Measuring Career Adaptability in a Sample of Italian University Students: Psychometric Properties and Relations with the Age, Gender, and STEM/No STEM Courses

Abstract

The continuous transformation of the labor market, characterized by great instability and uncertainty, and by rapid technological changes, has strongly influenced the construction and management of career paths. Nowadays, individuals are faced with careers that are fluid and boundaryless, characterized by discontinuity and a variety of organizations to deal with. In this scenario, the ability to adapt and react to continuous changes in the labor market and in organizations is now a priority for workers. This study presents the psychometric properties of the construct of Career Ability measured through Proactive Personality and Boundaryless Mindset as proxy variables in a sample of 579 adults enrolled at the University of Cagliari (Italy), or recently graduated therein. We aim to rate the factorial structure of the items and to evaluate their multi-group invariance regarding the gender variable. Moreover, the criterion and concurrent validity were assessed. The instrument shows good psychometric characteristics; factorial structure, factorial invariance in relation to the gender variable, concurrent, and criterion validities were confirmed.

1. Introduction

Economic and financial crises, and deep social changes within recent decades, have profoundly changed careers management. The concept of the working life span has certainly changed compared to the past; in fact, if it was previously possible and foreseeable to enter a few working contexts during one’s career (Eby et al. 2003), today individuals are faced with careers that are fluid and boundaryless (Briscoe and Hall 2006; Sullivan and Arthur 2006), characterized by discontinuity and a variety of organizations to deal with. The ability to adapt and react to continuous changes in the labor market and in organizations is now a priority for workers (Carter 2019). Therefore, those who are in a transition phase between higher education and the working world, such as university students, must be able, right from the start, to express this ability (Hamzah et al. 2021), which involves developing the propensity to be flexible, resilient, and take control of their careers (Alisic and Wiese 2020; Jackson and Tomlinson 2020; Zacher 2015). Several studies concur that the individuals who are best able to adapt to new circumstances are also those who can find work and build successful careers (Cortellazzo et al. 2020; Holtschlag et al. 2020; Kundi et al. 2020). The concept of career adaptability was postulated by Super and Kansel (1981) and developed by Savickas (2013), who defines it as a psychosocial resource, and a combination of attitudes, skills, and behaviors that help individuals adapt to the job they have chosen, in which self-knowledge and one’s personal characteristics become essential elements for developing this ability. Concern, control, curiosity, and confidence are the characteristics that help individuals to cope with the changes and unforeseen events that may occur in their professional career (Guan et al. 2016). According to Fugate and colleagues (Fugate et al. 2004), adaptability is defined as the ability that people have to change behavior, feelings, and thoughts with the aim of responding to the changes in the environment in which they are inserted. The individuals with this ability are able to manage ambiguity, uncertainty, are less exposed to stress, and more able to act outside predetermined boundaries (O’Connell et al. 2008). Some authors point out that proactive personality is an important component of adaptability (Cai et al. 2015; Fawehinmi and Yahya 2018; Hameed et al. 2020; Hou et al. 2014; Jiang 2017; Öncel 2014; Savickas and Porfeli 2012; Tolentino et al. 2014) together with boundaryless mindset (Chan et al. 2015; Stauffer et al. 2019).

Considering the potential importance of Career Adaptability measured through Proactive Personality and Boundaryless Mindset as proxy variables (McArdle et al. 2007), the purpose of this paper was to explore the psychometric features of the Italian version of the instrument in a sample of young adults, not yet employed and looking for professional integration, in the context of recent developments in the international labor market. Therefore, the study aimed to assess the factorial structure of the items and to evaluate their multi-group invariance regarding the gender variable (Byrne 2001, 2004, 2008, 2013). Moreover, the concurrent validity and criterion validity were assessed.

2. Literature Review

In recent years, studies have paid great attention to the construct of career adaptability, almost prefiguring the new scenarios triggered by the COVID-19 pandemic (Chen et al. 2020b). For students facing the transition phase, in order to enter work, learning to adapt in an ever-changing environment is recognized as one of the most important skills and one of the main factors related to job success (Rudolph et al. 2017; Spurk et al. 2013) in job search strategies (Koen et al. 2010), employability (Spurk et al. 2016) and performance (Zacher 2015). As Tolentino and colleagues (Tolentino et al. 2014), and Ginevra and colleagues (2018) suggest, adaptability allows individuals to face unexpected problems that are not easy to solve, enabling them to achieve a functional balance with the environment in which they are inserted. In a recent literature review, Chen and colleagues (Chen et al. 2020b) found that this specific field of study basically encompasses five main clusters (described in their study by the high frequency keyword grouping characteristics): boundaryless mindset, career adaptability scale, career construction, proactive personality, and life design. Despite this, the same review (Chen et al. 2020b) shows that the most important measurement scales on career adaptability have focused on different psychological dimensions (see for example: The Career Adapt Ability Scale (CAAS), by Savickas and Porfeli (2012); The Career Adapt abilities Scale—Short Form (CAAS-SF) by Maggiori et al. (2017); The Career Futures Inventory—Revised (CFI-R) Rottinghaus et al. (2012)), pointing out that researchers have developed different scales according to their own research. While the constructs of the boundaryless mindset and the proactive personality are among the most common in theoretical studies of career adaptability, none of the aforementioned scales uses them. Only one study (McArdle et al. 2007), to the best of the authors’ knowledge, used proactive personality and boundaryless mindset as proxy measures for adaptability.

The aim of the aforementioned study was to empirically test the employability model of Fugate and colleagues (Fugate et al. 2004) through three separate but inter-connected psycho-social dimensions: adaptability, career identity, and human and social capital. In that study, adaptability was measured through the proactive personality and Boundaryless mindset, as operationalized by Bateman and Crant (1993), and Briscoe and colleagues (Briscoe et al. 2006). The adaptability variables (i.e., boundaryless mindset and proactive personality), seem to be those that, in this time of great environmental turbulence and important change in the contract forms between workers and organizations, more fit with a no longer traditional career concept that allows people to operate in a VUCA environment (Pryor et al. 2008). The ability to cope with “disordered, complex, and unwritten problems” is now considered an essential learning outcome of higher education (Nelson Laird et al. 2009). Workers need to be comfortable with ambiguity; evaluate the quality of the available data; keep a clear view of the larger image; identify options when blocked, challenged or rejected; scan media broadly and efficiently; accurately identify the central issues in a conflict; and challenge conventional methods, systems, and thinking (Feller and O’Bruba 2009; Shaffer and Zalewski 2011). Traditional career theory was a useful way to help students make career decisions by narrowing, channeling, or focusing their attention on potential career choices, often following a strategy of matching students’ skills, attitudes, values, or personality traits with occupations or specific professions (Van Vianen et al. 2009). Today, however, in post-industrial economies, careers are often fragmented and protean, requiring workers to be willing as well as being able to stay within a flow of continuous growth and experimentation (Parker 2008; Van Vianen et al. 2009).

3. Components of Career Adaptability

3.1. Proactive Personality

Different studies link career adaptability to proactive personality (Cai et al. 2015; Fawehinmi and Yahya 2018; Green et al. 2019; Hameed et al. 2020; Hou et al. 2014; Jiang 2017; McArdle et al. 2007; Öncel 2014; Tolentino et al. 2014; Uy et al. 2015). The proactive personality can be defined as a stable disposition that makes individuals able to actively intervene in the environment, and more independent of situational factors and always looking for new opportunities for improvement and learning (Seibert et al. 1999; Wang and Wanberg 2017). Proactive personality finds its theoretical foundations in interactionism (Bowers 1973) and in social cognitive theory (Bandura 1986), for which there is a constant reciprocal interaction between the person, the environment, and behavior. The construct, which first appeared in a study by Bateman and Crant (Bateman and Crant 1993), was used in contrast to the concept of passive personality. People with proactive personalities are relatively independent of their surroundings; they are able to seize opportunities that present themselves and persevere in their actions until the result is achieved. Subsequently, other authors have defined individuals with proactive personalities as future-oriented, creative, self-sufficient, able to constantly improve, and have an optimistic orientation towards change, as well as being able to profitably enter the labor market (Brown et al. 2006; Campbell 2000; Fuller and Marler 2009; Greguras and Diefendorff 2010; Kim et al. 2009; Parker et al. 2006; Spitzmuller et al. 2015).

Still, other studies have linked proactive personality with a series of positive outcomes. The results showed how proactive personality affects learning ability and strengthens social capital and future orientation, self-directed career management, successful job searches, the results of the socialization process of newbies in organizations, career success, job satisfaction, well-being, and the effectiveness of the leadership process (Ashford and Black 1996; Bateman and Crant 1993; Briscoe et al. 2006; Crant 2000; de Guzman and Choi 2013; Frese et al. 2007; Ryff and Singer 2008; Seibert et al. 2001; Soresi et al. 2012; Spitzmuller et al. 2015; Valls et al. 2020; Uy et al. 2015; Zheng et al. 2020). Furthermore, being proactive supports a boundaryless career mindset (Briscoe et al. 2006; Jackson 1996; Lochab and Nath 2020; Mirvis and Hall 1996), and has a strong influence on the perception of career self-efficacy (Brown et al. 2006; Frese and Fay 2001; Kim and Park 2017). Bocciardi and colleagues (Bocciardi et al. 2017), for example, link the proactive personality to a high ability to make optimal career decisions and an overall improvement in professional life (Parker et al. 2006). Jiang (2017), Hameed, and colleagues (Hameed et al. 2020), and Öncel (2014), argue that proactive personality is able to influence the development of career adaptability in adult workers, while other authors (Cai et al. 2015; Hou et al. 2014; Tolentino et al. 2014) reach the same results in samples composed of university students who identify proactive personality as an antecedent of career adaptability. Rudolph and colleagues (Rudolph et al. 2017), analyzing 90 studies in their meta-analysis, identified proactive personality as one of the measures of adaptivity most associated with adaptability.

3.2. Boundaryless Mindset

In literature, another variable associated with career adaptability and proactive personality is the boundaryless mindset (Arthur 1994; Arthur and Rousseau 1996; Li et al. 2021; Spurk et al. 2013). Individuals with a boundaryless mindset conceive of their career as barrier-free and prefer organizations without traditional borders; they pursue career opportunities and relationships beyond a single employer and favor situations in which it is possible to physically move between jobs, professions, and different geographic areas (Sullivan and Arthur 2006). Hence, it is preferable to a career vision characterized by the lack of physical and psychological boundaries (Cheramie et al. 2007), and appears to be composed of two dimensions: physical mobility, more focused on the changeability of jobs and employers’ work (Gubler et al. 2014; Sullivan and Baruch 2009), and the psychological one, more oriented to maintaining relationships beyond organizational boundaries that do not necessarily involve physical mobility (Briscoe and Finkelstein 2009; Rodrigues and Guest 2010; Verbruggen 2012). Among the factors that contribute most to the boundaryless mindset are proactive personality, protean career, employability, and self-efficacy in job search (Brown et al. 2006; Fugate et al. 2004; Wiernik and Kostal 2019). Zhao and colleagues (Zhao et al. 2016) found that the boundaryless mindset is related to work performance abroad through the mediation of proactive resources, while Kundi and colleagues (Kundi et al. 2020) have shown that the boundaryless mindset is a predictor of career success. Subsequently, Li and colleagues (Li et al. 2021), carrying out a meta-analysis on contemporary career attitudes, verified the existence of a positive correlation between the boundaryless mindset and intrinsic and extrinsic career success, job satisfaction, and proactive personality. In their study, these authors also assert that, if on the one hand workers with a boundaryless mindset are highly appreciated by organizations, on the other hand those who are oriented towards larger networks may not be well regarded by companies, as they are more inclined to turnover. This generates the need to deepen the studies on contemporary career attitudes not only from the point of view of how beneficial these attitudes are for the individual career (Hall et al. 2018), but also how convenient or harmful they can be for organizations. Another criticism comes from Pringle and Mallon (2003) who underline that boundaryless career theory does not sufficiently take into account aspects such as social structures, the national context, gender, and ethnicity.

4. Proactivity Personality, Boundaryless Mindset and Demographic Characteristics

Several studies have investigated proactive personality and the boundaryless mindset from a gender and age perspective. These two constructs are strongly related to adaptability (Green et al. 2019; Hameed et al. 2020; Li et al. 2021; McArdle et al. 2007; Spurk et al. 2013), so it is important to understand whether being more independent from situational factors (Wang and Wanberg 2017) and having a career vision characterized by the lack of physical and psychological boundaries (Arthur and Rousseau 1996) can account for the current gap between men and women, or between young and older employees in terms of adaptability and profitable entry into the labor market.

4.1. Proactive Personality with Gender and Age

Spitzmuller and colleagues (Spitzmuller et al. 2015), in their meta-analytic review, showed no significant gender differences in the proactive personality, although the literature suggests that women have a lower power perception than men (Keltner et al. 2003), and ts it would make them less inclined to engage in proactive behaviors, such as challenging the status quo and overcoming obstacles to making meaningful change for women (Bateman and Crant 1993). Likewise, no significant difference by gender was observed in the studies of Travis and Freeman (2017) on a sample of university students, of Özkurt and Berkan (2018) concerning high school students, of Zhang and colleagues (Zhang et al. 2020) on a sample of graduating university students, who were moving from university to work, and of Hua and colleagues (Hua et al. 2020) in the analysis of an international students sample. At the same time, De Pater and colleagues (De Pater et al. 2009) highlighted how women have a lower proactive personality than male colleagues in research that considered a sample of interns.

McArdle and colleagues (McArdle et al. 2007) and Zhang and colleagues (Zhang et al. 2021) found no age or gender differences in terms of proactive personality in their respective research on Australian unemployment, on high-tech company employees, as well as Wang and colleagues (Wang and Wanberg 2017) in a sample of employee-supervisor dyads, and Thomas and colleagues (Thomas et al. 2010) in their comparative meta-analysis. Regarding age, some research has shown that older and younger workers do not differ in their levels of proactivity (Bertolino et al. 2011; Erdogan and Bauer 2005; Harvey et al. 2006; Seibert et al. 2001), while other studies (Truxillo et al. 2012) suggest that older workers would be perceived as lower than younger workers in terms of their proactive personality.

4.2. Boundaryless Mindset with Gender, Age and STEM/No STEM Degrees

Guan and colleagues (Guan et al. 2019), in their literature review, highlight how numerous individual moderators have been taken into consideration regarding the boundaryless mindset and, among them, age and gender, which would create further difficulties and barriers to career promotion (Forrier et al. 2009). Besides, different studies had already shown that one of the limits of the boundaryless mindset theory was just to not give due importance to such measures (Eby et al. 2003; Enache et al. 2011; Pringle and Mallon 2003).

Some studies have pointed out that, in general, women seem to have a higher psychological mobility, while men would have higher physical mobility and would be more motivated by economic factors, while women seem more oriented towards relational careers and more driven by the variety in work (Inceoglu et al. 2009; Mainiero and Sullivan 2005; Sullivan and Arthur 2006; Warr 2008). Segers and colleagues (Segers et al. 2008), investigating a large sample of European workers, not only confirmed the previous statements but found a significant difference in terms of age with respect to physical mobility, which sees the latter decrease with increasing age. Conversely, the Bednarska-Wnuk study (Bednarska-Wnuk 2020) found no statistically significant differences between men and women regarding boundaryless mindset in a sample of workers.

Even the Kostal and Wiernik studies (Kostal and Wiernik 2017), and Abid and colleagues (Abid et al. 2021), point in this direction, defining demographic differences in reference to boundaryless mindset negligible. For Briscoe and colleagues (Briscoe et al. 2006), there is no difference in gender, but a slight and significant positive relationship between age and boundaryless mindset was demonstrated in a sample made up of undergraduate business students, graduate students, and managers. In another study conducted on a sample of European Information Technology professionals (Gubler et al. 2014), younger respondents indicated a significantly higher preference for mobility than older ones, and the rejection of career opportunities were positively correlated to age.

Finally, different authors have taken into consideration the differences between STEM (Science, Technology, Engineering, and Maths) and NON STEM degree courses in reference to the development of soft skills, which can help facilitate the transition from university to the labor market (Dika and D’Amico 2016; Whalen and Shelley 2010; Wilson 2010; Xu 2013). Some research highlights that in STEM courses, employability skills are often underestimated and undertrained in educational paths (Rayner and Papakonstantinou 2016). What has been highlighted in these studies is that non-STEM disciplines programs, such as business studies, involve developing employability skills, while STEM programs, while successfully preparing STEM graduates with the academic knowledge necessary for the workplace, do not invest in soft skills development, and these students often lack interpersonal and transferable skills, practical experience, general workplace experience, or required business knowledge (Prinsley and Baranyai 2015).

According to several studies, training that implements not only specialized technical knowledge but also soft skills, such as adaptability and proactivity, would support job placement (McGunagle and Zizka 2018, 2020; Hartmann and Jahren 2015; Rayner and Papakonstantinou 2015) and would make future workers better able to find creative solutions to demands from workplace (Dyke-Ford and Teare 2006; Hartmann and Jahren 2015; Maxwell et al. 2010). Comparisons between different degree courses could help to differentiate the guidance programs in universities and to broaden knowledge relating to the factors facilitating placement.

5. Materials and Methods

5.1. Participants

The assessment was conducted on 579 adults enrolled at the University of Cagliari (Italy), or recently graduated therein, and 53.9% were women. The average age of the participants was 26 years (ranging from 19 to 60 years; standard deviation = 6.51). The 43% of participants attended STEM degrees. The average mark of their exams was 23.4 (sd = 6.8).

Participants were divided randomly in two groups in order to carry out two studies: a first study with explorative approach (n = 250), and a second with a confirmatory approach (n = 329; the second sub-sample had to be more numerous than the first, in order to have an adequate size to apply the confirmatory analyses required for the assessment of factorial invariance).

The sampling was nonprobabilistic. Participants, after receiving an invitation via their institutional university email, volunteered to participate in an online survey by self-completing a questionnaire administered in November 2019. Participation in the study was on a voluntary basis, and the data collected were anonymous and confidential. The participants were informed of details concerning the aim of the data collection, and they gave their consent to the data treatment. The protocol was administrated through the Google Drive forms, using the lists of students enrolled in the CareerDay of the University of Cagliari. All ethical guidelines were applied, following the procedures defined by the institutional research committee, by the American Psychological Association (APA), by the Italian Association of Psychology (AIP), and by the 1964 Helsinki declaration (with their and subsequent amendments). The descriptive statistics for all variables and subsamples are showed in

Table 1. Descriptive statistics.

Variable	Study 1 (n = 250)	Study 2 (n = 329)
Percentage of women	53.20%	52.58%
Age—Mean (sd)	26.260 (6.659)	26.049 (6.756)
Average of mark—Mean (sd)	23.680 (6.511)	23.138 (7.277)
Type of degree- STEM (%)	42.6%	43.5%
Dimensions Inquired
Proactive Personality—Mean (sd)	5.194 (0.715)	5.204 (0.820)
Boundaryless mindset—Mean (sd)	3.879 (0.561)	3.918 (0.551)

.

5.2. Instruments and Procedure

The items of the original scales were translated by the application of translation/back-translation process; these items were previously evaluated by a group of experts and a group of students in order to evaluate their comprehensibility. The final Italian version of the instrument is showed in the Appendix A.

The Proactive Personality (PP) scale was developed by Bateman and Crant (1993), and in this study we used the version of Seibert and colleagues (Seibert et al. 1999), validated in Italy by Trifiletti and colleagues (Trifiletti et al. 2009). This scale has been translated and adapted in many countries, always showing good psychometric properties [e.g., French version by Carrein (2011); Spanish version by Valor-Segura et al. (2020); Chinese version by Shang and Gan (2009); Korean version by Kim and Park (2017); Indian version by Lochab and Nath (2020); Turkish validation by Öncel (2014)]. The scale was characterized by 10 items and rated on a Likert scale from one (strongly disagree) to seven (strongly agree). One element was, for example: “If I see something I don not like, I will fix it.” The alpha reliability in our sample was 0.799. The Boundaryless Mindset Scale (BLM) was developed by Briscoe and colleagues (Briscoe et al. 2006) and validated in Italy by Lo Presti and colleagues (Lo Presti et al. 2011) as one of the factors of the Boundaryless Career Attitude Scale (Briscoe et al. 2006). This tool has been validated in many countries [for example, in Spain by Lochab and Nath (2020); in France by Stauffer et al. (2019); in Turkey by Çakmak-Otluoğlu and Bolat (2020); in the Philippines by Bernardo and Salanga (2019)]. The scale is defined by 8 items, rated on a 5-step Likert scale (from one—strongly agree—to five—strongly disagree). This is an example of a scale item: “I am looking for work assignments that allow me to learn something new.” The alpha reliability obtained in our sample was 0.851.

5.3. Data Analysis

Statistical data analyses were applied following a multi-stage approach. Specifically, participants were separated in two comparable subsamples.

In the initial phase, a Principal Component Analysis was applied in the first subsample.

Then, a Confirmatory Factor Analysis was applied for males and females separately; the model hypothesized two correlated factors.

Formerly, the factorial invariance was assessed through the Multigroup Confirmatory Factor Analysis (Byrne 2008; Hirschfeld and Von Brachel 2014), which allows one to simultaneously evaluate the data from different groups. This aim is accomplished by constraining some parameters to assume the equal values in the samples. By means of this technique, measurement invariance can be measured at different levels. A configural invariance specifies that the number of latent constructs and the patterns of factor loadings are comparable in the two groups (configural invariance). A weak invariance suggests that the earlier conditions are fulfilled and that we have metric invariance (i.e., the magnitude of factor loading is similar in two groups) (metric invariance). A strong invariance happens when the earlier requirements are observed, and scalar invariance is reached (i.e., the items’ intercepts are analogous across the groups) (scalar invariance). A strict invariance is reached once, over and above the previous conditions, residual variances are similar across the groups (strict invariance) (Hirschfeld and Von Brachel 2014). For each formerly cited model, the parameters were forced to be equal across the groups; the fit of each model was associated with the strong measurement invariance. So as to choose on the invariance of measurement, variations in the fit indices were detected; specifically the ∆CFI was considered (∆CFI < 0.01 is the recommended cut-off point to determine if a further constrained model indicates a substantial reduction in model fit regarding a fewer constrained model) (Chen 2007).

The Confirmatory Factor Analyses (estimator ML) were applied by the software R 4.0.2 (R Core Team 2021); precisely, the packages lavaan, semTools, and semPlot were used (Epskamp 2017; Jorgensen et al. 2016; Rosseel 2017). Furthermore, concurrent validity was evaluated by the computation of Pearson’s r coefficients. Specifically, we referred to concurrent validity as an assessment of the consistency of the measure with an immediately observable behaviour or event (e.g., in our case, classically, with the average mark of examinations carried out at the university) (Chiorri 2020; Cronbach and Meehl 1955; Guion and Cranny 1982).

Finally, the criterion validity was assessed by the application of a Multivariate Analysis of Covariance, using the age as covariate, the degree, and the gender as between factors (Chiorri 2020).

6. Results

6.1. Study 1

The data collected were assessed in order to evaluate the normality of distributions, and computing the indices of skewness (ranging from −1.248 to −0.089) and kurtosis (ranging from −0.813 to 0.861). Furthermore, a priori, we determine the minimum sample size required to achieve an acceptable level of statistical power for the factor structure under evaluation (Thompson 2004; Cohen 2013). Moreover, we verify that the data met the requirements for the application of the PCA (Kaiser-Meyer-Olkin index = 0.838) and the CFA.

We applied the Principal Component Analysis on the items, with Promax rotation (

Table 2. PCA, Component Loadings.

Items		PC1	PC2	Uniqueness
I am constantly on the lookout for new ways to improve my life.	PP1		0.353	0.782
Wherever I have been, I have been a powerful force for constructive change.	PP2		0.581	0.622
Nothing is more exciting than seeing my ideas turn into reality.	PP3		0.452	0.806
If I see something I don’t like, I fix it.	PP4		0.687	0.569
No matter what the odds, if I believe in something I will make it happen.	PP5		0.752	0.496
I love being a champion for my ideas, even against others’ opposition.	PP6		0.474	0.817
I excel at identifying opportunities.	PP7		0.673	0.534
I am always looking for better ways to do things.	PP8		0.570	0.598
If I believe in an idea, no obstacle will prevent me from making it happen.	PP9		0.744	0.499
I can spot a good opportunity long before others can.	PP10		0.692	0.468
I seek job assignments that allow me to learn something new.	BLM1	0.526		0.678
I would enjoy working on projects with people across many organizations.	BLM2	0.741		0.431
I enjoy job assignments that require me to work outside of the organization.	BLM3	0.717		0.503
I like tasks at work that require me to work beyond my own department.	BLM4	0.732		0.485
I enjoy working with people outside of my organization.	BLM5	0.829		0.368
I enjoy jobs that require me to interact with people in many different organizations.	BLM6	0.784		0.430
I have sought opportunities in the past that allow me to work outside the organization.	BLM7	0.627		0.615
I am energized in new experiences and situations.	BLM8	0.515		0.608

Note. Applied rotation method is Promax; PC = Principal component.

). In this analysis were observed two principal components (PC1—Scale PP, Eigenvalue = 5.421; Proportion of variance explained = 0.301; PC2—Scale BLM, Eigenvalue = 2.269—Proportion of variance explained = 0.126); these principal components showed a correlation of 0.429. The internal consistency (Cronbach’s Alpha) of the two dimensions is satisfactory (F1 α = 0.799; F2 α = 0.851).

6.2. Study 2

In the second subsample of participants, the confirmatory factor analyses (CFAs—in which the parameters were estimated by maximum likelihood method) were carried out. These analyses were applied three times in relation to the total subsample of the study, and furthermore (regarding the assessment of the factorial invariance in relation to the gender variable), separately for males and females.

We inspected the data fit regarding to two correlated factors. With the intention of evaluating the models, several classical indices were computed: the ratio of Chi Square and the degrees of freedom, that it is demarcated as acceptable if it is under three (Schermelleh-Engel et al. 2003; Wheaton et al. 1977); the Comparative Fit Index—CFI (Bentler 1990), that is considered acceptable if higher than 0.90 (Byrne 2001); the indices Root Mean Square Error of Approximation (RMSEA) (Steiger 1990) and Standardized Root Mean Square Residual (SRMR), that were considered adequate if lesser than 0.08 (Hu and Bentler 1999).

The CFAs presented good fit indices (

Table 3. Results of the application of CFAs in the study 2.

	Factor Loadings	df	Chi Square	Chi Square/df	p	RMSEA	RMSEA [90% CI]	SRMR	CFI	GFI
Total sample	From 0.33 to 1.01	134	254.154	1.896	<0.0001	0.053	0.043–0.063	0.074	0.967	0.962
Males	From 0.39 to 0.98	134	168.922	1.260	0.022	0.042	0.017–0.060	0.088	0.978	0.946
Female	From 0.40 to 0.82	134	135.425	1.010	0.449	0.008	0.000–0.039	0.073	0.999	0.964

Note: df = degrees of freedom; RMSEA (90% CI) = Root Mean Square Error of Approximation with Confidence Interval; SRMR = Standardized Root Mean Square Residual; CFI = Comparative Fit Index; GFI = Goodness of Fit Index.

), supporting the factor structure of the instrument.

Then, we carried out multigroup confirmatory factor analysis. This approach consents the measurement invariance of the scale to be measured across groups of participants (males and females) who are likely to have the similar levels of the latent constructs (Byrne 2008).

Table 4. Measurement invariance models regarding the two groups (males/females).

	df	Chi Square	p	Chi Square ∆	Df ∆	CFI	RMSEA	CFI ∆	RMSEA ∆
Model 1: configural	268	304.347	0.063			0.990	0.029
Model 2: metric	284	330.684	0.029	26.337	16	0.987	0.032	0.003	0.003
Model 3: scalar	300	340.417	0.054	9.733	16	0.989	0.029	0.002	0.003
Model 4: strict	318	355.659	0.072	15.242	18	0.990	0.028	0.001	0.001

Note: df = degrees of freedom; RMSEA (90% CI) = Root Mean Square Error of Approximation with Confidence Interval; CFI = Comparative Fit Index; ∆ = differences in fit indices between the unconstrained baseline model and the stronger constrained models.

shows the fit indices of the nested CFA carried out, from configural to strict invariance models.

The fits are good (Cheung and Rensvold 2002); the differences in fit indices amongst the unrestricted baseline model and the robust constrained models highlight for our data a strict factorial invariance, confirming the validity and usefulness of this assessment instrument, and furthermore the invariance regarding the gender variable. Specifically, this evidence of invariance implies that the males and females conceive of the constructs investigated in the same way.

Moreover, linear correlations (Pearson’s r) (

Table 5. Pearson’s r correlations.

		Age	Average Grade on University Exams	PP
average grade on university exams	r	0.006	1
	p	0.889
PP	r	0.025	−0.100 *	1
	p	0.550	0.016
BLM	r	0.125 **	−0.044	0.386 **
	p	0.003	0.299	0.001

Note = ** p < 0.01; * p < 0.05 (2 tailed); PP = Proactive Personality; BLM = Boundaryless mindset; r = Pearson’s r correlation.

) were computed between the assessed variables, the age, and the average grade achieved on university exams. These coefficients highlighted specifically that BLM have a significant positive correlation with age and PP; the average grade in exams show a negative significant correlation with PP.

To evaluate potential differences regarding the dimensions PP and BLM in relation to the socio-demographic variables, a Multivariate analysis of covariance (MANOVA) were carried out, having as covariate the age, as between factors gender and type of degree (STEM—NO STEM).

The analysis underlined a multivariate effect of the covariate age (Wilk’s Lambda _(2,565) = 0.981; p = 0.005; partial Eta Squared = 0.019), a multivariate main effect of the variable degree (Wilk’s Lambda _(2,565) = 0.972; p = 0.0001; partial Eta Squared = 0.028), and a multivariate main effect of gender (Wilk’s Lambda _(2,565) = 0.982; p = 0.006; partial Eta Squared = 0.018). The interaction between degree * gender is not significant (Wilk’s Lambda _(2,565)= 0.998; p = 0.519).

At the univariate level, the effect of the covariate age is significant for the BLM (F = 2.702; p = 0.003; partial Eta Squared = 0.016). Always at the univariate level, the main effect of the variable degree is significant for the scale PP (F(1,566) = 10.443; p = 0.001; partial Eta Squared = 0.018). Specifically, the students in the NO STEM degrees have higher scores than STEM students in this scale (STEM mean = 5.075, standard deviation = 0.747; NO STEM mean = 5.269; standard deviation = 0.789).

Furthermore, at the univariate level, the variable gender show a significant effect regarding the BLM (F(1,566) = 4.786; p = 0.029; partial Eta Squared = 0.008). Indeed, in this scale, females have higher scores (mean = 3.940; standard deviation = 0.525) than males (mean = 3.850; standard deviation = 0.572).

7. Discussion and Conclusions

This study explored the career adaptability construct, defined as the ability that people have to change behavior, feelings, and thoughts with the aim of responding to the changes in the environment in which they are inserted (Fugate et al. 2004). According to McArdle and colleagues (McArdle et al. 2007), career adaptability can be measured through proxy the two variables: Proactive Personality (Cai et al. 2015; Fawehinmi and Yahya 2018; Hameed et al. 2020; Hou et al. 2014; Jiang 2017; Öncel 2014; Savickas and Porfeli 2012; Tolentino et al. 2014), and Boundaryless Mindset (Chan et al. 2015; Stauffer et al. 2019).

Validating a scale capable of assessing career adaptability in the Italian context would allow us to explore the ability of individuals to adapt to different contexts and to the continuous changes they encounter during their life span (Carter 2019).

For these reasons, the purpose of this article was to explore the psychometric properties of the Italian version of the Career Adaptability using the Proactive Personality and the Boundaryless Mindset as proxy measures (McArdle et al. 2007).

A multi-stage approach was used for the data analysis and the participants were divided into two sub-samples. The first subsample was used for the analysis of the principal components with Promax rotation; as hypothesized, the presence of two main components, PP and BLM, with a satisfactory internal consistency emerged.

The second subsample was used to perform a Confirmatory Factor Analysis for males and females separately. This approach allows measuring the measurement invariance of the scale between groups of participants (male and female), who might show similar levels of latent constructs (Byrne 2008); specifically, the results confirm a strict factorial invariance. The scale highlighted good validity and usefulness, furthermore, the invariance with respect to the detected gender variable confirms that males and females conceive the constructs investigated in the same way. Subsequently a Multivariate analysis of covariance (MANOVA) was carried out in order to assess the existence of potential differences in the scores of PP and BLM. Assuming age as a covariate variable, the gender and type of degree (STEM—NO STEM) were taken into consideration. The analysis highlighted a multivariate main effect of the “degree course type” variable and a multivariate main effect of the gender. The main effect of the “degree course type” variable is significant for the PP scale, in which it is observed that NO STEM degree students are more proactive than STEM students.

This results is in line with previous researches, which highlight that employability skills continue to be a crucial gap in STEM paths (Harvey 2005; Maxwell and Armellini 2019; Maxwell et al. 2010). In academia, there probably still persists the idea that in some paths, the hard knowledge of a specialized technical type is the only one necessary to enter the working world. At the same time, there is a misperception that these skills can or should be learned on the job; if this were true, they would be called occupational skills or employee skills, not occupational skills (McGunagle and Zizka 2020).

Instead, other studies have highlighted the importance of good levels of proactivity in STEM degree students. For example, Major and colleagues (Major et al. 2012) found that students with a proactive personality demonstrated a greater commitment than students with a low proactive personality in STEM study paths; according to other authors, individuals with proactive personalities are well suited to modern career paths, such as those within STEM degrees (Fuller and Marler 2009) as they are more likely to see challenges as learning opportunities (Elliot and Harackiewicz 1996) characteristics that can be useful in the rapidly evolving technology fields.

Regarding gender, the results on career adaptability are conflicting and often considered inconclusive (Patton and Lokan 2001; Patton et al. 2004). In studies on student samples, females showed higher levels of career adaptability than males (Çizel 2018; Hartung et al. 2005), and more specifically regarding personal and emotional adaptability (Chen et al. 2020a). On the contrary, other studies, always on student samples, did not show gender differences (Cheung and Jin 2016; Ghosh and Fouad 2017; Hirschi et al. 2015; Koen et al. 2012; Rottinghaus et al. 2005; Tien et al. 2014; Zacher 2014). In the present work, the gender variable shows a significant effect with respect to BLM. Females appear to have higher levels of BLM and this finding contrasts with previous research that defines gender differences in relation to the construct as negligible (Abid et al. 2021; Bednarska-Wnuk 2020; Briscoe et al. 2006; Kostal and Wiernik 2017), although in line with others studies that attribute greater psychological mobility and motivation to the female gender (Inceoglu et al. 2009; Mainiero and Sullivan 2005; Segers et al. 2008; Sullivan and Arthur 2006; Warr 2008). These latter findings suggest the opportunity to provide ad hoc counseling programs for males and females students in order to overcome the difficulties linked in the first case to a low propensity to mobility, and in the second case to greater attention to relationships and to the change of intra-organizational role.

In conclusion, instrument shows good psychometric characteristics; factorial structure, factorial invariance in relation to the gender variable, concurrent and criterion validities were confirmed.

8. Practical Implication

Proactive individuals who perceive their career as unfettered by predetermined boundaries are more likely to be employable. University students are especially required to express this ability in the transition phase to the working world (Alisic and Wiese 2020; Hamzah et al. 2021; Jackson and Tomlinson 2020), considering that the success of this transition is linked to adaptability (Cortellazzo et al. 2020; Holtschlag et al. 2020; Kundi et al. 2020). An important aspect to consider is that many programs aimed at supporting unemployed job seekers have not proved effective (Nielsen Arendt et al. 2020; Card et al. 2018). It would be advisable to intervene before facing the transition phase to the labor market, especially during the university path, in order to foster an attitude to engage in adaptive behaviors before career transitions can serve as preparation and favor the success and quality of the work (Koen et al. 2010; Hirschi 2010; McGunagle 2016; McGunagle and Zizka 2018).

The scale validated in the present research would allow intercepting individuals who do not have good levels of career adaptability and who may have difficulty in entering the job market profitably. By having available data on the presence of a low proactive personality and low levels of boundaryless mindset, it would be possible to build guidance interventions capable of preventing critical issues and increasing the probability of success. Previous studies suggest that there are several psychosocial areas on which it could be possible to intervene to improve personal adaptability: for example self-efficacy (Blackmore et al. 2021; Dillahunt and Hsiao 2021; Guan et al. 2013), self-esteem (Amarnani et al. 2018; Cai et al. 2015); team work skills (de Guzman and Choi 2013; Ebenehi et al. 2016; Fawehinmi and Yahya 2018), learning goal orientation and, past and future temporal focus (Tolentino et al. 2014; Zacher 2015), resilience (Buyukgoze-Kavas 2016) problem solving and decision making (Kozlowski et al. 2001; Coetzee et al. 2015).

These interventions could be carried out both at the individual level with targeted counseling sessions, both at the group level with training moments, and at the organizational level through the implementation of professional orientation policies. Equally important are the accompanying paths to internships and post-graduate internships as a fundamental moment of encounter between the knowledge acquired in the study courses, life skills, and the necessary skills to achieve future optimal performance.

9. Limitations of the Study

Some limitations of this study must be considered. Primarily, the sampling was non probabilistic; it would be useful for the ongoing research to apply a stratified probabilistic sampling, to balance some relevant socio-demographic variables (as the geographical area). Additional, given the cross-sectional nature of this research, it is difficult to precisely assess the direction for the associations observed in our sample; it might be interesting to organize longitudinal studies to provide further suggestions about these relationships. Also, the assessment of further psychological dimensions might be useful to define specificities of Proactive Personality and Boundaryless mindset.