Identification and Analysis of Factors Influencing Turnover Intention of Pakistan IT Professionals: An Empirical Study

Information Technology is one of the fast-growing industries. However, it is facing a high employee turnover. Lack of management support, low level of compensation plans, poor relationship with supervisors, limited development and training programs, and work stress causes high employee turnover. Therefore, this study is conducted to identify and empirically evaluate the factors of turnover intention in Pakistan’s software and IT industries. For this purpose, a systematic literature review is conducted to identify turnover intention factors from the literature, and a survey is conducted to evaluate the identified factors empirically. In the survey, a total of two hundred and fifty professionals from Pakistan’s IT and software industry participated either through online Google Form or printed survey questionnaires. The collected data were analyzed using regression analysis performed in SPSS and Warp-PLS. According to our findings, IT professionals prefer to work long tenure in the organization if they are provided with better opportunities in terms of compensation plans, career growth, and management support. The findings of this study help the organization’s top management to retain skilled employees.


I. INTRODUCTION
Employee turnover is defined as "the number of employees is reduced by leaving the organization" [1]. One of the most reported factors in recent studies that impact turnover is "Turnover Intention". Turnover Intention is defined as "the deliberate and conscious willingness to leave the organization" [2], [3]. The significant loss faces by the organization due to turnover intention is the decline in performance and low level of productivity. These losses decrease the intention of employees to stay longer in the same organization. Many recent studies focus on employee turnover intention in different organizations and sectors, such as the hotel industry [4], hospital management [5], banking sectors [6] and airline industry [7], due to the shortage of talented employees, high competition, sustainable competitive advantages. One of the highest turnover intentions has to be seen in Information Technology (IT) and software industry.
More than 30% of significant growth has witnessed $150 billion in the IT and software industry in the last few years [8]. IT and software professionals are tremendously mobile and interconnected with the world. Due to internet connection, they can change or switch their jobs efficiently and work from one industry to another [9]. According to the Pakistan Software Engineering Board (PSEB) report, the annual revenue of the IT industry is projected from $1.23 billion in a Fiscal Year (FY) 2019-20 to $1.64 billion in FY 2020-21 [10]. Due to these remarkable growths, Pakistan and India's IT services have rapidly improved in internetworking and communication technologies and become the world's most prominent market leader [11]. The recently reported growth of increased revenue in the nation's industry is the US $74 billion FY in 2010 to the US $167 billion FY in 2018 and estimated to hit the US $350 billion by 2025 [12]. Due to the tremendous growth of the IT industry, IT professionals have a crucial role in their success. IT industry's success merely depends upon retaining valuable skilled and talented employees [13]. However, it is difficult to retain skilled employees for longer tenure and achieve the organization's benefits [14]. Numerous studies have addressed this issue by investigating the retaining strategies to fill this gap [15]- [17]. Many other factors impact the failure of skilled employees and increase turnover intention, such as lack of job security and satisfaction, high workload and stress, unclear career path, inefficient compensation plans and poor relationship with the team member and management [15], [18]. These affecting factors are considered in various studies; however, less attention has been given to organizational and personal demographic factors that highly influence the turnover intention of IT professionals. Furthermore, several studies investigate the high turnover intention in the IT sector due to the low level of productivity, the substantial direct or indirect cost in terms of recruitment and selection process, delay in project completion and loss of skilled employees, etc., [16], [19].
This study aims to fill these gaps by investigating the impact of the turnover intention of IT professionals in the IT and software industry. Turnover intention is one of the most explored and related topics because it is the major facing issue in growing industries. Furthermore, the conceptual framework is proposed to gain insight into IT professional challenges and retaining strategies. A Systematic Literature review (SLR) is conducted to identify the affecting factors related to IT professionals in the IT industry. Finally, an empirical investigation has been performed to evaluate the conceptual framework of the Pakistan software and IT industry. The main research objectives of this study are to identify the affecting factors of turnover intention from existing literature and to evaluate the impact empirically.
The remainder of the current research study is organized as follows. Section 2 discussed the literature review cited in recent related work. It provides recent literature on employee turnover intention and influencing factors that impact IT professional turnover along with the proposed hypotheses. In section 3, the research methodology that has been used to resolve the research quires of the current study. An SLR was conducted to investigate the factors influencing the turnover intention of IT professionals. A conceptual framework was proposed based on SLR containing extracted factors from literature. The proposed framework was empirically evaluated from Pakistan's small and medium-sized IT and software industries. Section 4 describes the detailed empirical evaluation and results of SLR. A detailed discussion has been provided based on the results regarding each research question in Section 5. Finally, the current study's conclusion, limitations, and future directions are presented in Section 6.

II. RELATED WORK
In this section, the recent research studies on employee turnover intention along with influencing factors such as recruitment and selection process, teamwork and management support, performance and career management, salary and compensation, employee commitment, job security, recognition, organization demographics, and personal demographics, are presented in detailed.

A. EMPLOYEE TURNOVER INTENTION
In recent studies, employee turnover intention has been investigated widely in different organizations, such as hospital management, airline industry, banking industry, hotel industry, and IT industry [4]- [6]. In the IT and software industry, the high turnover intention of employees has been observed due to different challenging issues such as lack of management support, inefficient compensation plans, and lack of job satisfaction [15], [17]. In the IT industry, talented IT professionals strive to keep looking for new opportunities in high brand name organizations that provide better salary and compensation plans. However, retaining skilled IT professionals is challenging because rival companies merely search for and attain these employees [20]. Many IT industries build new policies and strategies to control the turnover intention that are unavoidably for them because their growth and performance are directly linked to the retention of hightech key employees [19].
The authors in [3] proposed a conceptual framework to investigate the relationship between onboarding success and turnover intention of IT professionals by mediating the role of job satisfaction and workplace relationship quality. However, the authors in [3] have not considered other important factors that influence IT professional turnover intention, such as job security, employee commitment, and organizational commitment. Moreover, the employee retention strategies that significantly influenced the IT sector were missing. In another recent study, the authors in [21] proposed a conceptual model that elaborates the antecedents of employee intention to leave South African IT industries. The authors have only considered the limiting factors and specific areas of employees in their study. However, their study did not investigate the other affecting factors such as recruitment & selection, performance & career management, and salary & compensation considered by [22]. The authors in [23] proposed a conceptual model to investigate the impact of professional commitment on turnover intention by mediating the role of job satisfaction. The IT employees are more committed to their job if the organization provides challenging tasks and learning opportunities. The effect of job satisfaction on employee commitment increases the intention of IT professionals to stay long in their current organization. As a result, the IT professionals' turnover intention is decreased. The authors have considered only these two essential factors that significantly affect turnover intention. However, The recruitment process explores the right skilled employee, whereas the selection process hires shortlist recruiters based on their talent and skills [1]. The primary concern of HRM is to retain their talented employee for a long tenure, have better revenue growth, and increase their productivity [25]. After hiring skilled IT employees, training is provided to meet the organizational task's requirement and polish a new skill. If the recruitment and selection process is well organized, the right skilled IT professional will stay long. Some organizations tried different strategies to retain them. However, IT employees are not satisfied with their learning skills or career path [26]. As a result, they search for new opportunities and employment to achieve their career goals. A well-managed recruitment and selection process positively influences employee commitment to the firm and employees' performance.
So, we hypothesize that IT professionals' turnover intention is based on the recruitment and selection process.
Hypothesis 1: The high degree of recruitment and selection process, the lower the turnover of IT professionals.

C. TEAMWORK AND MANAGEMENT SUPPORT
Teamwork is a collaborative task in which multiple individuals work together on a specific job to achieve common goals. In the IT industry, most professionals work long-term if they accomplish tangible benefits such as a high salary, efficient compensation plan, higher job role, and high job satisfaction [22]. Management support plays an essential role in the retention intention of IT professionals. In some recent studies, authors argue that relationships between employees and employers play a vital role. In some cases, IT professionals leave the organization due to poor relationships with supervisors, co-workers, and management [27]. When the satisfaction level of IT professionals increases with their supervisors or managers, they feel comfortable sharing their issues and problems. Such a supportive environment encourages professionals to keep doing their job and not leave the organization [17], [28]. After evaluating the recent literature, we hypothesize as follows: Hypothesis 2: Teamwork and management support negatively affect the turnover intention of IT professionals.

D. PERFORMANCE AND CAREER MANAGEMENT
The retention of skilled employees broadly impacts on organizational and workplace environment. The working environment where career growth, training, and learning opportunities for IT employees are high helps them increase their performance and productivity [2], [24]. The authors in [15] investigate that the job satisfaction of IT staff has a positive influence on the organizational environment. The IT employees strive to find new career opportunities in their working organization or some high brand name companies. Furthermore, they were also looking for a friendly work environment that increases their stay for a long tenure. Most IT organizations provide different career opportunities within their organizational departments that help to increase IT professionals' growth and job engagement. IT organizations develop several strategies and policies, such as equal opportunities, unbiased appraisal, efficient bonuses plan, and full autonomy, that help their employees to engage with them and increase their intention to stay longer [21], [29]. After reviewing that, we hypothesize as follows: Hypothesis 3: The IT professional's performance and career management negatively influence turnover intention.

E. SALARY AND COMPENSATION
Compensation plays a critical role in IT professionals' turnover intention [30]. A compensation plan refers to the promotion, rewards system, social benefits, basic allowances, and retirement savings plan. IT professionals are constantly trying to obtain better compensation plans within or outside VOLUME 4, 2022 their organization. Organizations mostly give high compensation plans to skilled employees based on their performance, which must be enough to retain them [1]. In recent studies, the authors investigate that the primary cause of high employee turnover is a lack of an efficient compensation system or salary growth [14]. Reasonable compensation and salary growth plan attract talented and skilled employees to remain in the same organization. Most experienced and potential IT professionals always look at compensation plans while joining and selecting a new organization [22]. Therefore, we hypothesize that fair salary and compensation enrich the IT professionals and decrease turnover intention.
Hypothesis 4: There is a positive relationship between salary and compensation and a negative impact on IT professionals' turnover intention.

F. EMPLOYEE COMMITMENT
Employee commitment refers to the relationship between employees' involvement and experience with their working organization. The IT employees' commitment plays a critical role in the industry's success. In recent studies, the authors [23] argue that highly committed employees stay longer due to the perceived cost of leaving the firm and emotional attachment to their organization. According to the human capital theory, the high skilled employees have less intention to leave the current organization or search for a job elsewhere if the organization has provided efficient development and training skills [31]. In the past few studies, most researchers have investigated the impact of employee commitment on turnover intention [2], [23], [25].
When an employee is committed to their career or profession, they start searching for better opportunities and compensation plans in other organizations; as a result, they stay short tenure in the current organization [32]. Furthermore, the authors [24] investigate that a high level of organizational commitment increases employee commitment; as a result, employees' turnover intention will reduce. In a recent study, the authors [32] emphasized that when employees' goals are fulfilled in the current working organization, they are highly committed to their profession. So, it would be difficult for them to leave the organization. Therefore, we hypothesize that employee commitment has an impact on turnover intention.
Hypothesis 5: Employee commitment negatively affects the IT professionals' turnover intention.

G. JOB SECURITY
Job security is defined as an employee's perception that their current job, or essential jobs feature, is secure. It is also described as the assurance of a job that employee will continue their employment or same profession in the current organization [15]. In a recent study, the authors [33] investigate that the employees who feel secure about their job or profession, their productivity and satisfaction level increase; as a result, they have less intention to seek new jobs or opportunities. However, if employees perceive threats to their job, they merely search for new employment and leave the organization. In the most recent management literature [17], [34], the researchers explore the impact of high job security on employee turnover intention. The authors [17] examine that job insecurity influences employees to find alternate job offers or new opportunities in other organizations. However, high job security may be decreased if the current organization provides high employment stability or efficient development programs [34]. Therefore, we hypothesize that the high significance of job security influences the IT professionals' turnover intention. Hypothesis 6: The high level of job security harms the turnover intention of IT professionals.

H. RECOGNITION
Recognition refers to caring for and appreciating the employees' contributions. Employees need to increase their motivation level; as a result, their association will increase the benefit of organizational goals [35]. In recent studies, most researchers discovered that employees need better compensation plans and want to be appreciated by their supervisors or managers [30], [35], [36]. Recognition is the most effective method for rewarding key employees. The author [36] suggests that employees should be appreciated for their achievements or achieved objectives. As a result, they have confidence in their job and work roles.
Furthermore, highly valued employees have less intention to leave or find new opportunities. A recent study explored that most professionals leave the organization because they feel unsatisfied and unappreciated by management due to a lack of proper recognition and reward system [36]. In another study, the authors [30] stated that the organization or firm must ensure employees' job security, provide good compensation and full autonomy and improve their skills. As a result, they committed to their profession and stayed long tenure with the organization. So, we stated the following hypothesis that the recognized IT professionals harm turnover intention: Hypothesis 7: The high degree of recognition negatively influences the turnover intention of IT professionals.

I. ORGANIZATION DEMOGRAPHICS
Organization demographics refers to the number of employees in the IT department within the organization. In the past few studies, several authors have investigated that high brand and larger organizational employees have a low level of turnover intention as compared to small or medium enterprises [15], [21], [26]. Furthermore, the IT employees are concerned that their organization provides new training skills, high salaries, job security, and recognition for their performance [26]. The employees working in small or medium-sized companies have not appreciated their performance by the organization. These employees have a high turnover intention, and therefore they start looking for new employment in large enterprises to achieve their career goals [15]. The author [21] stated that high portfolio employees have high demand in their working organization due to top management roles, high talents, and skillsets. Rival brand organizations are always searching for these talented employees to increase their productivity and annual revenue cost [22]. Therefore, we hypothesize that the higher the organizational demographics, the lower the IT professional turnover intention.
Hypothesis 8: Organizational demographic factors negatively affect the turnover intention of IT professionals: (a) Organizational Size, (b) Department Size.

J. PERSONAL DEMOGRAPHICS
Personal demographics refers to employee age, gender, education level, and work experience. Employee age is an essential factor that plays a vital role in turnover intention [37]. In recent studies, authors investigated that young employees significantly impact turnover intention compared to employees above 45 years because young employees voluntarily change their job due to fast learning and training skills [37]. However, employees above 45 have more retention intention to remain in the same organization [38]. The employee gender gap has another factor that influences the high employee turnover. According to the past few studies [26], [39], female employees in the IT sector leave the industry 2.5 times more than male employees due to work-family conflict, lack of proper role and personal preference, etc. Women employees also have a high turnover intention when the rival organization offers better opportunities or higher salaries. According to [26], those female IT staff who has more than 10 years of work experience have a great tendency to leave their job as compared to male IT staff. According to [40], the employees whose education is low level (secondary school, diploma holders); have lower turnover intention compared to employees with a high level of education (bachelor's or master's degree holder). Therefore, HRM must maintain and develop a flexible compensation plan and working environment for male and female employees [39]. So, we hypothesize that personal demographic factors play an essential role in employee turnover intention.
Hypothesis 9: Personal demographic factors negatively influence turnover intention of IT professionals: (a) Age, (b) Gender, (c) Experience, and (d) Education level.
In recent studies, several authors investigate that job satisfaction mediates turnover intention and personal demographic factors [15], [21]. They also suggest that women employees have high job satisfaction when the organization appreciates their performance and provides efficient compensation plans. Furthermore, it has been seen that job satisfaction has a strong negative impact on turnover intention when IT employees feel satisfied with their firm [15]. Therefore, we also hypothesize that job satisfaction has a mediator role as follows: Hypothesis 10: Job satisfaction has a mediating relationship between personal demographics and turnover intention of IT professionals: (a) Age, (b) Gender, (c) Experience, and (d) Education level.

K. PROPOSED CONCEPTUAL FRAMEWORK
In this section, the proposed conceptual framework is presented along with its hypothesis. The following proposed framework is motivated by [22] and re-designed to investigate the affecting factors that influence the IT professional turnover intention in the IT and software industry. The IT professional's turnover is relatively high as compared to other sectors. The core construct factors such as recruitment and selection process, teamwork & management support, performance & career management, and salary & compensation plan are considered independent variables. These variables have a direct impact on the turnover intention of IT professionals. In the previous proposed conceptual framework [22], all these factors positively affect employee turnover and retention intention by considering the mediating role of job satisfaction, employee commitment, job security, and recognition. However, recent research studies consider these variables as independent, whereas job satisfaction has a mediating relationship between personal demographics and turnover intention. Therefore, the given proposed conceptual framework considered job satisfaction as a mediator role. Employee commitment, job security, recognition, and organizational demographics are independent variables that directly impact turnover intention. Personal and organizational demographic factors are extracted from SLR. By considering all these factors, the study is helpful for the top management of IT industries by providing deeper insight concerning factors. In Figure 2, the proposed conceptual framework is highlighted.

III. RESEARCH METHODOLOGY
This section presents the research methodology used in this research study. Initially, SLR was performed to identify the turnover intention factors of IT professionals in the IT industry. By following the SLR findings, a conceptual framework was proposed. The empirical analysis is used to validate the VOLUME 4, 2022 5 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication.

A. SYSTEMATIC LITERATURE REVIEW
SLR is an assessment of research questions used as an overt approach to collect and critically analyze data from the currently available studies related to a particular research topic [41]. SLR is also used for finding and extracting data from recent literature in a transparent, unbiased, and reproducible way. The most relevant data is obtained by applying different criteria to recent research studies. Furthermore, all these procedures of SLR are well organized and documented. Different proposed studies have been following the different ways of conducting SLR. Barbara Kitchenham's guidelines [42] recommendations are followed in this study. This guideline includes three review phases: planning, reporting, and conducting. These SLR phases are shown in Figure 4. The following upcoming sections briefly discuss these phases.

B. PLANNING THE REVIEW
In the planning phase, two research questions were addressed to achieve the objectives of our study. Relevant studies are searched and identified from appropriate electronic repositories to answer research questions. Furthermore, search string, inclusion, and exclusion criteria are well-defined, and quality assessment criteria were selected for primary studies.

1) Research Questions
A research question is the primary step of the SLR process. The current research study aims to identify the factors that impact IT professionals' turnover intention. The following research question is addressed to give motivation to our research: RQ1: What are the turnover intention factors addressed in existing literature?

2) Electronic Repositories
The primary studies published between 2016 to 2021 in the following electronic repositories are considered for further research.
The other electronic repositories are not included in our research study due to accessibility issues.

3) Search String
The search string is a sequence of keywords in which Boolean operators; AND & OR are used to create concatenation between keywords. All the search results were carried out based on title, abstract, and keywords. The digital repositories were searched against the following search string: • ("turnover" OR "attrition" OR "job change" OR "turn away" OR "job mobility" OR "job movement" OR "job flow" OR "career change") AND ("software developer" OR "IT professional" OR "IT personnel" OR "software engineer" OR "software coder") AND ("software industry" OR "IT industry"). A modified search string was used in each publication repository because of their attributes and distinct features to query for the selected primary study. The above search sting defines the intersection of articles that include job turnover intention, related factors, type of profession, and domain, either the software or IT industry.

4) Inclusion Criteria
The following inclusion criteria are formulated in this study: • The selected studies were published in a conference or journal. • The selected studies were written in the English language. • The time duration of selected studies lies between 2016 to 2021. • The selected study referred to turnover factors in the software or IT industry.

VOLUME 4, 2022
This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication.

5) Exclusion Criteria
The following exclusion criterion is used: • Those studies were excluded, not written in the English language. • Blog, slides, technical reports, book chapters, white papers, and workshop papers were excluded. • Studies that were inaccessible or unavailable online. • Reluctant or duplicated studies were excluded. • The studies were not relevant to the turnover intention of IT professionals.

6) Quality Assessment of the Studies
Eight quality assessment questions presented in Table 1 are used to identify the significance of selected studies. It also helps to obtain reliable results. Is the proposed study clearly defining scope and context? QA3 Is the proposed research discussed factors that influence job turnover intention? QA4 Does the proposed study propose a conceptual research model? QA5 Is the proposed solution validating by an empirical study? QA6 Are the stated research questions well answered? QA7 Does the conclusion of the study clearly define the aim and objectives? QA8 Does the proposed research provide future work and limitations?
The preliminary studies are included based on the quality evaluation for quality assessment. Table 2 depicts the above studies' score criteria as follows. A score of 1 is for those research studies that have appropriately answered the quality assessment questions. A score of 0.5 is for those research studies that partially answered the quality assessment questions. A score of 0 is for those research studies which have not answered the quality assessment questions.

C. CONDUCTING THE REVIEW
In this section, following steps are considered for conducting the review: • Primary studies selection. • Data Extraction. • Data analysis. All these steps are discussed below in detail.

1) Primary Studies Selection
A Tollgate approach has been used to refine the identified relevant studies through the primary study selection process. This approach includes five levels as follows: • Level 1: Search the appropriate studies based on the search string.
0.    Figure 5 depicts the primary study selection based on the tollgate approach. By using the search string, 7791 studies were obtained from relevant repositories. Initially, the screening process was performed to exclude those studies which were duplicated or irrelevant. The inclusion and exclusion criteria were performed on the remaining 6476 studies. Furthermore, those studies were excluded whether full content was inaccessible or not in English. The remaining 154 studies were assessed for the further process. After that, inclusion and exclusion criteria were performed on the remaining studies based on the relevance of the research topic. Finally, out of 54 studies, employees' turnover intention in different sectors was influenced. The remaining 54 studies were then assessed, out of which seven were excluded and did not identify IT employees' turnover intention in the IT or software industry. Finally, 47 selected papers were included in our data extraction process.

2) Data Extraction
Relevant data is extracted from the primary studies to achieve the research objectives. Those factors that impact on turnover intention of IT professionals, such as recruitment and selection, teamwork, supervisor support, salary and compensation, appraisal policies, and career management, are extracted. Job satisfaction is a mediator variable between personal demographics factor and turnover intention. Job security, recognition, and employee commitment are independent variables. Turnover intention is extracted as a dependent variable.

3) Data Analysis
Different employee turnover intention factors were extracted from 47 final studies. According to the recent literature, these employee turnover intention factors impact IT professionals staying or leaving the IT industry. Some of these factors have great significance in IT organizations. Moreover, the organizations have to face many critical challenges without them, i.e., high loss of revenue cost, low-level of productivity, etc. The most significant critical factors are the recruitment and selection process, team and career management, job satisfaction, employee commitment, job security, and turnover intention. The research questions were formulated against the data collected from the literature.

D. REPORTING THE REVIEW
In this section, reporting the review is performed after conducting the review phase of SLR. There are two steps in reporting the phase as follows: • Types of studies • Temporal distribution of studies  Figure 6. This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication.  Figure 7 presents the temporal distribution trend.

E. EMPIRICAL ANALYSIS OF CONCEPTUAL FRAMEWORK
In this section, an empirical investigation of the conceptual framework has been carried out using the quantitative research method. A survey method has been used to collect the data as it is a more convenient method to gather data. In the survey method, it is crucial to determine the target population. The targeted population was data analysts, programmers, developers, testers, designers, project managers, and CEOs in Pakistan's small and medium-size IT organizations. The sample size sufficient for conducting Partial Least Square -Structural Equation Modeling (PLS-SEM) analysis is 200 or above [70]. Therefore, a total number of 250 responses were used in this study. From the results of SLR, a survey questionnaire was formulated to evaluate the proposed framework. Close-ended questions were selected to collect data from the target population. The current study questionnaire consists of two main sections. The first section focuses on the respondents' personal and organizational demographic information. While the second section provides the factors affecting IT professionals' turnover intention in the IT industry. The fivepoint Likert scale was used which starts from 1 = "strongly disagree", 2 = "disagree", 3 = "neutral", 4 = "agree" and 5 = "strongly agree". It is essential to pre-test the questionnaire before distributing it as it helps to investigate the questionnaire's credibility and reliability. Four steps are included in pre-testing a survey questionnaire, i.e., content validity, questionnaire readability, pilot study, and final review [71].
The content validity of the survey questionnaire was performed to evaluate its authenticity and validity. It also involves observing the questionnaire's items by experts and confirming whether the target respondents could understand the questions or not. The questionnaire readability was neces-sary before conducting the survey. To evaluate the readability of the questionnaire, the grammar, fonts, sentence clarity, engagement, etc., were reviewed. The pilot study was conducted after reviewing the survey questionnaire's readability. Two expert team members of an IT organization reviewed the survey questionnaire to evaluate its validity.
Moreover, the survey questionnaire was revised, and review changes were applied based on the experts' opinions. After collecting the data from the pilot study, the Statistical Package for the Social Science (SPSS) version 26.0 was used for testing. The final survey was distributed to the target population for data collection. The data collection process was started from September 2021 to October 2021. The survey was circulated to around 490 targeted respondents through Google Form, LinkedIn, Facebook, and visiting the onsite software and IT industries.

IV. RESULTS AND FINDINGS
In this section, the findings of SLR are discussed in detail. The extracted data from SLR contains the affecting factors that impact employee turnover intention. After that, the proposed conceptual framework is evaluated by performing an empirical investigation. Finally, a comparison of SLR and empirical study is highlighted.

A. SLR RESULTS
This section extracted a comprehensive list of affecting factors influencing employee turnover intention. By conducting the SLR, 54 studies were selected as primary studies. The quality assessment and tollgate approach was applied to refine SLR results. So, a total of 47 primary studies were finally selected. A total of ten factors were extracted from these primary studies after performing the data analysis process. The number of occurrences was obtained for each extracted factor to calculate their frequency. Table 4 shows the identified factors along with their frequencies and rank.
The results of the SLR present two main aspects. First, the most recent studies were conducted in different countries for a specific region, and limiting factors were presented. Besides, few studies were conducted to compare the employees' turnover intention in the IT industries of developed countries. Several related turnover intention factors were reported in recent studies to decrease the turnover intention of IT professionals. However, no study was conducted to cope with all identified factors impacting employee turnover intentions. Secondly, there was limited literature on employee turnover intention related to Pakistan's IT and software industry. This empirical study aimed to identify the factors that critically impact IT professionals' turnover intention, as shown in Table 4.

B. RESULTS FROM EMPIRICAL STUDY
In this section, the empirical investigation of this study is discussed. The PLS-SEM is used to perform an empirical analysis. Moreover, Warp-PLS and SPSS were considered to evaluate and analyze the proposed conceptual framework.  Furthermore, PLS path modeling indicates the latent variable outcome score that gives symptomatic data about IT professionals' turnover intention. To answer RQ2, a survey was distributed in Pakistan's small and medium-sized IT and software industries. Also, each hypothesis was evaluated and analyzed along with its outcome result.

C. DEMOGRAPHICS ANALYSIS OF RESPONDENT
In this section, the demographic analysis of survey respondents was conducted using descriptive statistics. In this study, the participants' personal and organizational demographics information is gathered through a survey. All the respondents have experience in the IT or software industry. Table 5 depicts the participants' personal and organizational demographic information.
According to survey results, the male respondents were 210 (84%), and the female respondents were 40 (16%) who are currently working in the IT and software industries. This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and In this study, the targeted organizations are the small and medium-sized enterprises in Pakistan's IT or software industry. Therefore, the organizations which contained the number of employees between 0-10 were (6.8%), (16%) between 11-25, (29.6%) between 26-50, the number of employees between 51-100 and 101-200 was (24.8%) and (22.8%) respectively. The department size of organizations contains the number of employees between 0-10 (44%), employees between 11-25 (32.8%), the number of employees between 26-50 and 51-100 was (12.8%), and (6.8%), and between 101-200 employees were (3.6%). Table 5 shows the organizational demographic information of respondents. It also shows no respondents whose organizational size is more than 200 employees.

D. RELIABILITY ANALYSIS
The reliability of the questionnaire and the internal consistency of items were identified through the instruments' reliability analysis. The internal consistency of items was measured to check the association of items with their construct and other construct items within the scale. There are several statistical tests to measure internal consistency. Cronbach's alpha test has been applied to evaluate the internal consistency. It generally depicts the lower and upper bound value of reliability. In exploratory research, the values of Cronbach's alpha lie between 0.6 to 0.7 are acceptable, and a value below 0.6 is considered poor [70]. However, in the advanced stages of research, values between 0.7 to 0.9 are considered good [3]. Moreover, the value of Cronbach's alpha 0.8 is also suggested as good. If Cronbach's alpha value is greater than 0.95, it indicates a chance of duplication; therefore, the value is not accepted. The Cronbach's alpha value of the individual construct was separately explored for reliability analysis. The current study results are statistically significant because all the constructs' values of Cronbach's alpha are within the range. Table 6 shows the significant results of the reliability analysis of each construct.

E. DESCRIPTIVE ANALYSIS
This section discusses the descriptive analysis of the survey study in detail. Descriptive statistics of each construct item are measured using mean, standard deviation, skewness, and kurtosis. Each statistics score shows each item's level of satisfaction and summarizes the survey data. Each construct instrument is measured in this study through two research questions observed by a 5-point Likert scale. Recruitment and Selection (RAS), Team and Management Support (TAMS), Job Security (JSE), Recognition (REC), and Job Satisfaction (JAS) constructs were measured by four items. Salary and Compensation (SAC), Employee Commitment (EC), and Turnover Intention (TI) constructs were measured by three items. In addition, the Performance and Career Management (PACM) construct was measured using five items. The Likert scale extends from strongly disagree to strongly agree. The means score of each item should be in the range of 2.0 to 3.99. According to Kim et al. [73], the range score of skewness is -1.96 to +1.96; otherwise, it is viewed as unacceptable. In contrast, kurtosis's accepted range score value is -3 to +3. The individual items' statistics of each construct are given in Table 7.

F. CORRELATION ANALYSIS
Before performing SEM, the correlation analysis was performed and analyzed among the constructs. It helps to assess the association between dependent and independent variables. In this study, Pearson's Correlation Coefficient was used to determine the correlation among the proposed conceptual framework constructs. According to [74], the correlation coefficient was interpreted by the following guidelines: • The coefficient value lies at ±1; it shows a strong positive or negative linear relationship -both variables have linearly increased or decreased.  Table 8 depicts the correlation analysis matrix among dependent and independent variables' constructs in the proposed conceptual framework. According to the results, all these construct variables have a strong correlation coefficient value.

G. RESULT OF STRUCTURAL EQUATION MODELING
PLS-SEM is used in this study to analyze the proposed conceptual framework using SPSS and Warp-PLS. PLS-SEM is an inclusive multivariate analysis method containing two sub-models: measurement and structural model. The measurement model indicates the interrelation between the latent variable and collected data from the survey. The structural model shows the relationship among the latent variables. PLS-SEM is especially considered when the conceptual framework is complex, has many latent variables, and gives more reliable results [75]. SEM helps evaluate the relationship between exogenous and endogenous variables rather than taking them separately. Significantly, it helps to correctly analyze the proposed conceptual model and respec-tive proposed hypothesis with acceptable value [76]. The measurement model analysis and structural model analysis are discussed in detail.

1) Measurement Model Analysis
There are two types of latent construct measurement models; formative and reflective. The construct variables influence all the latent variables in the formative measurement model. While in the reflective measurement model, all the construct variables are influenced by the latent variables [77]. The proposed conceptual framework is a reflective type in this study. According to [78], path modeling is a generally developed system used for cause & effect relationships between latent variables. The measurement model analysis was conducted to evaluate the relationship between construct and latent variables. It assesses the collinearity between these constructs. There are two ways to validate the measurement model analysis; convergent validity and discriminant validity.

a: Convergent Validity
Convergent validity determines how well different latent variables of construct correlate positively or negatively. Two metrics are crucial for evaluating convergent validity: Average Variance Extracted (AVE) and standardized outer-loading of latent constructs' variables. Table 9 represents the crossloading of the constructs' latent variables. According to [3], a standardized outer-loadings value should be greater than 0.70; however, it does not mean the outer-loadings values below this threshold must be removed. If the range value of outer-loading is between 0.40 to 0.70, the removal effect on the AVE should be considered. If the AVE significantly improves, remove the items; otherwise, retain them in the model. In this empirical analysis, all the standardized outer loading values of constructs are greater than 0.70 given in Table 9.
The composite reliability values between 0.80 to 0.90 represent the high correlation of construct variables [3]. In Table 10, all the composite reliability values of the construct are within the range and acceptable. In addition, the AVE is measured to assess the average variance shared between 12 VOLUME 4, 2022 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and  the construct and latent variables. The AVE value of each construct is between 0.50 to 0.80. Moreover, the value of AVE should be higher than 0.50 [3]. Results have shown in Table 10 that all the constructs' AVE values are greater than a threshold value.

b: Discriminant Validity
In discriminant validity, different constructs of conceptual framework describe different concepts or phenomena. In addition, it measures the distinct behavior of each construct with other constructs. It also implies that each construct of the conceptual framework has unique and unrepeated characteristics which are not determined by other constructs.
There are two common methods to evaluate the discriminant validity.
Firstly, each factor of the latent variable's cross-loadings and outer loadings value is investigated. The outer-loadings of each construct's factor should be greater than any other construct's factor. Suppose the latent variable loads a higher 13 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication. [79] that remove the latent variable that does not measure respective construct loadings-measuring higher standardized loading of any other construct. Table 9 depicts that all the standardized loading of latent variables loads highest onto their particular construct. Secondly, the discriminant validity is evaluated using the Fornell-Larcker criterion. It states that the value of each respective square root of construct' AVE should be higher than the construct correlation with other constructs. In simple terms, the individual construct shares high variance with its latent variable than other constructs. Table 8 presents the constructs correlation matrix with the square root of AVEs on the diagonals. Each construct value on the diagonal is higher than the other constructs' correlation value, which validates the Fornell-Larcker criterion.

2) Structural Model Analysis
The structural model analysis is applied to evaluate this study's proposed conceptual framework and hypothesis. Figure 8 presents the structural model with outer loadings and path estimation coefficients. Blue circles represent the independent variables, red circles show independent control variables, and a white circle presents the dependent variable. The blue dotted line depicts the mediation effect between independent and dependent variables, discussed in detail.
The multiple regression analysis is performed to investigate the relationship between dependent and independent variables. Before giving the results, it is essential to evaluate the correlation of regression coefficient among construct variables, which shows the impact of the relationship. Equation  1 demonstrates the multiple regression model [22]; where 'y' is dependent variable, 'c' is constant, 'x 1 , x 2 , ..., x n ' are independent variables, and 'b 1 , b 2 , ..., b n ' are regression coefficients.
To evaluate the path estimation coefficient, R-square value, beta coefficient, t-statistics, and p-value of significance level, have been analyzed. The range of R-square values is -1 to +1. If the R-square value is close to +1, show the strong correlation among variables. An R-square value close to 0 shows a weak positive correlation, and -1 represents a strong negative correlation among variables. R-square value demonstrates how much variance in the independent variable is affected by the dependent variable [22]. The beta coefficient depicts the strength of each independent variable on the dependent variable. The t-stats for two-tail significance were considered to measure whether the relationship among variables was significant or not. The acceptable value of t-stats for twotail are to t > 1.96 at p < 0.05, t > 2.567 at p < 0.01, and t > 3.29 at p < 0.001 [80]. The R-square for turnover intention is 0.775, indicating 77.5% variance predicted by all independent variables. Hence, the structural measurement model presents statistically significant results. In Table 11, the beta coefficient, R-square, t-statistics, p-value, and hypothesis supported or not supported are presented.

3) Hypothesis Evaluation
In the current study, ten hypotheses were proposed to investigate the influence of independent variables on a dependent variable. The multiple regression analysis and PLS-SEM were used to test the hypothesis using SPSS and Warp-PLS. Table 11 presents the direct impact of individual independent variables on dependent variable (turnover intention).
The Recruitment and Selection (RAS) process has a high impact on the IT professional Turnover Intention (TI). The survey results show the high collinearity of RAS with IT professionals' TI. A strong negative impact of RAS on TI with R-square value (0.546), beta coefficient value (-0.739), t-stats at -17.284, and p-value<0.0001, respectively. The results show a significantly negative relationship between RAS and TI, supporting our hypothesis 1. The regression analysis was performed to predict the influence of Team and Management Support (TAMS) on the TI of IT professionals. The survey results establish the relationship between TAMS and TI with R-square value (0.616), beta coefficient value (-0.785), t-stats at -19.946, and p-value<0.0001, respectively. The results demonstrate that TAMS has a significantly negative relationship with TI. So, hypothesis 2 is accepted in this study.
The negative relationship between Performance and Career Management (PACM) and TI was observed using Rsquare and beta coefficient values, given as 0.615 and -0.785 with t-stats -19.888. As the PACM increases, the employees' intention to stay long-tenure in their organization. The results show a significant relationship between PACM and TI at p-value<0.0001. Therefore, hypothesis 3 is supported in this study. According to our findings, the beta coefficient and R-square values were obtained as -0.658 and 0.433, which shows the moderate negative impact of Salary and Compensation (SAC) on TI. The two tail significance level was supported with t-stats and p-value as -13.775 at < 0.0001. Hence, hypothesis 4 is accepted and supports our finding results.
The Employee Commitment (EC) has a negative impact on TI with R-square value (0.500), beta coefficient value (-0.707), t-stats -15.749 at p-value<0.0001, respectively. As the EC increases, the intention of IT professionals decreases to leave the organization. The results support our findings that EC negatively influences the TI of IT professionals. So, hypothesis 5 is supported in this study. The Job Security (JSE) negatively influences employee TI with R-square value (0.470), beta coefficient value (-0.685), and t-stats -14.825 at p-value<0.0001, respectively. The IT employees who feel secure about their job have less intention to leave the industry. So, the finding shows that hypothesis 6 supports our research study. The IT employees whom their supervisors or team leaders have appreciated; have more commitment to their job. Our survey results depict that the Recognition (REC) has a negative impact on the TI of employees. The 14 VOLUME 4, 2022 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication.   and Department Size (Dept_Size). The direct impact of Org_Size and Dept_Size are investigated with TI. Our survey results represent the R-square value (0.017), beta value (-0.131), t-stats -2.082 at p-value<0.038 for Org_Size and Rsquare value (0.022), beta value (-0.147), t-stats -2.344 at p-value<0.020 for Dept_Size. Both variables have a negative influence on the TI of IT professionals. Hence, the hypothesis H8a and H8b are significantly satisfied in this study.
The Personal Demographics variable has four subcategorical control variables; Age, Gender, Experience (Exper), and Education Level (Edu_lvl). Firstly, the direct impact of Age and TI was measured with an R-square value of 0.004, the beta value of 0.061, and t-stats 0.965 at p-value >0.335 (>0.05) respectively. The results show that Age has a nonsignificant impact on TI of IT employees, which does not support our hypothesis H9a. Secondly, the direct impact of Gender was evaluated with TI. The R-square and beta values were given as 0.001 and -0.032, showing a weak negative relationship. The t-stats and p-value were also measured to check the significance level as -0.505 at >0.614 (>0.05). The results show a non-significant relationship between Gender and TI, which rejects hypothesis H9b. Thirdly, the direct impact of the Exper and TI was observed. The R-square and beta values were given as 0.019 and 0.039, which shows the weak negative relationship between them. The t-stats and pvalue were obtained as -2.176 at <0.031, which presents the relationship's significance level. Hence, hypothesis H9c is accepted with a significance level of less than <0.05. Finally, the direct impact of the Educ_lvl was measured with TI to investigate their negative relationship. The R-square value 0.039, beta coefficient value -0.198, t-stats -3.185 at p-value <0.002 respectively. The results show that Educ_lvl of IT professionals increases, their TI to leave the organization decreases. So, hypothesis H9d is supported in the current study.
A bootstrapping method was adopted to evaluate the mediating effect of Job Satisfaction (JSA) between Personal Demographics and TI using SPSS Process Macro [81] as shown in Figure 9(a). First, the regression analysis was performed using Age to predict the significant effect on JSA. The results show the beta value = .039, t-stats = .260, p>.795, respectively. Next, the second regression analysis was performed while controlling the JSA (as mediator) between Age and TI. The result predict the Age has a non-significant impact on TI (beta value = -.139, t-stats = 1.709, p>.080). Based on 5000 bootstrapping samples, the result of the indirect effect show a non-significant relationship between Age and TI mediated by JSA (a*b = -.023, Bootstrap CI 95 = -.185 and -.142). Similarly, there was non-significant direct effect between Age and TI (beta = -.139, t-stats = 1.709, p>0.05). Table 12 shows the evaluation of mediation model between Age, JSA and TI. Hence, hypothesis 10a is rejected and does not support this study.
In Figure 9(b), the mediation effect between Gender and TI was investigated by controlling JSA. First, the regression analysis was performed using Gender to predict the significant effect on JSA. The results show a non-significant relationship with beta value = .061, t-stats = .361, p>.719, respectively. Next, the second regression analysis was performed between Gender and TI by mediating JSA. The result predict the Gender has a non-significant impact on TI (beta value = -.032, t-stats = -.353, p>.724). Based on 5000 bootstrapping samples, the result of the indirect effect show a non-significant relationship between Gender and TI mediated Similarly, there was non-significant direct effect between Gender and TI (beta = -.032, t-stats = -.353, p>0.05). Table  13 shows the evaluation of the mediation model between Gender, JSA, and TI. So, hypothesis 10b is not supported in this study. The mediation effect between Exper and TI was investigated by controlling JSA presented in Figure 9(c). First, the regression analysis was performed using Exper to predict the significant effect on JSA. The results show a weak positive significant relationship with beta value = .104, t-stats = 2.087, p<.038, respectively. Next, the second regression analysis was performed between Exper and TI by mediating JSA. The result predict the Exper has a non-significant direct impact on TI (beta value = -.026, t-stats = -.945, p>.345). Based on 5000 bootstrapping samples, the result of the indirect effect show a significant negative relationship between Exper and TI mediated by JSA (a*b = -.061, Bootstrap CI 95 = -.118 and .008). The total effect on TI [P M = (-.061) / (-.087)] was approximately 70.1% affected by mediator JSA and Exper. Similarly, there was a negative significant direct effect between Exper and TI (beta = -.087, t-stats = -2.176, p<0.05). Table 14 shows the evaluation of the mediation model between Exper, JSA, and TI. So, hypothesis 10c is supported in this study. Figure 9(d) shows the investigated results of the mediation effect between the Educ_lvl and TI by controlling JSA. First, the regression analysis was performed using Educ_lvl to predict the significant effect on JSA. The results show a weak positive significant relationship with beta value = .134, t-stats = 4.282, p<.000, respectively. Next, the second regression analysis was performed between Educ_lvl and TI by mediating JSA. The result predict the Educ_lvl has a non-significant direct impact on TI (beta value = -.002, t-  .081, t-stats = -3.185, p<0.05). Table 15 shows the evaluation of the mediation model between Educ_lvl, JSA, and TI. So, hypothesis 10d is supported in the current study. Finally, the multiple regression analysis was performed to predict the model fit using independent, control, and dependent variables. The R-square value was calculated, explaining the percentage of variance and strength of the relationship between independent and dependent variables. The range of R-square values is between 0 to 1. The R-square value of 0.75 shows a strong effect, 0.50 shows a moderate effect, and 0.25 shows a weak effect [3]. The structural model in Figure 8 indicates the R-square value of 0.755 that predicts the 75.5% variance in turnover intention by using the independent and control variables.

4) Triangulation of SLR and Empirical Study
This section discusses the triangulation results of the SLR study and empirical study. Figure 10 depicts the comparison VOLUME 4, 2022 17 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2022 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ of identified influencing factors through SLR and empirical study. Two types of datasets are used to evaluate the likeness and variations. In Table 16, the ranking of identified affecting factors through SLR and survey are shown. In this empirical study, a close-ended survey questionnaire was formulated to investigate and identify the affecting turnover intention factors in the IT and software industries. The positive occurrence responses were chosen from the collected survey data, i.e., agree and strongly agree.
Using SPSS, spearman's rank-order correlation was used to assess the significant difference between the SLR and empirical ranking of identified affecting turnover intention factors. The spearman's correlation coefficient value was obtained as 0.627, which indicated that the SLR and survey ranking were moderately correlated. The p-value was given as 0.039 (<.05); it demonstrates that the obtained correlation was statistically significant as implied in [82]. The spearman's correlation between the SLR rank and empirical rank is presented in Table 17.
This study compares the SLR and empirical study ranking results to investigate the critical affecting factors that impact the turnover intention of IT professionals. According to [82], the criticality of affecting factors was measured through the proposed criteria. If the frequency of affecting factors is >=50% in both SLR and empirical study, it is considered a critical factor. The current study results demonstrate that only one critical influencing factor in SLR and empirical study is TI (turnover intention). The findings also identified that eight affecting factors are considered critical except salary & compensation and organizational demographics among Pakistan's IT and software industries.
Therefore, the comparison results help the top management of IT and software organizations pay attention to these critical factors while developing new retaining strategies. Furthermore, the finding also highlights which critical factor has the most influencing impact and non-significant impact on IT professionals' turnover intention.

V. DISCUSSION
In this study, it has been observed that IT professionals have a critical role in the tremendous growth of Pakistan's IT and software industry. IT organizations have to face several challenging issues such as lack of skilled employees, decrease in production growth, the high turnover intention of skilled employees, high loss of annual revenue, etc. Among these challenges, one of the most cited challenging issues [19], [22], [25] is the high turnover intention of IT professionals in recent research studies. Therefore, it is necessary to explore the affecting factors of the turnover intention of IT professionals. In recent literature, several authors have developed different types of conceptual frameworks and strategies in the IT and software sector to identify the factors that impact the turnover intention of employees [16]- [18], [22]. However, according to our best findings, none of the studies has addressed these challenging issues. Moreover, no one has empirically investigated the impact of turnover intention factors in the Pakistan software and IT industries. The current study aims to identify and investigate the influencing factors of IT professionals in the IT or software industry to fill this gap.
An SLR was conducted to answer the RQ1. A total of 47 primary research studies were selected by applying inclusion, exclusion, and quality assessment criterion. After that, ten important affecting factors have been identified that impact turnover intention. Recent literature has investigated the different influencing factors of turnover intention in their specific area or region. Although some factors were considered common in all studies, several factors differed from one region to another region. In this study, the primary purpose is to identify all these factors and propose a conceptual framework model. Furthermore, the hypotheses were proposed to evaluate the conceptual framework.
The empirical analysis was performed to analyze the proposed hypothesis and answer RQ2. A survey was conducted to target IT professionals in the Pakistan software and IT industries. Out of 490 respondents, 96 survey responses were collected through online Google Form, and 184 survey responses were gathered through a printed survey questionnaire. Finally, the missing and unfilled questionnaires were excluded, and the remaining 250 responses were used for data analysis.
In the proposed conceptual framework, a total of ten independent variables were identified to explore the direct impact on turnover intention. In this study, we argue that the recruitment and selection process has a negative impact on the turnover intention of IT professionals. According to [25], the HR management strives to build a friendly working environment for new talented employees to increase the productivity growth of the organization. The past studies explore that the well-organized recruitment and selection process increases the interest of employees to stay connected with their organization. As a result, employees' intention to leave the organization was decreased. The new internees were focused on their learning new skills and induction programs 18 VOLUME 4, 2022 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication.   that help in the growth of career path [3]. According to our results findings, the recruitment and selection process negatively affects turnover intention (b = -.739, p-value < .0001), which supports our hypothesis 1. So, the result benefits the organization in developing new induction programs that engage the new employees to stay long.
The team and management support plays a significant role in the turnover intention of IT professionals. According to [27], the relationship of IT professionals with their supervisor, team members, and HR management has a critical impact on turnover intention. Several studies argue that a strong relationship would encourage the IT professional to feel secure and satisfied with their job [17], [28]. Supervisor behavior has positively increased the intention of team members to stay with their organization; as a result, their intention to leave the organization is negligible. Our research shows that team and management support is negatively influenced by the turnover intention of IT professionals (b = -.785, p-value <.0001). Therefore, hypothesis 2 is significantly supported by the recent literature findings. Above mentioned results highlight that IT organizations pay attention to team management and supervisor support for a better future.
According to [2], [24], the career growth, training, and learning opportunities of IT professionals decrease their intention to leave the organization. Several organizations did not pay attention to performance and career management, resulting in the loss of skilled employees. The management should acknowledge their key employees and increase the career opportunities within their department so that the job engagement of these key employees increases [21], [29]. Our survey results satisfied the literature findings that team and management support negatively impacts IT professionals' turnover intention (b = -.784, p-value <.0001). So, the proposed hypothesis 3 is accepted in this study. The authors in [36] explored that the lack of salary and compensation plans has been one of the most common reasons for the high turnover intention of IT professionals. Organizations have to concentrate on salary and compensation plans for talented employees to reduce turnover intention. Most employees highly consider compensation plans before joining a new firm. According to recent studies [1], [22], [30], salary and compensation have a positive relationship between them and a negative impact on turnover intention. Our findings significantly support the hypothesis 4 (b = -.658, p-value <.0001).
According to the Human Capital theory, the highly committed employees stay long tenure when their organization provides efficient development and training programs [31].
In the past few studies [2], [23], [25], many researchers explored the influence of employee commitment on turnover intention. They argue that when the employee commitment has increased, a decrease in turnover intention has been observed. In the current study, the survey result supports our proposed hypothesis 5, which states that employee commitment negatively affects the IT professionals' turnover intention (b = -.707, p-value <.0001). In [33], the authors investigate that the IT professionals who feel secure about their job, their satisfaction level increases; as a result, they have less intention to leave their working organization. In management literature, the authors investigate the impact of job security on the turnover intention of IT professionals [17], [34]. They argue that high job security provides high employment stability, which reduces employees' turnover intention. The result predicts that job security has a negative impact on turnover intention (b = -.685, p-value <.0001), which supports our proposed hypothesis 6. Therefore, the given results help the organizations to develop a comfortable and secure environment for their valuable employees.
According to [35], the IT employees not only need a better salary and compensation plans from their organization but also want to be appreciated by their management. The valuable employees stay connected to their organization and have less intention to find a new job somewhere else. The authors in [36] highlight the important aspect that proper recognition and reward system decreases the professional's intention to leave their job. The findings of this study observed that the recognized IT professionals have a negative impact on turnover intention (b = -.703, p-value <.0001). Therefore, the proposed hypothesis 7 significantly supports this study. In recent literature [15], [21], [26], the authors argue that the employees working in high brand and large enterprises have a low level of turnover intention. The IT staff working in small size organizations has less intention to leave. The findings of our survey results suggest that organizational demographics (organization size and department size) has negative impact on turnover intention of IT professionals (b = -.131 with pvalue < .038 and b = -.147 with p-value <.020). So, the proposed hypotheses 8a and 8b are statistically significant.
The direct and indirect impact of the personal demographics on turnover intention has been observed in this study. In the past few studies [37], [38], authors investigated that the young IT professionals have high intention to leave or change their job as compared to employees above 40 years. In this study, our results depict a non-significant impact of age on the turnover intention of employees (b = .061, p-value >.335). Therefore, hypothesis 9a is rejected and not supported in this study. The mediation effect of job satisfaction is also evaluated between age and turnover intention. In literature, the age factor has a small weak impact on turnover intention [33]. However, our study result does not support the mediation effect of job satisfaction between age and turnover intention (b = -.023, p-value >.05). Therefore, hypothesis 10a is also rejected. The gender gap of IT professionals has another aspect that impacts on turnover intention of IT professionals in recent literature [26], [39]. According to [26], the female employees in the IT sector leave the organization 2.5 times higher as compared to male employees. In our survey result, the male respondents were 84%, which shows that male respondents were higher than females. The direct impact of gender has non-significant effect on turnover intention (b = -.032, p-value >.614). So, the proposed hypothesis 9b is not supported. Furthermore, the mediating effect of job satisfaction is measured between gender and turnover intention (b = -.036, p-value >.05). The results depict that hypothesis 10b is rejected due to the non-significant effect on turnover intention.
The employees' experience has been gathered from survey data. According to [15], [21], the IT professionals having high experience have less intention to leave the organization.
In our results, most employees work in small and mediumsized organizations. Moreover, these employees have less intention to leave the organization. The results show that employee experience has been negatively influenced by turnover intention (b = -.137, p-value <.031) and accept hypothesis 9c. Furthermore, the indirect relationship between employee experience and turnover intention was evaluated by controlling the job satisfaction as a mediator. The results show the significantly full mediation effect between experience and turnover intention (b = -.061, p-value <0.05) and support our proposed hypothesis 10c as shown in Table 14. The education level of employees was measured through survey data. In our results, most of the employees completed their bachelor's degree, as shown in Table 5. The regression analysis was performed to assess the direct impact of education level and turnover intention. The result illustrates a weak significant effect between education level and turnover intention (b = -.081, p-value <.002). Finally, the mediating effect of job satisfaction between education level and turnover intention is also evaluated. The result shows the significant impact on turnover intention (b = -.079, p-value <0.05). Therefore, our proposed hypotheses 9d and 10d are significantly supported.
Conclusively, the current research study satisfied all relationships among dependent and independent variables except age and gender control variables. Furthermore, this study aims to identify the affecting factors of IT professionals' turnover intention in Pakistan's IT and software industry. Moreover, the evaluation of these affecting factors was done using the recent literature review and data analysis. The results of this study help the top management of IT and software development organizations pay attention to these affecting factors.

VI. CONCLUSION
In the past few years, IT and software organizations have paid attention to acquiring talented and skilled employees. It has been observed from the recent literature that high employee turnover intention is one of the critical issues. Due to the high turnover intention of employees, it is difficult for organizations to retain them. Therefore, the SLR was conducted to identify the critical factors influencing IT professionals' turnover intention. From the findings of SLR, the identified affecting factors are recruitment & selection process, team & management support, performance & career management, salary & compensation, employee commitment, job security, recognition, and organizational, and personal demographics. Furthermore, the conceptual framework was proposed along with research hypotheses to evaluate the affecting factors empirically. A survey has been circulated to 490 participants working in the IT or software industries.
The findings of this study demonstrate the negative impact of employee turnover intention on influencing factors in Pakistan's IT and software industries. The retaining of key employees starts with the recruitment and selection process. Well-organized team & management support encourage talented employees to stay connected with their organization. 20 VOLUME 4, 2022 This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication. Citation information: DOI 10.1109/ACCESS.2022.3181753 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ Salary and compensation play a significant role in attaining valuable employees because a compensation plan is the main attraction for employment. Organizations have to pay attention to employee commitment, job security, and recognition factors to reduce turnover intention. Personal demographic factors were evaluated using job satisfaction as a mediator with turnover intention. However, the non-significant results of age and gender were obtained, which depict the weak effect on employee turnover intention. The findings of this study are helpful for the top management of organizations to retain their skilled and talented employees.

A. LIMITATION AND FUTURE WORK
Generally, all the research studies have some limitations or drawbacks. So, this study also has a few limitations. First, in this study, the most significant factors of the turnover intention of IT professionals were identified from SLR and considered for further analysis. However, in the future, researchers may analyze a few other factors such as work environment, work-life balance, work-family conflict, and employee engagement, which may play a significant role in employee turnover intention. In addition, researchers may identify more factors from the industry through interviews. To overcome this limitation, the authors of this study performed triangulation of SLR and empirical analysis findings. Secondly, the data was collected only from Pakistan's small and medium-sized IT and software industries. Therefore, the study results cannot be generalized. Applying the same study to other countries and sectors is suggested to overcome this limitation. During the study, it was observed that psychological factors and personality traits that cause mental health issues might lead to high turnover. Therefore, researchers can explore this area for further research in the future.