Relationship closeness, habit and perceived subjective norms as predictors of volunteer role identity in school-based mentoring programmes

Abstract

Purpose

This study aimed to examine factors associated with volunteer role identity in mentors of school-based mentoring programmes.

Design/methodology/approach

Drawing on established theoretical models of volunteerism (the Role Identity Model), and research and theory on mentoring programmes, an integrated model of predictors of mentor volunteer role identity was tested. Seventy-one mentors (63 females, mean age 36 years) completed a survey with measures of habit, subjective norms, satisfaction with the mentor-mentee relationship, relationship closeness, social skills and mentor role identity. Path analysis was used for data analysis.

Findings

Fit indexes revealed an acceptable fit to the data. There were six significant paths. Habit and subjective norms were directly related to role identity. The association between mentor role identity and two further predictors, satisfaction with the mentor–mentee relationship and social skills was respectively fully and partially mediated by relationship closeness.

Practical implications

Findings can inform mentoring programmes in supporting mentors to develop a close relationship with their mentees and promote the development of a role identity as a volunteer among mentors. A stronger role identity is in turn expected to enhance mentor retention in the programme.

Originality/value

An important and novel finding of this study is that relationship closeness contributes to mentors developing a volunteer role identity. Also, for the first time, the importance for mentors of support from significant others in fostering sustained volunteer engagement has been examined.

Citation

Pozzi, M., Marzana, D., Marta, E., Vecina, M.L. and Aresi, G. (2021), "Relationship closeness, habit and perceived subjective norms as predictors of volunteer role identity in school-based mentoring programmes", International Journal of Mentoring and Coaching in Education, Vol. 10 No. 1, pp. 17-30. https://doi.org/10.1108/IJMCE-05-2020-0019

Introduction

Supportive, intergenerational relationships with non-parent adults have been recognised as key resources for the healthy development of children (Scales et al., 2006). However, because of changes occurring in families and societal norms, such relationships may be less readily available for children than in the past (Jekielek et al., 2002; Rhodes et al., 2005). Mentoring facilitates the establishment of a relationship between an older or more experienced non-parental individual and younger or less experienced mentee. The mentor provides guidance and support to the mentee over time (Kanchewa et al., 2017). From a positive youth development perspective, mentoring is a way to address the issue of the lack of supportive adults in the lives of many children (Lerner et al., 2014).

An increasing number of formal mentoring programmes across the world offer young people access to supportive non-parental adults (MENTOR, 2015). Programmes generally focus on children or adolescents at risk of poor academic, behavioural or health outcomes. Typically, mentors are trained volunteer adults who are matched with mentees by the programme through a variety of criteria (e.g. socio-demographic characteristics, geographical proximity, common interests). Mentors and mentees meet periodically at school or in other public places. The quality of the mentor–mentee relationship is considered a key active ingredient of mentoring programmes. (Rhodes, 2005; Rhodes and DuBois, 2008) theorised that the mentoring relationship contributes to children's socio-emotional, cognitive and identity development through role modelling and the provision of emotional support and positive feedback. Furthermore, a relationship with a mentor can serve as a corrective experience for children who may have unsatisfying relationships with other adults in their lives (Rhodes, 2005). Mentoring has been used in a variety of settings, including schools, as an effective intervention strategy leading to positive child outcomes across behavioural, social, emotional and academic domains (O'Rourke, 2016; Raposa et al., 2019).

Because the majority of mentors in today's formal youth mentoring programmes are volunteers, their retention is key to the functioning of interventions and the achievement of the intended outcomes for youth (Stukas et al., 2013). Supporting mentors to stay on the programme also translates into benefits for the programme because, for example, it reduces the burden on staff to recruit and train new mentors and reduces early match termination (MENTOR, 2015). Premature match termination has important implications because it is associated with unfavourable outcomes for youth. There is indeed evidence that a minimum of 24 weeks of match duration is required for the relationship to develop and determine any benefit (Grossman et al., 2012).

Volunteering can be defined as a form of organised prosocial behaviour planned and maintained over time that benefits strangers and develops within a structure. Mentoring programmes can be understood in light of theoretical frameworks of volunteerism (Finkelstein, 2010; Marta and Pozzi, 2007; Marta and Santinello, 2010; Omoto and Snyder, 1995; Penner, 2002). Because volunteer role identity is key to their commitment to and continuation of their work (Grube and Piliavin, 2000), this study attempted to integrate theoretical models developed in the general psychological literature on volunteerism to examine factors involved in the development of a volunteer role identity among mentors in school-based mentoring programmes.

Volunteer role identity and mentor commitment

Classical work in social psychology literature distinguished between three components of one's identity (Linville and Carlston, 1994; Stryker, 1986; Thoits and Virshup, 1997; Turner, 1978). Personal identity consists of self-definitions in terms of personal characteristics, social identity reflects the identification of the self with a social group or category and role identity is the definition of self as a person who performs a particular social role. The role identity model (Callero et al., 1987; Piliavin and Callero, 1991) was developed to conceptualise this latter component of one's identity. This theory states that individuals engage in voluntary actions because they have developed a strong volunteer or service identity. As volunteering becomes part of their identity (i.e. a role identity as a volunteer), it drives continued volunteer actions because the individual strives to behave in concert with the changed self-concept in a virtuous cycle. Prior research has consistently shown that the strength of one's role identity correlates with the voluntary donation of time, money and even blood (Grube and Piliavin, 2000; Lee et al., 1999; Piliavin and Siegl, 2007). Continued participation in a mentoring programme can facilitate a process of internalisation of the volunteer role, which is adopted as a component of the self, an identity or role identity (Piliavin and Callero, 1991). Such internalisation ultimately promotes volunteer engagement, identification with the programme and long-term commitment (Stukas et al., 2013).

Factors contributing to volunteer role identity

There are at least three key factors involved in the development of a volunteer identity. First, despite established theoretical frameworks of youth mentoring (Keller, 2005; Noufou, 2014; Rhodes et al., 2006; Taylor and Porcellini, 2013) considering influences of external social networks (e.g. mentees' family, programme staff and peers as part of the organisational culture) in supporting mentoring matches, little attention has been paid to the role of other social referents, including mentors' significant others (e.g. family members and close friends). The theory of planned behaviour (TPB; Ajzen, 1985, 2011) states that subjective norms (i.e. the perceived social approval of engaging or not engaging in a behaviour) are essential in the process of deciding to volunteer (Brayley et al., 2015). In the mentoring context, this means that when mentors feel the support of significant others, they are more likely to stay engaged in the programme and continue mentoring. This theory has received empirical support across many fields, and its value as a theory for explaining volunteering has been demonstrated in a variety of populations and settings, including young adults (Hyde and Knowles, 2013; Marta et al., 2014) and elderly people (Warburton and Terry, 2000; Warburton et al., 2001). The TPB has also been validated within the context of related helping behaviours such as charitable giving (Smith and McSweeney, 2007; van der Linden, 2011) and blood donation (Reid and Wood, 2008).

Second, the concept of habit can be used to explain repeated behaviours. Habits are generally understood as semi-automatic performances of well-learned behaviours (Charng et al., 1988; Wood and Rünger, 2016). The more an individual performs a behaviour, the more likely it is that the behaviour will become a habit. Thus, past repeated behaviours can be an indicator of habit. Mentoring is a form of repeated behaviour: mentors meet multiple times with their mentees over a relatively long period, and they can be part of a number of matches over time. Importantly, repeated behaviours (i.e. having been a mentor for some time) can contribute to the construction of a role identity as a volunteer mentor.

Third, Stryker (1980) highlighted the importance of social relationships in the definition of one's self and role identity. In mentoring programmes, the mentor–mentee relationship is a key factor, and youth development goals can be achieved only to the extent that the mentor and youth forge a strong connection characterised by mutual respect, trust and empathy (Rhodes, 2005). The quality of this relationship can contribute to mentors developing a self-concept that includes being a mentor and, consequently, a role identity as a volunteer. Even though relationship quality is a multidimensional construct (McMorris et al., 2018), closeness is considered its organising construct throughout mentoring literature (Nakkula and Harris, 2013). It is a key indicator of relationship quality (DuBois and Neville, 1997; Herrera et al., 2007; Parra et al., 2002) and can affect the likelihood of achieving mentoring intended outcomes (Bayer et al., 2015). Importantly, mentor–mentee relationship closeness can contribute to volunteer role identity development. Relationship closeness can be, in turn, determined by mentors' social skills (i.e. the ability to provide verbal and non-verbal responses that are adequate to the situation; Herrera and Karcher, 2013; Lorr et al., 1991; Parra et al., 2002) and satisfaction with the relationship (Herrera et al., 2007). More specifically, in the Omoto and Snyder (2002) model of volunteerism, the interpersonal relationships between volunteers and recipients of their services (i.e. mentors and mentees in the case of mentoring programmes) have been examined. Mentoring programmes are essentially relationship-based interventions, and volunteers have high expectations for the quality of such relationships. Mentors enter the relationship with a strong desire to make a positive difference in the lives of young people, though they may be easily discouraged if they feel unable to develop strong ties with their mentees (Spencer, 2007). The degree of their satisfaction in regard to this relationship is key to maintaining the commitment to the relationship and the programme overall (Omoto et al., 1998).

In sum, there is evidence that mentoring relationship closeness serves as an important mechanism for change in mentee outcomes, though little is known in regard to relationship closeness outcomes for mentors and more specifically how it affects their mentor role identity.

Aims and hypotheses

This study aimed to examine factors associated with the volunteer role identity of mentors. The role identity model (Callero et al., 1987; Piliavin and Callero, 1991) and theories describing factors involved in repeated behaviours (i.e. TPB) provided the overall theoretical framework. Figure 1 depicts the overall conceptual model. We expected habit (Hp1) and subjective norms (Hp2) to be positively associated with developing a mentor role identity. In addition, by drawing on the literature on the importance of match quality in mentoring programmes, we expected the association between satisfaction with the mentor–mentee relationship (Hp3) and mentors' social skills (Hp4) with mentor role identity to be, respectively, fully and partially mediated by relationship closeness (Hp5). For sensitivity analysis purposes, an alternative model considering satisfaction with mentor–mentee relationship as the mediator between relationship closeness and mentor role identity was considered (alternative Hp5).

Method

Data collection

Researchers contacted organisations that were conducting school-based mentoring programmes in Italy. These were the Società Umanitaria (Programma Mentore), Telefono Azzurro (One-to-One Programme) and some Catholic parishes. (For an example of how these programmes work, see Aresi et al., 2020.) Programme managers were informed about the study and asked to disseminate a survey questionnaire to mentors within their organisation. The university ethical protocol was followed: This includes obtaining signed informed consent from participants (both programme managers and volunteers in this case), the right for participants to withdraw at any time and anonymisation of any data collected. All mentors were informed of the research aims and were asked to voluntarily participate in the study. Participants completed a paper-and-pencil self-report questionnaire. Participants were not compensated for taking part in the research.

Measures

Habit

Habit was defined as a “semi-automatic performance of a well-learned behaviour”, and past repeated behaviours are an indicator of habit. Habit was measured by a single item: “How long have you been volunteering as a mentor?” Participants indicated how many years they had been mentoring.

Subjective norms

Two ad hoc items (“People I care about think it is correct to give time to help young people in need” and “People I care about think it's a good thing I'm volunteering”) were used to measure subjective norms. Participants responded on a five-point Likert scale (1 = not at all; 5 = very much).

Satisfaction with the mentor–mentee relationship

Satisfaction with the relationship was measured using an ad hoc item: “In general, I am very satisfied with the relationship with my mentee”. Participants responded on a five-point Likert scale (1 = totally disagree; 5 = totally agree).

Relationship closeness

Relationship closeness was measured using a 26-item scale that included a modified version of the Relationships Closeness Inventory (RCI, 16 items; Berscheid et al., 1989) and Sternberg's (1997) Triangular Love Scale (TLS, 10 items). The development of the RCI was based upon the concept of interdependence as proposed by Kelley et al. (1983). The RCI includes three subscales (frequency, diversity and strength). In our study, the strength subscale was used to assess the degree of perceived influence each relationship partner had over the other. The intimacy subscale (10 items) of the TLS (Sternberg, 1997) was used. Participants responded using a five-point Likert scale (1 = strongly disagree; 5 = strongly agree). Item examples are “My mentee influences the way I feel about myself” (RCI modified scale) and “I have a good relationship with my mentee” (Sternberg intimacy subscale).

Social skills

Social skills were measured using a three-item scale drawn from the social skills inventory (Riggio, 1986). These items measure social skills, such as social expressivity, social sensitivity and social control, within the verbal/social domain. Participants responded on a five-point Likert scale (1 = totally disagree; 5 = totally agree). A sample item is “I usually take the initiative to introduce myself to strangers”.

Mentor role identity

Mentor role identity was measured using the eight-item Role Identity Scale (Callero et al., 1987). Participants responded on a five-point Likert scale (1 = strongly disagree; 5 = strongly agree). Items examples are “To be a mentor means more than to do mentoring service”, “To mentor is an important part of my identity” and “I'm proud to belong to my mentor group”.

Data analysis

Path analyses were used to investigate predictors of relationship affective quality and mentor role identity in accordance with the hypothesised theoretical model (see Figure 1). The software Amos 5 (Arbuckle, 2003) was used. Guidelines indicate the adequacy of a path analysis model, which is determined by a chi-square test (χ²). This test evaluates how well the covariance matrix implied by the model fits the covariance matrix of the observed data. However, because the chi-square is heavily influenced by sample size (Bollen and Long, 1994), additional fit indices are generally used (Bentler, 1990; Jöreskog and Sörbom, 2004). The comparative fit index (CFI) has been proposed as the most appropriate fit index because it has a 0–1 range, has a small sampling variability and is not affected by sample size (Bentler, 1990; Garrett et al., 1994). In addition, the root means square error of approximation (RMSEA) was used to identify the acceptable error level. Indices lower than 0.080 are considered acceptable (Bollen, 1989). Its 90% confidence interval index is also reported. The mediating effect of relationship closeness was tested using the accelerated-bias-corrected bootstrap estimation procedure, which yields the most accurate confidence intervals (CIs) for the indirect effects (MacKinnon et al., 2004). In the procedure, the given sample size was randomly resampled 10,000 times with replacement, and then 10,000 estimations of the indirect effect were calculated. When the 95% CI for an indirect effect did not include zero, the indirect effect was significant. An a priori power analysis was conducted to determine the minimum sample size. Results indicated that recruitment of 67 mentors was required to detect >0.30 significant associations among variables at p < 0.05 with 80% power.

Results

Participants

A total of 112 mentors were contacted; 80 agreed to take part in the study. Nine participants with missing data were excluded, leading to a sample of 71. The mean age was 36 years (SD = 16.41; range 18–68), and 45 were female (63%). A total of 42.2% of the participants had completed a master's degree, 42.3% high school degree, 14.1% had a lower degree and 1.4% had completed a PhD programme. The majority were single/not married (62%). The remainder were married, divorced or widowed. A total of 22.5% had a full-time job, 25.4% were retired or unemployed, 25.4% were students and 15.5% both worked and studied at the same time. Seven percent were working part-time, and 4.2% were looking for a job. The majority (68.1%) were involved in volunteerism from 1 to 5 hours per week, 31.9% from 6 to 15 h per week.

Path analyses testing proposed model

Descriptive statistics and correlations between variables are presented in Table 1. With the only exception of the variable habit, the data showed a normal univariate distribution given that most skewness and kurtosis values fell within the range of −1.0 to +1.0. The variable habit presented values of asymmetry and kurtosis above the threshold because of the presence of two outlier cases of mentors with >20 years of experience. Inspection of scatter plots demonstrated there was no major violation of the linearity of the relationship between the independent and dependent variables. Residuals were not correlated to any independent variable (all ps > 0.05). Multicollinearity was not an issue because there was no absolute correlation coefficient of >0.7 among two or more predictors. This is confirmed by values of the variance inflation factor for each predictor being below 10, the highest being 1.2.

Figure 2 shows results of the model. The chi-squared test was not significant. This indicates that the empirical data did not differ from the theoretical model (χ² (df) = 12.604 (9); p = 0.181). The χ²/df ratio was deemed acceptable (<2) (χ²/df = 1.400). CFI of 0.924 was acceptable, as well as the RMSEA (0.076 [0.000 −0.165]. There were six significant paths. Mentor role identity had two direct predictors: volunteering habit (Hp1) and subjective norms in regard to this behaviour (Hp2). The association between mentor role identity and two further predictors, satisfaction with the mentor–mentee relationship (Hp3) and social skills (Hp4) was respectively fully and partially mediated by relationship closeness (Hp5). The indirect effect of satisfaction with the relationship on role identity was significant (indirect effect = 0.087, p < 0.01, 95% CI = [0.016, 0.208]), as well as that of social skills (indirect effect = 0.073, p < 0.05, 95% CI = [0.002, 0.164]). The direct effects of satisfaction with the relationship and social skills were not significant (β = −0.163, p > 0.05), whereas the direct effect of social skills was significant (β = 0.367, p < 0.001). All hypotheses were confirmed.

Test of the alternative model considering satisfaction with mentor–mentee relationship as the mediator between relationship closeness and mentor role identity (alternative Hp5) resulted in a poor fit (χ² (df) = 15,673 (8); p = 0.047). The χ²/df ratio was deemed acceptable (<2) (χ²/df = 1.959), though CFI (0.838) and RMSEA (0.117 [0.012 −0.203]) were below acceptable levels. The association between satisfaction with the relationship and mentor role identity was not significant. Therefore, no evidence of such mediation effect was found.

Discussion

The work of mentors is crucial to the achievement of mentoring programmes' intended outcomes. By drawing on established theoretical frameworks of volunteerism (Callero et al., 1987; Piliavin and Callero, 1991) and research and theory on mentoring interventions (Rhodes, 2005), this study examined factors associated with the development of a role identity as a volunteer by mentors in school-based mentoring programmes.

This is the first study that has attempted to empirically test a social identity model – the role identity model – in the mentoring context. Findings indicate that volunteer role identity is related to the extent mentors have developed the habit of being engaged in volunteering and their perception that important people around them support and approve their commitment. Continued participation in a mentoring programme can be understood as a habit developed over time, and results of this study suggest that having been a mentor for some time may facilitate a process of internalisation of the volunteer role and therefore be related to greater commitment to the programme in the long run (Piliavin and Callero, 1991; Piliavin et al., 2002). Previous research has demonstrated that opportunities for involvement in prosocial activities, such as mentoring, play an important role in the development of role identity as a volunteer (Flanagan, 2004; Piliavin et al., 2002). These activities offer individuals opportunities to connect and collaborate with others and promote solidarity (Flanagan, 2004), thus further establishing a role identity as a volunteer (Flanagan et al., 2005).

In accordance with applications of TPB to voluntary action (Brayley et al., 2015), results of this study underscore the importance of mentors feeling supported by others. Modelling and encouragement are important factors in the development of social and role identities (Dovidio et al., 2006). Modelling refers to the idea that people observe relevant others (Nesbit, 2012; Perks and Konecny, 2015; Quaranta and Dotti Sani, 2016). Engagement in prosocial behaviour and approval and encouragement by significant others to perform such behaviour are believed to induce prosocial behaviour in the observant. Results of this study indicate mentors' social identity as a volunteer is related to feeling of support from significant others. These results offer a new perspective to understand mentors' commitment to the programme because it complements research revealing the relevance of the quality of the programme and the support offered to mentors (Herrera and Karcher, 2013; MENTOR, 2015). Higher levels of staff monitoring and support have been found to be associated with greater relationship quality because mentors who are struggling in their relationship may receive guidance and support (Herrera et al., 2013). Social approval sought from important people in mentors' lives is less instrumental and more emotional but can still be important to sustain their engagement and commitment to the programme.

In this study, the connections between volunteer role identity and two further predictors, satisfaction with the mentor–mentee relationship and social skills were mediated by mentoring relationship closeness. The mentoring literature (Bayer et al., 2015; DuBois and Neville, 1997; Grossman and Rhodes, 2002; Langhout et al., 2004, 2014; Rhodes et al., 2005; Spencer, 2006) highlights the importance of the mentor–mentee relationship in achieving positive outcomes for youth. An important and novel finding of this study is that relationship closeness also contributes to mentors developing a role identity as a volunteer. Importantly, mentors use social skills acquired in their contexts of growth and training to benefit the mentoring relationship (Felice and Tagliavini, 2004; Herrera et al., 2007; Parra et al., 2002). In accordance with previous studies in this field (McMorris et al., 2018), this highlights the importance for programmes of developing mentor social skills and continuously sustaining them, because the expressivity, sensitivity and sense control in the relation with a fragile child enhance the sense of closeness in the mentor–mentee bond.

In addition, results of this study are consistent with those of the general volunteerism literature showing that mentors gain personal satisfaction from having voluntarily dedicated themselves to people in need of support (Marzana et al., 2010). Our results expand upon such literature by demonstrating that this in turn can foster mentor–mentee relationship closeness. Rhodes (2005) concluded that positive mentoring relationships are characterised by the lack of disappointment and the presence of satisfaction for the relationship with the mentee.

Implications for research and mentoring practice

Findings of this study have several implications for future research and practice. First, mentor role identity can play an important role in volunteers' commitment to and retention in mentoring programmes, and therefore in the achievement of greater match length. This is important because avoiding early termination is crucial to achieve mentoring programmes intended outcomes for youth (DeWit et al., 2016; Grossman et al., 2012) and reducing volunteer turnover, which in turn diminishes the burden on staff to recruit and train new mentors (MENTOR, 2015). Strategies that foster mentors' role identity as volunteers could be incorporated into standard practice for orienting and training new mentors and for providing ongoing support and supervision. After orientation and initial training, support offered to mentors often declines (Stukas et al., 2013); instead, it is important to focus on the experience of mentoring and encourage mentors to enhance their skills to successfully develop effective mentoring relationships and feel satisfied. For example, it might be useful to offer mentors participation in support groups that contribute to reinforce their role identity as mentors through the creation of a social network with shared experiences and expectations of mentoring. Future research could further examine how role identity is developed over time and empirically test whether an enhanced identity as a volunteer predicts longer match duration and long-term commitment to the programme.

Second, the role of the social environment in fostering sustained volunteer engagement has been well examined (Rossi et al., 2016), though it has received little attention in the mentoring literature. This study examined the effect of the approval from family and close friends in the development of a volunteer role identity. Future research could investigate the effect of approval from other social referents, including fellow mentors or other people (e.g. colleagues). In accordance, mentoring programmes may develop strategies to help mentors feel supported by important people in their lives, for example, by organising activities and social events with them.

Third, the subjective benefits mentors get from their voluntary work can foster their engagement in the programme. For example, they can derive pleasure and satisfaction by expressing their prosocial values, learning new social skills, receiving the gratitude of their mentees, being productive and creative and enriching their personal identity and their social network (Livi et al., 2019). Giving (i.e. volunteering) is indeed a way of getting (Piliavin and Siegl, 2007, 2015). Mentoring programmes should therefore facilitate the matching between mentors' personal motivations and tasks that result in satisfaction and positive volunteer outcomes (Clary et al., 1998; Finkelstein, 2008; Hustinx, 2010; Tschirhart et al., 2001). In addition, as with other types of relational volunteering (Pozzi et al., 2017), it is necessary to consider the quality of the mentor–mentee relationship and the particular experience of mentors (Bayer et al., 2015; Rhodes, 2005; Rhodes and DuBois, 2008). It is of fundamental importance to supervise the progress of these relationships and the capacity of mentors to deal with any difficulties

Limitations

The present study has several limitations that are noteworthy. First, the cross-sectional design of the study constrains the interpretation of causal effects. Further research with longitudinal designs using larger samples is needed to better examine the causal relationships and mediation effects. Second, the relatively small sample size and relatively high proportion of single respondents and respondents with certain religious affiliations, restrict representativeness and have implications for the potential transferability of the findings. A third limitation is that this study employed only self-report measures, which might be susceptible to response bias (e.g. social desirability). Moreover, the measures for habit, subjective norms and satisfaction were developed ad hoc and future research would benefit from the use of validated measures. Multiple methods for evaluation would be beneficial in future research, including data on to what extent intentions to stay on the programme reflect actual behaviour afterward. Fourth, satisfaction with the relationship was measured using a single item which may not have captured the full breadth of the construct and limited the possibility of addressing reliability issues.

Conclusions

This study empirically investigated the applicability of theoretical models of voluntary action to mentoring programmes. Findings demonstrate the importance of match quality and closeness as a key component of mentors' engagement and their role identity as volunteers. This underscores the importance of considering mentors primarily as volunteers whose motivation and engagement need to be nurtured over time. Mentoring programmes can develop guidelines and practical strategies drawn from the literature on volunteerism to promote the development of role identity amongst their volunteer mentors.

Figures

Figure 1

Theoretical model of predictors of mentor role identity

Figure 2

Path analysis model

Table 1

Means, standard deviations and correlations between variables

	M	SD	1	2	3	4	5	6
1. Habit	5.13	4.46	1	0.135	0.124	0.052	−0.027	0.309**
2. Subjective norms	3.87	0.71	–	1	0.139	0.248*	0.201	0.350**
3. Satisfaction with the mentor–mentee relationship	3.83	0.94	–	–	1	0.341**	0.147	0.052
4. Relationship closeness	2.63	0.46	–	–	–	1	0.299*	0.397**
5. Social skills	3.24	0.61	–	–	–	–	1	0.450**
6. Mentor role identity	3.80	0.75	–	–	–	–	–	1

Note(s): **p < 0.01; *p < 0.05