their goals easily (Hamidi et al. 2018). Thus, it can hypothesise as: H4: Digital literacy has positive relationship with self-efficacy

Digital Literacy, Self-Efficacy and Readiness of Digital Transformation
According to Khalique and Singh (2019) as well as Cetindamar, Abedin and Shirahada (2021), individuals with a low confidence in digital knowledge, ability and capabilities to use digital technologies to perform their tasks will tend to have low readiness for the digital transformation. Besides having an understanding on the information technology and control procedures, digital talents must also be capable of confidently exchanging data with machines and integrated systems (Deja, Rak & Bell, 2021). Based on the above arguments, self-efficacy is essential and serves as a bridge between digital literacy and digital transformation readiness. Thus, it can hypothesise as: H5: Self-efficacy mediates the relationship between digital literacy and readiness of digital transformation.

Digital Organisational Culture and Attitudes Towards Digital Transformation
Trushkina, Abazov, Rynkevych and Bakhautdinova (2020) assert that the digital transformation involves not only an increase in demand for digital skills in the labour market, but also the necessity to develop and implement a set of strategies to convert digital organisational culture in the context of the rapid expansion of digital technologies and information in the organisation. With the establishment of a digital organisational culture, the organisation can provide sufficient digital technology infrastructure, encouragement and information exchange among employees in order to boost the employees' motivation, keen to take risk, refine their skills, acquire new ones and prepare for the digital transformation (Wu, Huang, Huang & Du, 2019;Alofan, Chen & Tan, 2020;Carmona, Gomes & Costa, 2020;Singh, 2021;Khin & Kee, 2021). Hence, it will strengthen their digital talent's views and attitudes towards the adaptation of digital transformation (Panichkina, Sinyavskaya & Shestova, 2018). Thus, it can hypothesise as: H6: Digital organisational culture has positive relationship with attitudes towards digital transformation

Digital Organisational Culture, Attitudes Towards Digital Transformation and Self-Efficacy
Through the establishment of a digital organisational culture in the organisation, the organisation can provide the infrastructure for information technology, information systems and a set of digital vision, mission and objectives that serve as the foundation for improving the employees' self-efficacy (Jang et al., 2018). Thus, the organisation is able to strengthen the digital talent's views and influence their attitudes towards digital transformation (Panichkina, Sinyavskaya & Shestova, 2018). In other words, employees will develop an interest in becoming more immersed in digital technologies and a curiosity about how digital transformation will affect their work performance (Khin & Kee, 2021) and encourage high self-confidence among themselves (Jang et al., 2018). Therefore, it is suggested that a digital organisational culture fosters favourable attitudes towards digital transformation among digital talents, which eventually increases their self-efficacy. Thus, it can hypothesise as: H7: Attitudes towards digital transformation mediate the relationship between digital organisational culture and self-efficacy

Digital Organisational Culture, Autocratic Leadership and Attitudes Towards Digital Transformation
Based on the argument of Akor (2014), the leadership style chosen by a supervisor has a significant impact on the employees. Caillier (2020) asserts that when managers engage in disempowering behaviours, they decrease their employees' sense of self-efficacy, psychological control and influence within their work environment as well as destroy the culture of the organisation. The extant literature indicates that when organisation strongly emphasises on autocratic leadership, a negative climate arises within the organisation that makes the digital organisational culture to reciprocate weakening the organisation (Dyczkowska & Dyczkowski 2018), which is translated into worsen employee attitudes and behaviour towards the digital transformation (Dyczkowska & Dyczkowski, 2018;Katou, Budhwar & Chand, 2020). Conversely, when organisation places less emphasis on autocratic leadership, a positive climate arises within the organisation that makes the digital organisational culture to reciprocate strengthening the organisation (Dyczkowska & Dyczkowski, 2018;Katou et al., 2020), which is translated into improved employee attitudes and behaviour towards the digital transformation (Katou et al., 2020). Thus, it can hypothesise as below. All the hypotheses are presented in Figure 1.

Research Design and Data Collection Method
By considering the research objectives and data collected, the quantitative research design is adopted in this research to evaluate the determinants of readiness of digital transformation (Eyisi, 2016). Furthermore, Singh Setia (2016) advocates that cross-sectional study should be adopted in this research in order to gather a large amount of data in a short period of time, and data will be acquired only once from the target population.
Besides, online self-administered survey is adopted in this study because this method allows data to be collected quantitatively (Malhotra, 2019). Primary data will be collected from the digital talents in Malaysia via online selfadministered survey to ensure the timing of assessments is up-to-date and align with the study follow-up period (Abbondanza & Souza, 2019). With the convenient of the online survey and respondents' network, the respondents can be accessed easily (Andrade, 2020).

Sampling Design
As to determine the readiness of digital transformation among digital talents, each organisation should begin by assessing the readiness of its own internal employees. As such, with the objective of promoting digital transformation readiness throughout Malaysian digital talents, the targeted respondents for this study are the employees from the various industries in Malaysia. For the eligibility to participate in this survey, the respondents must be Malaysian working adults with aged 18 years and above as well as currently expose to the digital technologies in their workplaces. This research uses the judgemental sampling approach to get a sample of employees from various industries in Malaysia. Bryman & Bell (2015) claim that a judgement sample aims to choose sample components that are typical of the population. The research employs discretion in this study to choose a certain number of individuals from various businesses who are most suited to provide the research with the much-needed insights (Cohen, Manion & Morrison 2018). Upon evaluating numerous sources of literature (Cochran 1948;Hair et al. 2018;Singh & Pal 2014;Nunnally 1978;Lei & Wu 2007;Hair, Bush, & Ortinau, 2009), a sample size of 450 respondents will be targeted in this research.

Measurement
The questionnaire design is divided into two parts. Section A focuses on demographic questions that consisting of 6 demographic questions, such as gender, age, education level, profession, types of industry and time spent on digital technology at working. A total of 6 constructs being measured in Section B, including dependent variable ijbm.ccsenet.org International Journal of Business and Management Vol. 18, No. 4; (readiness of digital transformation) and independent variables (digital organisational culture, digital literacy, selfefficacy, attitudes towards digital transformation) and moderator variable (autocratic leadership).
In term of measurement, digital organisational culture (DOC) is measured by 4 items that adopted from Zhen, Yousaf, Radulescu and Yasir (2021). Both digital literacy (DL) and self-efficacy (SE) are measured by 4 items respectively in which all of the questions are adopted from Deja, Rak and Bell (2021). Whereas, attitudes towards digital transformation (ATT) are measured by 3 items that adopted from Venkatesh et al. (2003) and Meske (2019). Autocratic leadership (AL) is measured by 8 items that adopted from Akor (2014). Lastly, readiness for digital transformation (RDT) is measured by 4 items that adopted from Nasution et al. (2018). All the items in Section B will be evaluated by 7-point Likert scale, ranging from (1) Strongly Disagree to (7) Strongly Agree. The detailed of the measurement can be referred to Appendix A.

Data Analysis Method
For the data analysis methods, the statistical programs SPSS version 28 and SmartPLS version 3.2.8 are adopted in this study to perform the statistical analysis. Preliminary analysis (non-response bias, common method variance, and multivariate normality), descriptive analysis (demographic and constructs), measurement model and structural model will be explicated in the data analysis and finding sections.

Preliminary Data Analysis
In this research, the non-response bias analysis revealed that there is no non-response bias because the six demographic variables of the early and late respondents are not significantly different from each other as the pvalue for all demographic variables are larger than 0.05 (Behar-Horenstein & Feng, 2017). The percentage variance retrieved from the Harman single-factor test is 39.48 percent, which is less than the threshold value of 50% (Podsakoff et al., 2003). Therefore, the common method bias issue does not exist in this research. For the multivariate normality, the result of Mardia's normality test indicated that the data is not normally distributed as the b-value of the multivariate kurtosis is 57.837553, which is higher than threshold value of 3 (Yuan, Bentler & Zhang, 2005).

Constructs and Correlational Analysis
As shown in Table 2  Note. **Correlation is significant at the 0.01 level (2-tailed).

PLS-SEM Assessment
PLS-SEM software is employed as the multivariate normality is not fulfilled in the research (Graber et al., 2002). Measurement model and structural model are developed with PLS-SEM (Hair et al., 2017). Figure 2 shows the conceptual structural model which being used in PLS-SEM.

Assessment of Measurement Model
Since the proposed framework is a reflective model, factor loadings, construct reliability, convergent validity, and discriminant validity will be examined through PLS-SEM (Hair et al., 2017). According to Table 3, all the factor loadings of the indicators are ranging from 0.761 to 0.930, which can be considered as satisfy as all the loadings are higher than 0.7 and majority of the loadings are higher than the threshold value of 0.708 (Hair et al., 2017). Besides, Table 3 shows that the inter-item consistency reliability values of Cronbach alpha within the range from 0.869 to 0.933, which larger than the suggested value of 0.7 (Nunnally, 1978). Furthermore, Dijkstra-Henseler's rho (ρA) also achieved a satisfactory reliability value which ranges from 0.870 to 0.945, which are exceeding the threshold value of 0.7 as recommended by Dijkstra and Henseler (2015). Moreover, the values of Composite Reliability (ρc) are ranging from 0.919 to 0.945, which were higher than the recommended value of 0.7 (Nunnally and Bernstein, 1994). Based on the result generated, the research achieved overall reliability with high internal consistency reliability.
For the Average Variance Extracted (AVE), all constructs are ranging from 0.682 to 0.863, which are exceeding the threshold value of 0.5 (Hair et al. 2017). Apart from that, the t-values of all factors are ranging from 32.328 to 136.193, which are above 1.96 (95% confidence level) (Ramayah et al., 2018). In shorts, all the factors are significantly loaded towards their respective latent constructs. In summary, the measuring model used in this study had convergent validity with the data.   Fornell and Larcker's (1981) and cross loading. According to Table 4, the results show the values of correlation among the latent variables are lower than the threshold value of 0.85 as suggested by Kline (2011), and there is no value of 1 being included in the upper and lower level of the confidence interval (Preacher & Hayer, 2008). Thus, it can conclude that the latent measurement constructs are clearly discriminant with each other. In conclusion, the measurement model achieved adequate reliability, convergent validity and discriminant validity.

Assessment of Structural Model
For the structural model, assessment of collinearity (VIF), significance of the path coefficient, determination of coefficient (R 2 ), effect size (f 2 ), predictive relevance (Q 2 ) based on blindfolding, and advanced predictive relevance conducted via PLSpredict would be analysed. Assessment of collinearity (VIF) test must be assessed before the evaluation of structural model as it can be used to ensure that the research does not contain any potential bias in the regression results. According to  According to the result generated from the assessment of R 2 , the R 2 value for the RDT is 0.270, illustrating that 27.0% of the total variance of RDT is explained by its exogenous variable (SE). Furthermore, 32.6% (R 2 = 0.326) of the total variance of SE is explained by its exogenous variables (ATT and DL). Apart from that, DOC explained approximately 41.1% (R 2 = 0.411) of the total variance of ATT, as according to the result generated.
Additionally, the assessment of f 2 is adopted in order to measure the effect size of exogeneous and endogenous constructs as stated by Ramayah et al. (2018). Cohen (1988) indicates that a threshold value of 0.35 is used to define a large effect size, 0.15 is used to define a medium effect size, and 0.02 is used to define a small effect size. According to the result generated from the assessment of f 2 , SE (f 2 =0.37) has a large effect size on RDT.
Furthermore, ATT (f 2 =0.08) has a small effect size on SE, whereas DL (f 2 =0.26) has a medium effect size on SE.
Finally, DOC (f 2 =0.15) has a medium effect size on ATT. With the results mentioned, the exogeneous constructs possesses significantly different levels of effect size on the endogenous constructs.
Furthermore, Stone-Geisser's Q 2 method was utilised in this research to determine the exogenous construct's predictive value for endogenous constructs after the exogenous construct's effect size on endogenous constructs was determined (Geisser 1974;Stone 1974;Shmueli et al. 2019). Q 2 value larger than zero suggests that the exogenous factors are highly predictive of the endogenous construct, as established by Fornell and Cha (1994) and Shmueli et al. (2019). According to the result generated from the assessment, the Q 2 values for ATT (Q 2 = 0.331), RDT (Q 2 = 0.201), and SE (Q 2 = 0.231) were all larger than zero, indicating the model's adequate predictive significance (Fornell & Cha, 1994;Hair et al. 2017;Shmueli et al. 2019).
For more accuracy of the predictive relevance, PLSpredict assessment is employed as an advancement of the Q 2 assessment. According to Table 6, the Q 2 predict values for all indicators are greater than zero for the PLS-SEM. If the Q² value is positive, the prediction error of the PLS-SEM results is smaller than the prediction error of simply using the mean values. In that case, the PLS-SEM models contains predictive relevance (Shmueli et al. 2019).
Furthermore, since the data collected for this study are not normally distributed, MAE will be employed to assess if PLS-SEM less than LM [PLS-SEM<LM]. According to Shmueli et al. (2019), MAE error would be recognised for the predictive relevance effect if the research is not normally distributed. As comparison about the MAE values, the findings conclude that the PLS-SEM analysis produces a medium prediction power as the result has majority of the indicators (7 out of 10 indicators; PLS-SEM<LM). In conclusion, majority of the indicators fulfil the requirement [Q 2 predict>0; MAE error values are negative (PLS-SEM<LM); thus, moderate predictive powers are existed for the RDT model (Shmueli et al., 2019).

Direct Effect Test
In this research, a total of four hypotheses being developed to examine the direct relationship between the variables. According to the results showed in Table 7, SE (β= 0.407, t =7.431, p < 0.05) is found to have a significant positive direct effect on RDT, indicating that H1 supported. Besides, ATT (β = 0.243, t = 5.036, p < 0.05), and DL (β = 0.444, t = 9.238, p < 0.05) are also found to have significant positive impacts on SE, indicating that H2 and H4 are statistically supported. Furthermore, H6 is also supported as DOC (β = 0.388, t = 6.376, p < 0.05) posits a significant positive effect on ATT. Besides, all of the direct hypotheses do not have zero strapped in between the upper level and lower level of the 95% confidence interval. Finally, the findings of the PLS-SEM bootstrapping approach show that all four direct hypotheses are significantly supported.

Mediation Effect Test
There are three mediation hypotheses being developed to investigate the indirect relationships between the variables. As exhibited in Table 7, H3 (β =0.127, t = 4.434, p < 0.05) and H5 (β =0.232, t = 6.353, p < 0.05) are supported to have significant indirect effects on RDT. Moreover, the mediation effects on SE are validated as H7 is supported with β =0.235, t = 6.407, and p < 0.05. Additionally, there is no zero straddle in between the upper level and lower level of the confident interval for all the mediation hypotheses. Therefore, it can conclude that all the three mediation hypotheses are statistically supported.

Moderation Effect Test
Moderation (H8) is statistically significant, as indicated in Table 7, with findings of (β= -0.154, t = 3.207, p < 0.05). Furthermore, there is no zero straddle in between the Confident Interval's lower and upper levels. At the meantime, R 2 and f 2 are measured between the direct relationship between DOC and ATT. With the comparison, R 2 of ATT has changed about 1.9% (additional variance) with the addition of the interaction term (DOC*AL), illustrating that the effect size of the moderating effect is minor (0.0325) as referred to the guidelines suggested by Cohen (1988). According to Figure 4, the upper line, which represent a high level of the moderator construct AL, has a flatter slope while the lower line, which represent a lower level of the moderator construct AL, has a steeper slope (Hair et al. 2017, p. 269). This make sense since the interaction effect is negative. As the rule of thumb and an approximation, the slope of the high level of the moderator constructs AL is the simple effect (0.388) plus the interaction effect (-0.154), while the slope of the low level of the moderator constructs AL is the simple effect (0.388) minus the interaction effect (-0.154) (Hair et al. 2017, p. 269). Hence, the simple slope plot supports the previous discussion of the negative interaction term as higher AL in the organisation, entail a weaker relationship between DOC and ATT, while lower levels of AL in the organisation lead to stronger relationship between DOC and ATT (Hair et al. 2017, p. 269). As referring to Table 7, the analysis yields a p-value of 0.001 for the path DOC*AL  ATT.
Overall, these results provide clear support that AL exerts a significant and negative effect on the relationship between DOC and ATT (Hair et al. 2017, p. 269). The higher the autocratic leadership, the weaker the relationship between digital organisational culture and attitudes towards digital transformation.

IPMA
Apart from the significance of hypotheses, it is vital and meaningful to adopt Importance and Performance Matrix Analysis (IPMA) to extend the findings of the basic PLS-SEM results with the latent variable scores (Hair et al., 2016). IPMA is used to identify the total effects (importance) and the average values of the latent variable scores (performance) of the specific endogenous construct (Ramayah et al., 2018). The results are contrasted using the IPMA procedure via PLS-SEM in order to identify the most influential area for digital talent readiness in Malaysia (Hair et al., 2017).
According to Figure 5 and Table 8, the outcomes indicate that the IPMA of Readiness of Digital Transformation (RDT) reveals that the self-efficacy (SE) does have a high-performance and high importance index score. According to Sethna (1982), SE is the important variables thus, should be concentrated as in Figure 5. Precisely, this aspect would be related to self-confidence and self-esteem of the individual in the workplace (Deja, Rak & Bell 2021). Without self-efficacy, an employee will find it difficult to make tough decisions, get people to communicate with them candidly, and be open to feedback. Hence, an employee will always doubt his decisions and find himself becoming defensive (Khalique & Singh 2019).

Goodness of Fit
Corresponding to the obtained findings, the SRMR values are 0.05 and 0.07, suggesting that the model has a high degree of fit for both the saturated and estimated models, as according to the Henseler et al. (2015) advised threshold value of SRMR is less than 0.08. Additionally, the NFI values are 0.869 and 0.863 for both the saturated and estimated models respectively, despite the fact that NFI values should be greater than 0.90 to be regarded acceptable (Bryrne 2016). The NFI scores in the study suggest a lack of model fit for both the saturated and estimated models. For the RMS theta measurement, the proposed model in this study does not have good fit as the value is 0.172, which does not meet the threshold value of 0.12 as proposed by Henseler et al. (2016), who said that a value closer to zero indicated a better fit. Therefore, this research is appraised with the reliability and validity of the measurement, the significance of path coefficient, the prediction ability, and the explanation ability of the model to ensure the amount of random error is acceptable in the research as there is not perfectly good-fit in PLS-SEM (Henseler et al., 2016).

Conclusion
In conclusion, this research filled the research gap by investigating the determinants for the readiness of digital transformation from the employee's perspective. This research provides theoretical contribution by concluding that there are direct and indirect relationships among digital organisational culture, digital literacy, attitudes towards digital transformation, self-efficacy and readiness of digital transformation from the employee perspective. In addition, the findings also discover that the higher the autocratic leadership, the weaker the relationship between digital organisational culture and attitudes towards digital transformation. As part of the efforts in implementing digital transformation in the organisation, the organisational top management is encouraged to enhance the digital literacy among their employees by adopting favourable digital organisational culture and transformational leadership management style. There are some limitations in this research. Most of the determinants for the readiness of digital transformation are derived from the perspective of employees. Further research can be carried out to what extent the social psychologies of the employees affects their readiness of digital transformation. Due to the diversity of industries being targeted in this research, it is recommended to focus on the small and mediumsized enterprises (SMEs) that contributing to the nation growth in the country. ATT3 I know about a lot of different digital technologies.

AL1
My manager is often over bearing in his regular inspection of my work.
Akor (2014) AL2 My manager does not accommodate any kind of domestic excuse interfering with my duties.

AL3
My manager believe that I will work best in a situation where I am given clear and direct instruction on my job.

AL4
My manager wears an officious look most of the time.

AL5
My manager rules with an iron hand.

AL6
My manager does not readily accept new ideas.

AL7
My manager takes decisions arbitrarily.

AL8
My manager does not explain his actions.

Copyrights
Copyright for this article is retained by the author(s), with first publication rights granted to the journal.
This is an open-access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/4.0/).