ram's seminal work on mentorship defined the subject in business literature (Kram, 1985;Kram, 1983). Mentorship research has focused on what the protégé gets from the mentorship relationship, how the relationship operates, and more recently what the mentor gets from the mentorship relationship (Ghosh & Reio, 2013). Mentorship has been recognized as a critical organizational function for Peter Drucker addressed both the promise and concern of leader development initiatives by noting that "there may be 'born leaders,' but there are surely too few to depend upon them." (Smith, 2003). Drucker's point reinforces the importance of leader development activities for organizations while subtly reminding the community of organizations that hoping or waiting for leadership to emerge is likely futile.

A second area of research focused on lifeseason and career-stage has been applied to the workplace to show how individuals pass through different stages of their lives, and at each stage they have different needs. Life-season and career-stage research has been used to inform why some people choose to become mentors, but little empirical work has been done to connect life-stage research with the needs of protégés. As individuals progress through different stages of their career (and life), they encounter new challenges even as they have achieved mastery of earlier challenges. We posit that while mentorship continues to be valuable at every career stage, the needs of the protégé evolve. Different mentoring attributes gain in importance at different career stages.

The primary purpose of this study is to propose and test a theoretically based model of mentoring that explains the developmental needs of protégés based on the current stage of a protégés' life or career.

The traditional perspective of mentoring generally holds that a seasoned senior executive will guide and sponsor a young protégé over the course of his career (Levinson D. D., 1978;Kram, 1985). Recent reviews of social science literature identify some 40 to 50 identify definitions of mentorship (Ghosh & Reio, 2013; Haggard, Dougherty, Turban, & Willbanks, 2011; Crisp & Cruz, 2009;Eller, Lev, & Feurer, 2014). More recent literature includes the recognition of mentoring relationships in the context of a network of multiple mentoring dyads (Kram, 1985;Higgens & Kram, 2001), peer-to-peer mentoring (Dennison, 2010), and reverse mentoring (Marcinkus Murphy, 2012).

Kram's seminal work in mentoring identified key mentor roles or functions in providing developmental support to protégés within career and psychosocial contexts (Kram, 1985). Later researchers refined these functions described by Kram to define and test specific mentor roles such as "Acts as a role model", "Teaches and advises on organizational politics", and "Teaches job skills." (Dreher & Ash, 1990;Turban & Dougherty, 1994;Steinberg & Foley, 1999). Steinberg and Foley's exploratory factor analysis of 16 mentor functions and behaviors yielded three distinct mentoring dimensions including personal development, career sponsoring, and job coaching. These results indicate that the roles of mentors fall within three broad categories of mentor attributes (Steinberg and Foley, 1999).

Daniel Levinson, a major contributor to the early formation of mentoring knowledge, is credited for developing the Theory of Life Structure Development which describes the evolution of adult development through four primary life structure stages: pre-adulthood (age 0 -22), early adulthood (age 22 -45), middle adulthood (age 45 -65), and late adulthood (age 65 and beyond). The Theory of Life Stage Development holds that social, psychological, and career activities vary as an individual progresses through life stages. Levinson describes the process of progressing through the stages of life structure as similar to passing through the seasons of a year (Levinson, et al, 1978;Levinson, 1986).Two of Levinson's eras, Early Adulthood and Middle Adulthood, provide relevant, substantive bases for examination in this study as these two eras cover a period that is generally consistent with a traditional professional career covering an age range from 22 to 65.

Similar to Levinson, Erik Erikson articulated individual development of their identity through an eightstage process highlighted in his Life Cycle Theory. Erikson viewed each of these stages as essential steps that must be resolved in order to develop as an adult. By overcoming the associated psychosocial crises that each stage presents, individuals are prepared to graduate to the next stage in the life cycle (Erickson,1959;).

Levinson's Life-Season model and Erikson's Life Cycle model suggest that as individuals progress through chronological development stages, individuals will have different needs. A mentor would better serve a protégé by recognizing the Life-Season of the protégé and adjusting mentoring behaviors according to the needs of the protégé, which will at least in part be the result of the protégé's Life-Season.

Super and Kram articulate Career-Stage models that are similar to Levinson and Erikson's Life-Season models, but focused on professional careers. The career stages of Super's model include exploration, establishment, maintenance, and decline. An individual progresses through their career as they complete each progressive stage through maintenance then career progression declines as an individual enters the disengagement phase. Each of the phases consists of specific psychological tasks and developmental needs required to prepare the individual for the following stage, which ultimately culminates in retirement or some other These three dimensions of mentoring and their interrelated associations are depicted in figure 1. The model consists of the 16 mentor functions identified in the literature and their corresponding functional dimensions. Each of these dimensions represents a functional component of the mentoring concept. Mentoring received by a protégé may consist of components of one or a combination of these dimensions. (Kram, 1985;Dreher and Ash, 1990;Turban and Dougherty, 1994;Steinberg and Foley, 1999). withdrawal from the career path (Orstein, S., Cron, W., & Slocum, J., 1989; Super, 1990;Super, 1957;Sullivan, 1999). Kram's descriptive theory of mentoring references Levinson's life development theory and posits that mentoring needs and concerns will change over the course of a career. She expects that varying self, family, and career concerns of protégés will affect the nature of mentoring relationships as individuals will seek specific developmental assistance from their mentors consistent with their current life circumstance. Kram describes mentoring differences in early, middle, and late career stages based on her interviews of protégés (Kram, 1985). In the early career-life stage, Kram expects that individuals will seek mentoring relationships that will enhance their vocational ability and establish a level of competence as they begin their careers. Individuals in the middle career-life stage are not expected to be concerned with competence, but rather their prospects for professional ascendancy in the organization or field; consequently, protégé's in this stage will be more likely to seek mentor support in the form of career advice or sponsorship in order to facilitate their professional progress. Finally, protégés in the late career-life stage are not expected to be primarily concerned with either competence or upward career mobility, but instead they are motivated to ensure that their work efforts have meaning or make a valuable contribution toward a greater purpose. Mentors of late career-life protégés are expected to provide validation or broad career guidance to their mentors (Kram, 1985).

Levinson, Super, Erikson, and Kram clearly establish that developmental needs and processes vary according to an individual's life-career season. Consequently, the following model (Figure 2) is proposed to depict the mediating effect of life-career season, borrowing from Levinson's metaphoric characterization, on mentoring as a component of leadership development.

This model extends the previously described leadership development (Figure 1) and mentoring (Figure 2) models by depicting the moderating effect of career season on the concept of mentoring. The direct application of existing life and career development theory to the model indicates that the level of helpfulness for each dimension of mentoring will vary across groups of individuals according to their career stage.

In addition to establishing a case that life-career season will impact the developmental functions of mentoring, the theoretical foundations established by Levinson, Super, Erikson, and Kram also provide a basis to understand which dimensions of mentoring and corresponding mentor attributes will be most beneficial to protégés according to life-career stage.

Early careerists and young managers are expected to seek the job coaching dimension of suggest that midcareerists will find this dimension and associated mentor functions most beneficial to their development. At this life-career developmental stage, individuals have achieved competence in their chosen field and now seek to achieve higher levels of responsibility, status, and rewards. Protégés will find mentor attributes corresponding with sponsorship functions such as career advice, job attainment, and protection that will enable them to obtain higher levels of professional achievement.

Finally, the mentoring dimension and component mentor attributes of personal development is expected to be most desired by late careerists. At this point in the life-career development season, individuals have already achieved competency and professional achievement and are now seeking a level of professional satisfaction or fulfillment in their profession. Erikson described this stage as wisdom. Similarly, Levinson, Kram, and Super described end of cycle developmental needs where individuals seek validation in their efforts from mentors.

Although there are compelling theoretical bases for establishing specific dimensions of mentoring as potentially most beneficial to distinct life-career groups, we should be aware that all mentor functions and dimension are expected to be helpful in varying degrees for all life-career groups. For instance, it is not unreasonable to expect that early careerists may seek to find higher meaning in their work and even late careerists require vocational support. In these cases, mentors would be called upon to provide developmental support that corresponds with the need of the individual protégé.

V.

Data used in this study was drawn from extant Department of Defense survey data obtained from a commissioned officer branch of the United States Army, the Medical Service Corps ("MSC"). Authorization in both deployable combat units such as the 101 st Airborne Division (Air Assault), and in non-deployable healthcare delivery organizations such as Walter Reed Army Medical Center. They serve in general management roles (CEO, COO equivalents), as well as providing highly specialized scientific services as researchers, practitioners, and providers (Ginn, 1996).

The dataset consists of 1,868 respondents surveyed from approximately 3,900 MSC officers. The first group of officers (n= 970) consisting of 2 nd Lieutenants, 1 st Lieutenants, and Captains will be grouped as "Company Grade" consistent with the Army model (Dalessandro, 2016). The second group of officers (n=897) consisting of Majors, Lieutenant Colonels, and Colonels will be grouped as "Field Grade" consistent with the Army model (Dalessandro, 2016). One respondent, a Flag Rank officer, was omitted from the analysis.

The primary variables of interest in this study are drawn from a 16-item operationalization of mentor functions and behaviors based on Kram's (1985) description of mentor functions. These items were originally developed by Dreher and Ash (1990) and adapted by Steinberg and Foley (1999) for a military population.

Respondents were asked to rate the following attributes or behaviors (Table 1) on a five-point scale ranging from 1 (Not Important) to 5 (Essential) with the following question: "How helpful do you believe each of the following MAs is in terms of providing effective mentoring to you at this time?" It is important to note the emphasis on the impact of mentoring to the individual him or herself at the present, not about their views on mentorship in general.

We test the data using confirmatory factor analysis and test of mean vector components of structural equation modeling.

Structural equation modeling provides a robust mechanism for model testing particularly with a multi-dimensional construct and latent variable (Tabachnick & Fidell, L., 2001; Williams, L., Edwards, J., & Vandenberg, R., 2003). In the course of conducting the factor analysis procedure, we evaluate as appropriate: outliers among cases, sample size, normality and linearity of variables, multicollinearity, and outliers among variables (Tabachnick and Fidell, 2001).

Table 2 depicts the distribution of responses for all respondents among the sixteen mentor attributes (MA) in the survey instrument. Additionally, exploratory factor analyses demonstrate that these variables load on multiple factors and provide scarce additional explanatory value in terms of accounting for variance. Furthermore, Steinberg and Foley (1999) also found that these mentor attributes loaded on multiple dimensions.

Evaluation of these four mentor attributes from a theoretical perspective indicates that these functions are either redundant to other mentor attributes or are implicitly incorporated within other mentor attributes. For instance, MA7, Provides Career Guidance, appears to exist as a general proxy for the process of mentoring and could reasonably fall within either or all of the three dimensions of mentoring. Similarly, MA8, Invites You to Observe Activities at His or Her Level, could reasonably be expected to fall within all three proposed dimensions of mentoring. Finally, MA5, Provides Support and Encouragement, and MA6, Provides Personal and Social Guidance, represent mentor functions that are implicitly incorporated in several of the other mentor attributes. An illustration of this implicit affect could involve a senior officer securing a premier assignment for a protégé. While this action is explicitly a function of MA10, Assists in Obtaining Future Assignments, the mentor's interest and effort in securing this position demonstrates support, encouragement, and career guidance to the protégé.

In order to adjust for the cases of multiple factor loadings and highly inter-correlated variables, mentor attributes 5 -8 have been omitted from the analysis and the following, enhanced model of mentoring will be evaluated for the effect of Life-Career Season (Figure 3). This model remains consistent with the basic findings of Steinberg and Foley (1999). Respondents assigned a uniform random distribution greater than 0.5 were assigned to group one while the remaining respondents were assigned to group zero.

Evaluation of the study was conducted by multiple two-group confirmatory factor analysis. Cross-validation of the twelve-item, three-factor model was confirmed within both Company Grade and Field Grade subgroups prior to conducting tests between groups. Both cross-validation results achieved root means square error of approximation (RMSEA) of less than 0.07 and significant Satorra-Bentler Scaled Chi-Square scores.

Multi-group confirmatory factor analysis results between Company Grade and Field Grade respondents are depicted in Table 4. The Company Grade group accounted for 48.88% of the model chi-square and the Field Grade Group accounted for 51.11% of the model chi-square. The RMSEA of 0.0692 and the Comparative Fit Index (CFI) of 0.985 indicate that the twelve-item, three-factor model is a reasonable fit (Tabachnick and Fidell, 2002). Based on these results, there appears to be equivalence and factorial invariance between Company Grade and Field Grade respondents for the twelve-item, three-factor measurement model. Specifically, factor form, factor loadings, and factor construct exhibit reasonable fit between groups. Figures 4 and 5 depict the factor loadings of the model in each of the two groups. Each of the remaining factor loadings exceed the commonly recommended criteria of 0.40 (Stevens, 2002). Evaluating the means of latent variables demonstrate whether there may be differences across dimensions of mentoring based on life-career stage. Comparisons of the mean vectors between Company Grade and Field Grade respondents were generated to evaluate theoretically expected differences between groups. Results of tests for invariance among latent variable means are depicted in Table 5. Results of the tests of latent variable mean differences indicate that statistically significant differences exist between groups for each of the three dimensions of mentoring in the mentoring model. The 0.15 estimate for the Personal Development dimension indicates that latent variable mean for Field Grade respondents is higher than Company Grade respondents. This result is consistent with what we expect from the theoretical basis of this study as Field Grade officers or late life-career stage individuals are expected to seek validation of their lives' work from their mentors in a way that enriches their perceptions of their complete life and/or career path. The -0.14 estimate for the Career Sponsoring dimension indicates that the latent variable mean for Field Grade respondents is lower than Company Grade respondents. This result is generally consistent with the theoretical basis of this study as Company Grade officers as similar early\midcareerists are expected to seek developmental support from their mentors that facilitates attainment of higher levels of professional and personal achievement; however, some Field Grade officers, particularly, recently promoted Majors, might also be considered mid-careerists. Finally, the -0.16 estimate for the Job Coaching dimension indicates that the latent variable mean for Field Grade respondents is lower than Company Grade respondents. This result is also consistent with the theoretical basis of this study as Company Grade officers and early\mid-careerists are expected to seek development support that enables the protégé to achieve competency in their designated field.

Building on previous theoretical and empirical work, this study provides empirical evidence to support the theoretical expectation that the needs of protégés change with life-season and career-stage. Using a sample drawn from the Medical Service Corps, we refined a list of mentoring activities and behaviors, but generally confirmed the existing understanding of the dimensions of mentorship. Having confirmed the dimensions of mentoring, we added a layer of nuance that reflected theoretical expectations that life-season and career-stage would mediate the degree of importance of particular mentoring activities. While protégés generally value all three dimensions of mentoring throughout their life-career seasons, the relative value of the three dimensions changes depending on where the protégé is in terms of her/his life-career season. Early careerists seek mentors who can help them develop job skills and sponsor them in career opportunities; later careerists seek mentors for personal development for presumably softer skills. The idea that a more senior protégé would value a mentor role modeling more than an early careerist implies that later careerists understand that executive skills cannot be learned mechanically.

From a practice perspective, this conclusion is useful for individuals who look to engage with potential protégés as a mentor to ensure that the would-be mentor is providing the right mix of mentoring behaviors given the protégé's life-season and career-stage. This conclusion is also useful for organizations that are looking to create formal mentoring programs. Formal mentoring programs should be structured to take into account life-season and career-stage of the proteges, and assigned mentors should be prepared to address life-season and career-stage appropriate issues.

The primary limitation of this study is generalizeability to other populations. The respondent population consists of active-duty United States Army officers. Army officers may not provide appropriate representation of the population of management and strategy practitioners as a consequence of the inherent nature of their profession.

Furthermore, the demographic representation of respondents was 75% male, 25% female and 76% Caucasian, 24% Non-Caucasian which may not accurately map onto other populations.

Similarly, the study may have generalizeability issues within the United States Army as this sample was drawn from a branch in the Army that oriented to combat service support, specifically delivery of healthcare, rather than direct combat support or combat. One consequence of this orientation is that the study population has a greater proportion of women than the rest of the Army, and also a higher level of education. The advantage of using a sample from the Medical Service Corps is the diversity of occupations, training, and roles of MSC officers. This may make the results more generalizable to a civilian population than a sample drawn from a combat arms branch such as the Infantry.

Follow on research might apply the refined 12mentor activities to a population in a different industry and/or with a different demographic mix to confirm generalizability.

This study advances the literature on mentorship as a component of leadership and management development by confirming empirically the theoretical expectations that life-career season mediates the mentoring relationship. Protégés identify different needs depending on their life-career season. Protégé's mentoring needs are mediated by their life-career season. Implication of these findings include reconsideration of mentor-protégé matching processes, time and content focus of mentor-protégé exchanges, and further research to better understand development processes of mentors to effectively address protégé developmental needs. As proteges seek advice, counsel, and\or support from potential mentors, or alternatively, as potential mentors seek to provide development, consideration of life-career season effects may enhance chances for optimal pairing. Similarly, given the time dynamics and competition for time allocations, mentors and proteges may be better served by focusing time or allocating available opportunities for development on those life-career season dimensions most likely to generate value for the protégé at the present time. Finally, if life-career season dimensionality exists for protégé development needs and expectations, additional research and practice enhancement could emerge that may focus on the suitability and development of potential mentors to appropriately apply developmental processes according to life-career season. For instance, one mentor may be perfectly suited to apply job coaching or training development to a protégé, but be inadequately prepared for either sponsorship or personal development aspects of mentoring.

This page is intentionally left blank