What Really Matters? Turnover Intention Factors of Software Developers

This study's purpose is to identify the factors influencing turnover intention in software developers. Its relevance is related to the shortage of software developers in the labor market both within and outside the EU, and their relatively short employment periods in individual companies, which causes high costs and knowledge outflow. This study therefore aims to analyse the links between different combinations of causal conditions: human resource management climate, affective and calculative commitment, work engagement - dedication, horizontal trust, gender and tenure of software developers. In addition, the study analyses the outcome [the presence or absence of turnover intention] among software developers tested, using a qualitative comparative analysis and fuzzy-sets method. Thus far, scholars have assumed that affective commitment is a fundamental factor in discouraging turnover intention. However, this study shows that calculative commitment, rather than affective commitment, plays such a role in case of software developers. Still, it is worth noting that HRM process, affective commitment, and work engagement are noticeably lower among individuals planning to change their job. The study shows that the relationship between factors is nonlinear, highlighting the need for a nuanced understanding of the interplay among various conditions. The findings offer valuable insights for organizations seeking to tailor strategies to retain software development professionals.


INTRODUCTION
In a highly competitive working environment, attracting and retaining skilled IT professionals is a significant challenge [1].This is especially true for software developers [2], who are an extremely mobile professional group due to the fact that they can provide work from anywhere connected to the internet.It is also characteristic of them to move to a new company every few years [3,4].Previous studies characterise software developers as seeking challenging jobs, frequently leave their organisations when their expectations are not met, and emphasise that their competencies are highly transferable and enable them to change jobs easily [5].At the same time, it is worth noting that the costs of leaving and replacing departing software developers for organisations are high due to their highly specialised skills [6].The causes of employee turnover are multiple and complex, and still poorly understood [7].However, a considerable amount of research indicates that higher turnover intention leads to higher actual turnover rate, so turnover intention may be a valid proxy of actual turnover behaviour.[8].Moreover research also indicates that this relationship is stronger if cost and uncertainty relating to leaving are lower [9].
Previous research on the causes of turnover intention software developers focus on factors such as, for example, human resource management (HRM) practices [10], job satisfaction [11], work stressors [12], psychological contract [13], and organisational commitment [14] and others.However, a review of the literature identified that limited studies have assessed to calculative commitment [15,16].The literature highlights that the reason why software developers change jobs frequently is because they have a higher commitment to the profession than to the organisation [16].Hence, it is conjectured that software developers' commitment to the organisation has a largely calculative character.It is also worth analysing the impact of affective commitment on software developers' turnover intention in order to identify the real importance of this factor especially in comparison with the affective commitment, considered to be the most valuable component linking the employee to the organisation.
The above also prompts us to test the effect of horizontal trust [among co-workers] on software developers' turnover intention to the importance of horizontal relationships in this profession.Although a lot of research has been devoted to the impact of specific HRM practices on turnover intention, so far none has tested the HRM process.It is also worth examining the impact of the dedication component of work engagement as a defining factor of attitudes to professional tasks within the organisation.
Taking these issues into account, the existing body of work recognises a gap in research into the determinants of turnover intention among software developers.This is all the more so as many authors [17] suggest the need for further research into additional antecedents and moderators in turnover research.To date, no studies analyse if the combined effect of all these variables lead to the presence or the absence of turnover intention of software developers.This study therefore aims to analyse the links between different combinations of causal conditions: HRM climate, affective commitment and calculative commitment, work engagement -dedication, horizontal trust, gender and tenure of software developers.In addition, the study also analyses the outcome [the presence or absence of turnover intention] among software developers tested, using a qualitative comparative analysis and fuzzy-sets method.Finally, our study has four primary contributions: I This study fits in with the need to search for new predictors of turnover intention among IT professionals to which software developers belong, to identify the specific characteristics of this professional group [18].For example, due to the calculative type of software developers' attachment to the organisation, the research model tested calculative commitment versus affective commitment in order to gain deeper insights regarding the nature of software developers' attachment to the organisation [15].
II This research was conducted on a homogeneous group of software developers that has not been sufficiently studied in the context of turnover intention.Thus, the present study also responds to calls for occupation-specific research [19].
III The present study also responds to calls for research to study turnover intentions of IT professionals within different cultural contexts [20], due to the scarcity of this type of research in the Polish context.
IV From a practical point of view, understanding the factors influencing the turnover intentions of software developers will also help enterprises in developing HR-related strategies and actions aimed at retaining skilled software developers.A review of the extant literature provides the theoretical basis to build the propositions this study explores.

LITERATURE REVIEW
Turnover intention represents one's voluntary intent to leave an organization to which they are presently attached to.Typically, it emerges when an employee perceives another employment option as more attractive and his/her expectations of the employer are not sufficiently met.Research analysing turnover intention determinants spread in different directions in the search for ways to explain these employee impressions and feelings that precede turnover [21].Research in this area therefore focuses on the search for turnover intention predictors that explain this phenomenon [22].Worth citing in this context is a classification based on Ghapanchi and Aurum's [20] study of a group of IT workers where the following potential predictors are listed: person-related factors, including demographics, health conditions or employment history; psychological and environmental factors; and organisational and work-related factors such as job characteristics and fit, role behaviour and stressors, intra-organisational relationships, job satisfaction and organisational commitment.A significant group of researchers have tested the relationship between HRM practices and turnover intention [23].Among the antecedents of turnover intention, HRM practices are the most studied [24].Few studies have been devoted to HRM process.The HRM process is understood as communicating the values of the HRM philosophy and policies to employees, shaping the 'image', the rank of the HR function, the message regarding the contribution of the HR function to the organisation's performance or the behaviours desired and rewarded by the organisation [25].Due to individual differences, individual employees may perceive these messages somewhat differently.Indeed, it has been shown that there are considerable differences in the perception of the same work environment by different employees.It is therefore reasonable to assume that the perception of HRM activities may vary: firstly, according to the position occupied by the individual in the organisation; secondly, it may depend on individual factors [including psychological factors] as well as on processes within the organisation.All this makes it impossible to put an equal sign between the practices implemented in an organisation and their perception by employees [26].Hence the important role of the HRM process, i.e. the way in which information about the opportunities offered by HR policies is disseminated.The messages communicated to employees should therefore be: understandable, consistent and should be characterised by an alignment between statements and actions.Any changes resulting, for example, from the need to adapt the system to the competitive environment, changes in the labour market or in connection with system improvements should be communicated to employees as soon as possible, together with an explanation of the reasons.The following proposition is therefore put forward.
P.1.HRM exerts a negative effect on software developers turnover intention.
Organizational commitment is the employee's psychological attachment to organization, which may have a different effect on the behavioural orientation of individuals, especially when the organization organisation does not meet employees' expectations.There are many conceptualisations of organisational commitment, the most well-known being by Meyer and Allen [27].This conceptualisation points to the existence of three types of commitment: affective, which describes an emotional relationship with the organisation based on the organisation's attractiveness to the employee; normative, which describes a sense of obligation to stay with the organisation due to moral obligation; and continuance, which results from the desire to avoid switching costs.In contrast, another conceptualisation emphasises the existence of two components calculative, which is instrumental and based on the calculation of a bundle of both financial and non-financial benefits, and usually associated with the beginning of the employment period, and affective, which develops during the employment period [28].The significant interest in the phenomenon of organizational commitment is related to its relationship with turnover intention.Strongly committed employees wish to remain in their organization and work for them [27].In this context, many researchers point to the affective component's highest significance, whose high level causes an employee to choose to remain in the organization.The research results also confirm the existence of a negative dependence between affective commitment and turnover intention [e.g.29].Previous research seems to suggest low organizational commitment among software developers [30].It can be assumed that duration attachment has lost its meaning for these software developers, who do not experience the problem of finding a favourable employment alternative when they decide to leave the organisation.Therefore, in this research we test two types of commitment affective and calculative.
P.2.Affective commitment exerts a negative effect on software developers turnover intention.P.3.Calculative commitment exerts a negative effect on software developers turnover intention.
Schaufeli et al [31] point out that work engagement is a positively work-related state of mind that is characterised by high levels of energy, passion, dedication and full concentration on work.It is seen as the opposite pole of work burnout.Three components are distinguished within it: vigour, understood and measured as the level of energy and psychological resilience at work, dedication, resulting from pride in one's work and a willingness to identify with it, and absorption, manifested by concentration on work and difficulty in disengaging from it.To date, little research has addressed the impact of work engagement in the context of software developers.A study by Ramaprasad and colleagues [24] on a sample of IT professionals with more than 3 years of work experience found a negative impact of work engagement on turnover intention.In this research model, only the dedication component was tested, which seems to be of a more permanent nature and driven by an assessment of the importance and attractiveness of the tasks performed, and may therefore bind the employee to the organisation.
P4. Work engagement -dedication exerts a negative effect on software developers turnover intention.
Software developers, like other IT professionals, are characterised as strongly identifying with their professional group and willing to work with people they consider to be experts.Horizontal trust implies confidence in the honest intentions and competence and reliability of colleagues.Horizontal trust relationships stimulate cooperation and reduce conflict [32].They are a prerequisite for effective communication and information exchange in organisations and teams [33].They also positively influence the level of organisational commitment and the perceived effectiveness of work performance, and reduce the turnover intention [34].A study of software development teams [35] showed the impact of trust on the speed of new product launches and the cost reduction of these activities.The above became the basis for the following proposition: P.5.Horizontal trust exerts a negative effect on software developers turnover intention.
The calculative approach based on side-bet theory [36] emphasises the employee's investment in the organisation, such as the time it takes to learn and understand the organisation, acquire skills specific to the work environment, and establish relationships with colleagues [36].An employee's achieved status, position, salary, established friendships -unite the employee with the organisation, which, combined with the fear of change and loss of existing privileges, makes the employee more inclined to stay with the organisation i.e. lower intention to leave as the length of employment increases.It was therefore decided to test the following proposition in relation to software developers.P.6.Tenure influences software developers turnover intention.
Most research to date indicates that women do not generally differ in their intention to leave from men, but that women are slightly more likely than men to leave their jobs for personal or family reasons, especially during the initial period of employment [37].The above and the lack of such research among software developers prompted the following proposition.P.7.Gender influences software developers turnover intention.

Data collection and sampling
The study obtained data through a self-administered questionnaire to software developers, participants of postgraduate studies in IT project management in three consecutive editions.Participants were invited to an anonymous survey in which participation was voluntary.In total, 227 correctly completed questionnaires were obtained.The age of the respondents ranged from 22 and 56 years, with an average age of 32 years.Each person had higher education.Of this group, 82% worked in the private sector.The respondents were mainly employed as specialists [66%] and as managers [34%].
Less than 45% of the respondents had 0-3 years of work experience in the company.Almost 60% of the respondents worked in companies employing more than 500 employees.The smallest group [1%] was made up of employees of companies employing 0-10 employees.Others: 11-50 employees -15%, 51-250 employees -18%, 251-500 employees -6%.The respondents' answers allowed for the identification of five constructs: I human resources management (HRM) process, II affective commitment, III calculative commitment, IV work engagement, V horizontal trust.Additional variables included in the study are dichotomous variables that describe the gender [0-female, 1 -male] and work experience in the company [1 -under 3 years 1, 0 -above 3 years].The items used in the process of generating constructs are listed in Table 1.
Table 1 also includes selected statistics used to verify the reliability of the model constructs.Cronbach's alpha coefficients are all above 0.7 and at the same time below 0.9.This means that the internal consistency of constructs is satisfactory [38].The values of the average variance extracted [AVE] for the variables exceed the 0.50 threshold.Thus, the construct's ability to capture variance in relation to measurement error is acceptable.Additionally, since the strongest correlation between constructs is 0.34 [calculative commitment with work engagement] and all AVE values exceed 0.50, the Fornell-Larcker criterion is adequate and discriminant validity is established at the construct level [39].

Analysis of necessary conditions
Typically, prior to Fuzzy-set Qualitative Comparative Analysis, one assesses whether any of the causal conditions might be deemed necessary.In a necessary condition analysis, the interpretation  of the results is based on consistency.A condition is considered necessary when its consistency value exceeds 0.9 [40].It is good practice to examine the presence of necessary conditions for both the outcome variable and its negation [41].In the case of this study, the resulting variable is the turnover intention, and its negation [Table 2].
The analysis demonstrates that none of the conditions are pointed as necessary.Therefore, we can conclude that none of the analysed conditions will be present in all the combinations of conditions leading to the observed outcome.
Still, the first conclusions can be drawn on the results obtained.The biggest differences in the level of consistency can be noticed for the calculative commitment variable.Among employees with turnover intention, calculative commitment is significantly lower than among those who are satisfied with their workplace.The second finding is related to horizontal trust, which can be seen as quasi-essential for intention to stay in the organisation, as its value is close to 0.9.Thus, lack of horizontal trust can lead to turnover intention.However, the effect may not be significant because high consistency of horizontal trust also comes with the turnover intention.In the case of all five constructs, it is noticeable that higher consistency corresponds to the desire to keep the job, and lower consistency corresponds to the desire to change it.The opposite situation applies to construct's negation.

fsQCA
Fuzzy-set Qualitative Comparative Analysis (fsQCA) is a systematic and rigorous method used to analyse complex, multivariate data in social science research [42].It serves as a bridge between qualitative and quantitative approaches, allowing researchers to study intricate causal relationships that traditional quantitative methods might overlook [43].The fsQCA enables to derive insightful and contextually findings that contribute to a deeper understanding of complex social phenomena.
Technically the method involves transforming interval scale variables into fuzzy sets, where membership scores range between 0 and 1, denoting the degree of association with a particular set.The pivotal 0.5 crossover point signifies the threshold for ambiguity with respect to membership [44].
To perform fsQCA, the calibration of constructs was needed, in case of this study, with recommended breakpoints set at 0.95 and 0.05 for full membership and non-membership, respectively.The truth table obtained aids in the identification of various configurations and their associated causal conditions and outcomes.The configurations are prioritized on their consistency, ensuring that they account for a significant proportion [at least 80%] of the observed cases.The next step involves logical minimization.
Logical minimization in Fuzzy-set Qualitative Comparative Analysis involves simplifying complex sets of conditions and outcomes by identifying the minimal combinations of factors necessary to produce the observed outcome.By reducing the complexity of the configurations, logical minimization aims to uncover the essential causal pathways and conditions that are indispensable for the occurrence of the outcome [45].
Finding sufficient conditions is a crucial goal of fsQCA.Sufficient conditions indicate that a particular set of factors, when present in a particular configuration, is enough to lead to the desired result, and the outcome does not depend on the presence of other factors or conditions.Empirical analysis requires calculating raw coverage, unique coverage, and solution coverage to assess the explanatory power of each configuration [Table 3, Table 4].Raw coverage measures the extent to which each term within the solution explains the outcome's membership, while unique coverage quantifies the proportion of the outcome's membership explained solely by each term.Solution coverage evaluates the overall explanatory power of the complete solution in accounting for the outcome's membership [46].The results of the analysis are included in Table 3.
The solution consistency of .807 is above .75which is the minimum threshold suggested by Fiss [47].According to the analysis (Table 3

DISCUSION
The results help to identify the causes of turnover intention software developers and the variables that are important for avoiding turnover intention in this occupational group.The results obtained indicate that none of the seven causal conditions need to be present or absent for turnover intention to arise.Nevertheless.horizontal trust is identified as a quasi-essential condition for the intention to stay in the organisation.The high consistency value close to 0.9 suggests that the presence of horizontal trust is important for reducing turnover intention among software developers.At the same time, it is worth noting that this factor is not sufficient for staying in the organisation.However, the results suggest that an employee in whose team relationships are not based on trust with colleagues may be more inclined to leave the organisation.This result confirms the importance of trust in software teams in this case, as a factor that also binds them to the organisation [3,15,49].At the same time, this result confirms P5.
The fsQCA results reveal that calculative commitment alone does not stand out as a necessary condition for turnover intention.Instead, the absence of calculative commitment is identified as a factor contributing to turnover intention, emphasizing that the lack of a calculative commitment may increase the likelihood of turnover.Therefore, P3 was supported in this study.
As in the previous case, affective commitment alone is not identified as a necessary condition for the turnover intention.However, when considering different combinations, the absence of HRM and lack of affective commitment for individuals with short tenure emerge as contributing factors to turnover intention.Additionally, it reveals a nuanced relationships where the affective commitment, calculative commitment, and employee engagement collectively Note: G = present, = absent, blank spaces = "don't care" condition.Consistency cut-off: 0.8; Frequency cut-off: 2.00.
contributes to no turnover intention.This suggests that affective commitment, in conjunction with other conditions, may play a role in influencing turnover intention among software developers, particularly those with shorter tenure.This aligns with the notion that software developers' attachment to an organization is more driven by calculative factors than affective ones.This confirms the results of previous studies and indicates that affective commitment are not the dominant commitment type for software developers [4,15].For this group, there is a trend indicating a greater focus on professional and personal goals than on organizational goals, resulting in a decline in commitment to the organization in favour of a larger commitment to the profession [16].This means that commitment changes the addressee to the profession or professional development and this appears to be an inevitable consequence of the changes in an employee's relationship with the organization [48].However, the fact that affective commitment is losing importance in this professional group is consistent with the characteristics of this group [1,2,4].
The analysis suggests that employee engagement, specifically dedication, is not a necessary condition for turnover intention.However, the absence of employee engagement emerges as a contributing factor for turnover intention, indicating that a lack of dedication to work may be associated with an increased likelihood of turnover among software developers.Therefore, P4 was also supported in this research.
Moreover analysis indicate that HRM processes alone do not emerge as a necessary condition for turnover intention.The absence of a consistent pattern suggests that HRM processes may not be a standalone driver of turnover intention among software developers.However, further investigation into the combinations of conditions leading to turnover intention reveals that the absence of HRM, no work engagement, and no horizontal trust as well as the absence of HRM, and no affective commitment collectively contribute to turnover intention.This highlights the complex interplay of various factors, and the role of HRM processes may be contingent on the presence of other conditions.The result suggests that the HRM process, i.e. the belief that HRM contributes to meeting employees' needs, is more important for employees with longer tenure (more than 3 years).The results also suggest that in certain contexts, individuals with longer tenure may exhibit a reduced likelihood of turnover, aligning with the idea that tenure contributes to organizational commitment, thus providing some support for P6 The analysis indicates that gender alone is not identified as a necessary condition for turnover intention.However, specific conditions, such as no employee engagement for women with short tenure and no employee engagement for men with long tenure, are associated with turnover intention.This nuanced finding suggests that gender may interact with other factors, such as employee engagement and tenure, in influencing turnover intention among software developers.
In summary, the results underscore the complexity of turnover intention among software developers, with multiple factors interacting to influence their decision to stay or leave an organization.The findings provide valuable insights for organizations aiming to develop targeted strategies to retain software development professionals.Future research could delve deeper into the nuanced relationships among these factors and explore additional dimensions that may impact turnover intention in this dynamic and competitive field.These results are also in line with the scholars that argued for the need to focus on the turnover intention predictors that reflect the specificity of this occupational group [18].
I am sure that my colleagues wish me well.I am sure I would receive help from my colleagues.If I had an idea that was different from the ideas of others, I could share it in peer-to-peer conversations.My colleagues are happy to share ideas and information with me.Source: own calculations.

Table 2 :
Analysis of necessary conditions

Table 3 :
Analysis of sufficient conditions, outcome: turnover intention.