CareerSteer – careers test for career choice                                     

        Informal mentoring, white collar and blue, in a British city.
           A  cross-sectional survey concentrating on informal  mentoring 
        was  completed by 404 workers of various occupations in  Norwich.  
        As  in previous studies, mentoring was correlated  with  positive 
        outcomes  for  employees.   Kram's  (1985)  factorial  model   of 
        mentoring relationships was supported, except in the case of  the 
        blue  collar  workers; differences inherent in  the  blue  collar 
        sample,  however,  provided some evidence in favour of  a  causal 
        link between mentoring and the outcomes.
           As  in  other studies, gender differences  were  insignificant 
        except  in the expected areas of income and, in this  study,  job 
            Levinson's (1978) characterisation of mentors as hierarchical 
        superiors  was  supported.   A  correlation  was  found   between 
        potential   proteges'   social  orientation  at  work   and   the 
        probability of becoming mentored.  
           Formal mentoring was not perceived as influential in  careers, 
        although it was considered to be useful for induction purposes. 
            Mentoring, for the purposes of this study, can be  understood 
        as  a relationship - often interactive - between an  older,  more 
        experienced  person  within an organisation and a  younger,  less 
        experienced one, serving a variety of developmental functions for 
        both  of  them (Kram, 1985).   Levinson's model of  adult  career 
        development  (1978) postulates the likely influence of  a  mentor 
        during  the  stage  of  early  adulthood;  one  aspect  of   this 
        relationship  is that it should become more equal as the  protege 
        acquires skills.  Adherents of Erikson's epigenetic theory of ego 
        development (1959) would also expect older successful persons, in 
        a  period  of  'generativity', to actively want  to  be  mentors, 
        passing on the fruits of their experience.
           Kram,      the     leading     theorist      of      mentoring 
        relationships, postulates two superordinate mentoring dimensions, 
        career oriented and psychosocial.   The career oriented dimension 
        derives from the functions of sponsorship, coaching,  protection, 
        exposure,  and  the provision of challenging work  for  proteges; 
        they   largely  relate  to  experience,  rank,   and   influence.   
        Psychosocial  functions  comprise  role  modeling,   counselling, 
        acceptance,  confirmation,  and  friendship; these  are  said  to 
        foster "a clear sense of professional identity" as well as aiding 
        confidence and competence.  
           "Relationships   that   provide  both   kinds   of   functions 
        are  characterized  by greater intimacy and  commitment  and  are 
        viewed  as more indispensable, more critical to development,  and 
        more  exclusive than other relationships at work.   Relationships 
        that  only  provide career functions are  characterized  by  less 
        intimacy and are valued primarily for the instrumental ends  that 
        they  serve in an organizational context.  More often  than  not, 
        relationships   provide  a  subset  of  the  possible   mentoring 
        functions, and, in general, career functions are more  prevalent" 
        (Kram, 1985).   
           Other    models   have   been   postulated.     Schein,    for 
        example, provides a model (1978) concentrating on roles more than 
        functions;  these  include  the roles  of  sponsor,  role  model, 
        teacher/coach/trainer,  leader, protector, opener of  doors,  and 
        talent  developer.    Factor analytic studies by Noe  (1988)  and 
        Scandura (1992), however, have broadly supported Kram's model.
           Dreher     and    Ash    (1990)    found     that     business 
        graduates experiencing extensive mentoring relationships reported 
        receiving  more  promotions, had higher incomes,  and  were  more 
        satisfied   with   their  pay  and  benefits  than   were   those 
        experiencing   less   extensive  mentoring   relationships.    No 
        significant gender relationships were identified, except for  the 
        seemingly  inevitable  income differential.  Other  studies  have 
        reached similar conclusions (e.g. Fagenson, 1989).
           Dreher  and  Ash (1990) have noted the problems in  trying  to 
        identify causal relationships.  'High-fliers' may be adopted more 
        readily as proteges; Whiteley et al (1992) discuss social origins 
        as  a possible variable.  In other words, mentoring may not be  a 
        cause of success.

           Interpersonal  attributes are also thought to have  a  bearing 
        upon  the formation of mentoring relationships.  On the  side  of 
        the  mentor,  Levinson  (1978) and Clawson  (1980)  considered  a 
        hierarchical  superior  to  be typical,  Whiteley  et  al  (1992) 
        considered  the  role  of the manager  within  the  organisation, 
        whilst  Olian et al (1988)  found that differences  in  managers' 
        interpersonal skills significantly affected protege attraction to 
        the  potential mentor, organisational integration of  the  mentor 
        only  proving  significant in the event of  weaker  interpersonal 
        skills  in  the mentor. Considering proteges, Olian et  al  found 
        that  younger  proteges  were  more likely  to  be  attracted  to 
        potential  mentors.   Fagenson  (1992) found  that  proteges  had 
        significantly  greater  needs  for  power  and  achievement  than 
           Another  point  is  that research tends  to  be  American,  is 
        usually based upon studies of managers (often alumni of  graduate 
        business schools), and tends to emphasise formal mentoring  (e.g. 
        Noe,  1988).   Although  some studies  (Berlew  and  Hall,  1966; 
        Clawson,    1980)   have   considered   the   'facilitation    of 
        organisational socialisation' - or induction - American mentoring 
        studies generally relate to major career influences
                             THE SCOPE OF THIS STUDY
           Perhaps it is not entirely coincidental that a British  review 
        of  the  mentoring literature (Collin, 1979) asked  the  research 
        question  of whether or not there were typologies of  development 

        for  non-managers, and if any mentoring roles existed within  the 
        formal  hierarchy.  She proposed that the mentor personifies  the 
        company's  'psychostructure' and acts the midwife in the  process 
        of  socialisation.    One study of MBAs suggested  that  informal 
        mentoring  was more effective than a formal system, although  not 
        necessarily as fair (Iles and Mabey, 1993). 
           If we take into account the point that mentoring systems  have 
        been  adopted  less  frequently in Britain  than  in  the  U.S.A. 
        (banking and further education would appear to be exceptions), it 
        would  appear to be unrealistic to make a direct comparison  with 
        American  models.  Given this inability to test directly  whether 
        or not such systems work similarly across national cultures,  the 
        research may usefully study different phenomena.  
           It is not suggested, however, that differences ininternational 
        culture  are insignificant (c.f. Hofstede, 1980).   Rather  that, 
        contrary to Roche's (1979) U.S. survey of 1,250 executives, where 
        two  thirds had a mentor or sponsor in their early  careers,  and 
        Kram's  assumption  of  the development of  subordinates  as    a   
        mainstream activity within American managerial tasks, the current 
        British industrial climate is not conducive to the development of 
        subordinates  beyond the parameters of obvious short-term  needs.  
        Formal mentoring systems are rare; their usage will be looked at, 
           This  study  concentrates  upon  informal  mentoring,  over  a 
        broad(if  not  thoroughly representative) sample of  the  working 
        population  within  an English city.  A factor  analysis  of  the 

        typology of mentoring relationships will test Kram's (1985) model 
        in this rather different setting.  
           The only measure of intrapersonal attributes was the  question 
        about attitudes to work ('social orientation').  This was applied 
        to Judith Coles' idea of receptivity to mentoring.
        1   That mentored individuals should have more job  satisfaction, 
        higher  wages and greater career opportunities than  non-mentored 
        2   That gender differences will not prove significant (with  the 
        probable exception of income differentials).   
        3   That  the  levels  of mentoring  will  vary  greatly  between 
        professional and working class careers.  
        4   That official mentoring will not appear to be widespread, and 
        that  it  will not be viewed as particularly  beneficial  by  its 
        Further questions to be asked:
           Are Kram's factor analytical qualities valid in Britain?
           Will the mentor typically be a hierarchical superior?   

                         RESEARCH DESIGN AND METHODOLOGY
        The questionnaire
           A correlational cross-sectional survey (see Appendix A),  this 
        distinguished   the   following   demographic   groups:   gender, 
        occupation, age, and income.  Three attitudinal scales (questions 
        2.1  -  2.3) were cruded tests of  job  satisfaction,  subjective 
        attainment  (c.f.  Lawrence, 1984, on the  'social  clock'),  and 
        social  orientation  at work.  The latter, as well  as  measuring 
        affiliation  needs,  also  served as  an  indirect  indicator  of 
        receptivity to mentoring.
           Those   who  considered  themselves  to  have   been   greatly 
        influenced  in  their careers were asked to  continue  with  some 
        details  of the mentoring relationship and then a global  measure 
        of mentoring practices.  Questions represented different  factors 
        postulated by Kram (1985); see Appendix B.  The measure was drawn 
        from Dreher and Ash (1990), who themselves selected items used by 
        Noe  (1988)  and Whiteley et al. (1988).  Minor  amendments  were 
        made to assist comprehension by a wider range of respondents than 
        in the previous studies.  The researcher added a 19th.  question, 
        from  the  Noe study, in order to include an  essential  part  of 
           Brief  questions  on  formal  mentoring  were  asked  of   all 
        respondents,  with  room  for comments.  It  seemed  likely  that 
        formal  mentoring  would  be scarce in the  British  context,  so 
        little space was devoted to this issue.  The issue was placed  at 
        the  rear of the questionnaire to avoid influencing choices  made 
        about informal influences.
        The sample
           Questionnaires were distributed via simple random sampling  of 
        the  working population of Norwich.  The guiding priorities  were 
        to   gain  information  speedily  and  over  a  broad  range   of 
        occupations  and incomes.  Accessibility of subjects  within  the 
        time   and   personal  budget  available  meant  that   a   truly 
        representative  range was out of the question.   Random  sampling 
        was  therefore  simple rather than proportional,  although  still 
        intending to be numerically representative.
           The Census of 1991 considered the population of Norwich,  aged 
        between  16  and  65, to be 76,864.  The  working  population  of 
        Norwich will be less when unemployment is taken into account, but 
        with the well-known phenomenon of commuting - into Norfolk's only 
        city - the working population probably ranges between one and two 
        hundred  thousand people.  Krejcie and Morgan (1970) recommend  a 
        sample  size  of  384 for a population of between  75,000  and  a 
        million;  Roscoe  (1975)  recommends a maximum of  500  for  such 

           One  city  was  chosen to avoid the  confounding  variable  of 
        economic  fluctuations across geographic areas.   Certain  points 
        should  be noted about the context, however. The city of  Norwich 
        is  primarily populated by caucasian people; indeed, it  has  the 
        dubious distinction of being called the 'last white city' by some 
        neo-fascists.  Racial factors such as those considered by  Thomas 
        (1990)  therefore could not be covered; arguably,  this  excludes 
        another potentially confounding variable.
           A distinction of the county of Norfolk is that its  relatively 
        slow  pace  of life is popular and the area  is  therefore  often 
        known  as the 'graveyard of ambition'.  The implications of  this 
        for  such factors as mobility and promotion should be  considered 
        when  considering  the generalisability of this  study  to  other 
        parts of Britain.
        Data collection
           901  questionnaires were distributed, 438 (48.61%)  returning, 
        of  which 3 were discarded as unusable.  404 arrived on time  for 
        analysis.   Questionnaires  were distributed  personally  by  the 
        researcher  in  the case of individuals and  very  small  groups, 
        otherwise  via managers and other officials.  With the  exception 
        of  a few organisations which made alternative arrangements,  all 
        individuals  received  a questionnaire with  (attached  by  paper 
        clip),  an  envelope, stamped first class and  printed  with  the 
        researcher's  Norwich address.   Where requested, the  researcher 
        would  summarise verbally the contents of the  questionnaire;  in 
        larger organisations, the researcher requested distribution to  a 
        range of hierarchical levels.  Follow-up calls were only made  at 
        the suggestion of questionnaire recipients (a rare occurrence).
           Questionnaires   went  out  in  the   following   proportions, 
        essentially based upon accessibility:
        banks 20%; post office 12%; social workers 12%; fire service 11%; 
        diverse  civil  servants  10%;  hospital  staff  6%;  police  6%; 
        education   4%  estate  agents  2%;  recruitment   agencies   2%; 
        voluntary\charitable  organisations  2%;  railway  employees  1%;  
        others,  including self-employed people, taxi drivers, sales  and 
        bar staff, various professionals, and retired individuals  12%.
           Of   the   404  questionnaires   analysed,   the   approximate 
        proportions were: 
        banks  26.5%; hospital\education\social work 14.5%;  post  office 
        9.5%;  police  8.5%;  civil servants 8.5%;   fire  5%   rail  2%;  
        miscellaneous 25.5%.
           Particularly   for  the  purposes  of  factor  analysis,   the 
        occupational  groups  were conflated into blue collar  (125)  and 
        white collar (279) workers. 
           Six  pilot questionnaires were distributed to friends  from  a 
        variety of work backgrounds.  Various criticisms of format led to 
        minor  changes in the questionnaire, including the  inclusion  of 
        HND  as  an academic qualification level.  The relevance  of  the 
        subject matter to particular workers was questioned.
           Of  particular  interest  was  the  idea  of  'receptivity  to 
        mentoring'  as  a likely influence on the take-up  of  mentoring.  
        Judith  Coles,  one of the pilot respondents, raised this  as  an 
        issue;  as  discussed, other researchers have  looked  at  rather 
        specific  attributes  influencing  the  mentoring  process.   The 
        researcher  decided to use the question about social  orientation 
        at  work  as an indirect measure of this; those who saw  work  as 
        divorced  from social aspects were likely to be the  people  who, 
        according to Coles, tended to pass over potential mentors.
           Lack  of  time  prevented a  thorough  pilot  distribution  of 
        questionnaires.  An analysis of the first 50 mentoring scales for 
        those  with  mentoring relationships, indicated  strong  internal 
        consistency; a subsequent Cronbach Alpha reliability test on  the 
        total  177 cases  factor analysed gave a coefficient of .8967  on 
        the first 18 items of the mentoring factor scales, .9027 with the 
        19th. question added.
           Consistency  of the questions on formal mentoring  was  highly 
        unsatisfactory  (.1595).   The  attitudinal  measures  were  poor 
        (.5963),  although  the removal of the social  orientation  scale 
        from the calculations gave the fair result of .6350.  

        Analysis methods
           Factor  analysis was used on the mentoring  behaviour  scales, 
        studying  all of those with mentors and also groups divided  into 
        blue and white collar occupations and also by gender.  The direct 
        method  of  analysis was Principal Components,  with  an  oblique 
        rotation technique (Oblimin).   Kaiser's criterion (Eigenvalue  > 
        1)  was  chosen  for deciding upon the number of  factors  to  be 
        extracted,  supported  by the scree test  (Cattell,  1978).   For 
        clarity,  it was decided to limit analysis of coefficients  to  a 
        minimum  of  .5.    Structure matrices were  usually  chosen  for 
        analysis  (a  pattern  matrice was used in  the  female  category 
        because  it  resembled  the number of  factors  in  the  original 
        statistics); these appear in Appendix H.
           Non-parametric   methods  were  generally  chosen  for   tests 
        involving Likert scales.  Although not based on interval data, it 
        was felt reasonable to apply parametric testing to income given a 
        relatively normal distribution.  
           Similar  considerations allowed the use of parametric  testing 
        on  the aggregated scores of the mentor behaviour  scales,  which 
        served as a measure of the quality of mentoring.  This  dependent 
        variable will be referred to subsequently as 'addfacts'.
           Additional mentoring independent variables were the  existence 
        of  mentoring  or  otherwise,  the number  of  mentors,  and  the 
        duration of the most important of these relationships.

        1   That mentored individuals should have more job  satisfaction, 
        higher  wages and greater career opportunities than  non-mentored 
           This  was supported overwhelmingly; those individuals who  had 
        been mentored differed from those who had not in terms of  income 
        (Kruskal-Wallis  p.  <.0001),  job satisfaction  (p  <.0002)  and 
        subjective estimates of career level achievement (p <.0002).  
           It  should be noted, however, that the same  statistical  test 
        uncovered  similar relationships between academic  qualifications 
        and the dependent variables (income p <.0001; job satisfaction  p 
        <.005).   Although  this  test  failed  to  find  a   significant 
        relationship with career level (chi square 8.4451 p <.1333),  the 
        Sign test did (p <.05 2-tailed).  Unsurprisingly,  qualifications 
        and mentoring correlated (Spearman r= .0946; p <.05 1-tailed).  
           Levels of perceived attainment varied interestingly,  however, 
        when distribution was examined in the light of qualifications.  O 
        level  standard  respondents  appear  to  have  fared  the  worst 
        (Appendix  E).   Those  with  no  qualifications  (Appendix   F), 
        although  clearly having the expected large proportion  who  have 
        failed to achieve, have a higher mean than those with 'A'  levels 
        or  higher  (Appendix G) and have outstripped them  in  terms  of 
           The  quality  of mentoring, as measured by  the  summation  of 
        factorial   item   scores   ('addfact'),   did   not    correlate 
        significantly  with  income  (Pearson), but was  related  to  job 
        satisfaction   (Spearman  .2491,  p  <.001)  and  to   subjective 
        attainment  (Spearman  .2009,  p <.01).  The  number  of  mentors 
        produced a similar set of relationships (Spearman: n.s.; r=.2491, 
        p <.001; r=.2009, p <.001).
        2   That gender differences will not prove significant (with  the 
        probable exception of income differentials).   
           As  Appendix  D shows, males and females responded  in  almost 
        equal  numbers.   As  expected, there was  a  significant  gender 
        difference  in income (independent t-test p <.001).  Gender  also 
        correlated  with  job satisfaction (Mann-Whitney  p  <.001);  the 
        expected levels measure was not significant.  
           The  hypothesis  as  a whole is  supported  by  the  available 
        evidence.    Male   and  female  respondents   did   not   differ 
        significantly in terms of 'addfact' (independent t-test),  number 
        of mentors (Mann-Whitney) or in the proportion of those  mentored 
        (Mann-Whitney; statistics in Appendix D).
           The   relationship   between  gender   and   affiliation,   as 
        represented   by   the  social  orientation  measure,   was   not 
        significant  (Mann-Whitney).   The  results  were  supported   by 
        separate analyses of mentored and non-mentored individuals.

        3   That  the  levels  of mentoring  will  vary  greatly  between 
        professional and working class careers.  
           The  null  hypothesis may not be rejected here.   Of  the  404 
        respondents, 125 were blue collar workers, 279 white collar.   Of 
        the 179 respondents (44.31%) who claimed to have had  significant 
        personal  influences in their lives, 56 were blue collar  (44.8%) 
        and 123 were white collar (44.09%).
           The  differences between the occupation categories on  quality 
        of  mentoring  (addfactors) and duration of  the  most  important 
        relationship  were  insignificant  (Mann-Whitney).   Blue  collar 
        workers claimed to have had more mentors (Mann-Whitney, p <.001).   
           Interestingly,  whilst  white collar  workers  were  generally 
        better  qualified  academically  (Mann-Whitney  p  <.0001),  blue 
        collar  workers  were better paid (Kolmogorov-Smirnov p  <.02  2-
        tailed).    Differences  pertaining  to  job   satisfaction   and 
        subjective expectations proved insignificant.   
           Of  particular interest for its implications for hypothesis  1 
        is that blue collar workers with no academic qualifications,  who 
        had  also  reached  above their  expected  level  of  attainment, 
        claimed  to  have  had  significant  personal  career  influences 
        (Spearman, p <.02 2-tailed).

        4   That official mentoring will not appear to be widespread, and 
        that  it  will not be viewed as particularly  beneficial  by  its 
           Of  the  404  cases, 61 (15.09%)  had  had  formal  'mentors'; 
        another  61  had  had  'buddies' or  some  other  such  allocated 
        assistance.  Correlations between these categories and  perceived 
        usefulness were not significant; respondents' opinions appear  to 
        have  been supported by insignificant correlations  with  income, 
        job satisfaction and career level assessments (Spearman).
           A negative correlation with the existence of informal  mentors 
        occurred  in  both cases (r = -.1249 p <.02, r = -.0984  p  <.05; 
        Spearman  2-tailed).   Comments from those who  had  found  these 
        formal roles useful generally revealed that they were essentially 
        induction procedures, run for a few weeks or months, rather  than 
        ongoing career assistance.   This seemed particularly  widespread 
        in  the blue collar occupations, and rather reflects the view  of 
        the  British  commentator, Audrey Collin, that the mentor  is  an 
        agent of socialisation.  
           Are Kram's higher order factors valid in Britain?
           The  factor analyses conducted in this study tend  to  support 
        Kram's  general grouping of factors into psychosocial and  career 
        oriented  higher order factors.  The initial statistics  for  the 
        factorised group as a whole were as follows:

                 Eigenvalues             % of variance   cumulative %
        VC1      7.14827                 37.6            37.6
        VC10     2.14487                 11.3            48.9
        VC11     1.32133                  7              55.9
        VC12     1.05072                  5.5            61.4
           After  rotation,  the first factor  contained  5  psychosocial 
        elements (mainly 'counselling') and 2 apparently  career-oriented 
        elements.  The latter, however, comprise sharing of the  mentor's 
        history with the protege (question 14) and similar attitudes  and 
        values  (18); these particular elements, chosen by Noe (1988)  to 
        represent the coaching and role model factors could quite  easily 
        be interpreted as being fundamentally psychosocial in function.
           The  second  factor comprises 4 career  oriented  elements  (3 
        exposure/visibility and 1 challenging assignments).
           Factor  three  comprises  6  career-oriented  elements  and  2 
        psychosocial  elements.   Again, classification seems to  be  the 
        source  of  the apparent discontinuity:  the  'counselling'  role 
        involves  discussions of competence and promotion (question  13), 
        with  'acceptance/confirmation' being encouragement to behave  in 
        new  ways  on  the  job' (16).    Interpreting  these  as  career 
        oriented   rather   than  psychosocial  does  not  seem   to   be 
        constraining facts in favour of theory unduly.
           The fourth factor comprises three career functions.   
        Factors correlation matrix:

                    Factor 1    Factor 2    Factor 3   Factor 4
        Factor 1    1.00000
        Factor 2     .23823     1.00000
        Factor 3    -.47601     -.28740     1.00000
        Factor 4     .28727      .32986     -.24768    1.00000
           The   strong   negative  correlation  between   factor   1   - 
        psychosocial  - and factor 3, a strongly career-oriented  factor, 
        are also very suggestive of Kram's higher order division.
           When  the  white  collar  occupational  segment  was   studied 
        independently, Kram's higher order functions were supported  with 
        even greater clarity.      

                 Eigenvalues             % of variance   cumulative %
        VC1      6.30568                 33.2            33.2
        VC10     2.31200                 12.2            45.4
        VC11     1.35617                  7.1            52.5
        VC12     1.19400                  6.3            58.8
        VC13     1.01946                  5.4            64.1
           Factor  1 comprised 4 psychosocial elements, with question  14 
        again; another apparently career-oriented element is question 18, 
        pertaining  to similar attitudes and values.  The 5  elements  in 
        Factor 2 seem to cover coaching and advancement; the inclusion of 
        questions 13 and 16 again do not seem to contradict this.  Factor 
        3  is  career  oriented, comprising exposure  and  visibility  (3 
        elements)   and  challenging  assignments  (1).     Factor   four 
        comprises  two   role  model  elements,  the  fifth  factor   two 
        protection elements.   

            The  Blue  Collar  category,  however,  shows  a   remarkably 
        different topology.  The initial statistics are as follows:
        Question       Eigenvalue       % of variance   cumulative %       
        VC3            7.14418          55              55
        VC6            1.12195           8.6            63.6
        Correlation -.56988
           Both  factors,  as shown in Appendix H, contain a  mixture  of 
        psychosocial    and   career-oriented   elements    (in    Kram's 
        formulation);   the  breadth  of the  mixture  does  not  suggest 
        blurring  of interpretive classification.   The nine elements  of 
        Factor  2 all feature amongst the thirteen in Factor 3;  all  are 
        negative    correlations,    presumably    indicating    bi-polar 
           Although gender divisions show fewer changes from the topology 
        of  the whole sample, differences do exist:   Initial  statistics 
        follow for the men.
        Question       Eigenvalue       % of variance   cumulative %       
         1             7.43755          39.1            39.1   
        10             2.20094          11.6            50.7 
        11             1.35305          7.1             57.9 
        12             1.07563          5.7             63.5
        13             1.01525          5.3             68.9
           Factor  one has 2 unambiguous career-oriented  elements,  role 
        model  (question  17, example of how things should be  done)  and 
        sponsorship  (7, promotion of career interests).  Factor  two  is 
        exactly the same as that for the whole factorised sample.  Factor 
        three is very similar (minus question 8; plus 10 and 7).   Factor 
        four  retains the 'protection' elements, 7 having  been  subsumed 
        into the previous factor.  Factor 5 contains two  career-oriented 
        elements  (6  and 8).  As in the main matrix, there  is  a  large 
        negative correlation between factors one and three.
        Factor correlation matrix:
                    Factor 1   Factor 2   Factor 3  Factor 4  Factor  5
        Factor 1    1.00000
        Factor 2     .21636    1.00000
        Factor 3    -.49934    -.21376    1.00000
        Factor 4     .21271     .18390    -.16940   1.00000
        Factor 5     .23983     .01944    -.21288    .05559   1.00000
           106  of the 123 blue collar workers (86%) were men.   Although 
        causation  is  unclear,  the relationship  between  worktype  and 
        gender (Chi-square .45894, p >.00001) suggests that the nature of 
        the blue collar jobs may be an influential variable in explaining 
        the  difference  in factor 1, given the great disparity  of  this 
        category  from all other factor analyses, the even gender  spread 
        in  the  study as a whole and the solid adherence  of  the  white 
        collar   sample  to  the  sample  norm.   A  discussion  of   the 
        qualitative  responses  to  formal mentoring  (hypothesis  4)  is 
        supportive of this argument.
           Initial statistics for the female category are as follows:

        Question       Eigenvalue       % of variance   cumulative %       
         1             7.04670          37.1            37.1              
        10             2.33318          12.3            49.4
        11             1.39170           7.3            56.7
        12             1.26948           6.7            63.4
        13             1.06172           5.6            69
        Factor correlation matrix:
                  Factor 1   Factor 2    Factor 3    Factor 4   Factor 5
        Factor 1  1.00000    
        Factor 2   .23175    1.00000 
        Factor 3   .09018     .03620     1.00000
        Factor 4  -.42860    -.34951     -.10744     1.00000
        Factor 5  -.07441    -.01403     -.04889      .05991    1.00000
           The  first  two  factors, responsible for  half  the  variance 
        covered  by  the large Eigenvalues, are similar  to  the  overall 
        mentored   sample.   Differences  occur  in  the  lower   factors 
        (Appendix H holds the details).
           Will the mentor typically be a hierarchical superior?   
           This was generally the case:
        line managers / supervisors            49.2%
        more senior managers                   20.3%
        older / more experienced colleagues    15.8%
        colleagues generally                    5.6%
        very senior managers (e.g. directors)   2.8%
        others (e.g. consultants, wives)        6.2%

           As  the  line chart in Appendix C indicates, it  is  generally 
        within  the  line manager's ability to provide the  best  quality 
        mentoring (by the addfacts measure).
        Additional results
           The  social orientation measure correlates with the  existence 
        or otherwise of a mentoring relationship, as predicted by  Coles' 
        theory of receptivity to mentoring (Mann-Whitney p <.001).

           The  support  for  the first  hypothesis,  that  mentoring  is 
        related to outcomes, reflects much of the literature referred  to 
        earlier.   This result appears to have generalised from  American 
        studies  of  formal mentoring to a primarily  informal  system  - 
        perhaps  this  is  rather an oxymoron - in Britain.   It  is  not 
        surprising  that  factor analysis reveals a closer  adherence  by 
        white  collar  occupations to Kram's (1985) dichotomy  of  career 
        orientated and psychosocial factors; this classification has more 
        in common with the American respondents. 
           Although  the blue collar category was originally included  to 
        examine  class differences (which appeared insubstantial  in  the 
        light  of hypothesis 3),  the results pertaining to it  may  well 
        have  a bearing on a regular problem of cross-sectional  studies: 
        the  direction  of  causation.   This  perennial  is  of   course 
        suggested by the correlation between qualifications and outcomes;  
        have qualifications or accompanying impressions of being a 'high-
        fliers'  attracted  mentors.  Whilst this may  be  a  confounding 
        variable in the white-collar occupations, this appears to be less 
        likely  in  the blue-collar category, where a  less  academically 
        qualified  contingent  seems  not  to  differ  significantly   in 
        mentoring  levels.   The correlation between expected  levels  of 
        attainment and claims of significant personal career  influences, 
        when  pertaining  to those without any  academic  qualifications, 
        does rather suggest that 'who you know matters more than what you 
        know' at certain stages of some careers.

           Unfortunately,   no  such  additional  evidence  allows   such 
        commentary  to be applied to the relationship  between  mentoring 
        quality (as measured by 'addfacts') and subjective measures.
           Results  relating to gender differences appear to reflect  the 
        American studies.  No differences in mentoring outcomes occurred; 
        the  usual income differential, however, was also accompanied  by 
        one  relating  to  job satisfaction.   No  great  differences  in 
        mentoring experiences emerged from the factor analysis.
           Differences in the male responses to factor analysis  probably 
        reflect  the situation in blue-collar occupations.  A  difference 
        occurs  in  the topology of the mentoring, although not  in  non-
        factorial  measures.  Kram's factors do not appear to  cross  the 
        class barrier.  The resultant 'feel' of the analysis (c.f. Child, 
        1990,  on the reliance on subjective judgement relative to  other 
        statistical  methods)  is one of a very personal attention  to  a 
        protege's  advancement; perhaps division between advancement  and 
        interpersonal issues is more characteristic of white collar  work 
           This may be interpreted as supporting Kram's assertion  (1985) 
        that  relationships  providing both both kinds of  functions  are 
        particularly beneficial for development.   The complete  blurring 
        of definitions in the blue collar classification means,  however,  
        that  the  model  has no construct validity in  this  domain  and 
        should not, therefore, be used for extrapolations here.

           Kram's  model is supported generally, however;  its  extension 
        beyond both country of origin and the usual respondent population 
        is  impressive.  The strong negative correlations between  strong 
        psychosocial and career-oriented factor clusters are particularly 
        indicative  of her superordinate factors.  Her claim (1985)  that 
        career  functions  are  generally more prevalent  may  be  domain 
        specific, however, given the formal systems in the U.S.  In  this 
        study,  Factor 1, with the largest share of the variance by  far, 
        has a predominance of psychosocial loadings.
            British employment - if the Norwich survey is  representative 
        -  does  not  appear  to support formal  mentoring  in  terms  of 
        influencing  careers beyond organisational  socialisation  (which 
        would  appear to be primarily task-related).  Given the  research 
        on  efficacy  of  formal and informal systems  (Iles  and  Mabey, 
        1993),  this may not be problematic.  Many of the  recipients  of 
        formal  mentoring did seem to appreciate its value as a  tool  of 
        induction.   It  just  does  not  comprise  the  general   career 
        influence  portrayed  by  Kram and supports  the  assertion  that 
        British employers are only likely to organise systems for meeting 
        short-term needs.   
           Whilst  Iles  and  Mabey  cast doubts  upon  the  fairness  of 
        informal allocation, this study would suggest that the likelihood 
        of  becoming  mentored  lies  somewhat  in  the  personality,  or 
        possibly conception of work,  of the would-be protege.  A measure 
        of  social  orientation  that  was designed  to  look  at  gender 
        differences in affiliation needs (not significant) was also  used 
        as  an  indirect measure of receptivity to  mentoring.   This  is 
        supportive  of  Judith  Coles'  observations:  her  own   mentor, 
        although  invaluable to her, was not 'discovered' by  several  of 
        her  colleagues; when she herself became a mentor, some used  her 
        experience, others 'did not want to know'.  
           In  terms  of  mentor  characteristics,  the  pre-eminence  of 
        hierarchical  superiority  is  in  line  with  Levinson's  (1978) 
        conception of the mentoring role.

        Textual problems
           One of the flaws in the questionnaire itself was the original, 
        ambiguous,   classification  of  the  global   mentoring   scales 
        questions  by  Noe (1988).  As noted  earlier,  the  respondents' 
        answers  generally clustered meaningfully; Noe's  categorisations 
        sometimes  lacked  validity, failing to measure that  which  they 
        purported to measure.
           In  making alternative interpretations, the researcher may  be 
        accused  of, at worst, constraining the facts to fit the  theory, 
        or - still problematic - making comparison with previous  results 
        rather difficult.  Future replication is not a problem,  however, 
        given  the  explanation in the text of which  elements  were  re-
           The  researcher's own questions left something to be  desired.  
        These  were  of low reliability.  A proper pilot  study  was  not 
        conducted  (a result of temporal and fiscal pressures).  The  job 
        satisfaction  measure  was so crude as to be unworthy  of  debate 
        relating to theory.  The only excuse that may be offered is  that 
        the  need  to study the other focal points without  producing  an 
        unwieldy questionnaire rather encouraged this dabbling.  

           The  use of an affiliative measure - and a riskily  colloquial 
        one at that - to indirectly measure receptivity to mentoring  was 
        of dubious validity.  It really covered attitudes to work. It was 
        shown  to lack consistency and further study would be  needed  to 
        see  what  is the underlying factor correlating with  take-up  of 
           Some  textual  errors succeeded in surviving the  small  pilot 
        scheme.  At one point, the text refers to influences on  careers; 
        more  than  one respondent pointed out that some  influences  are 
        malevolent,  the very opposite of the benevolent mentoring  under 
        examination.  A worse error was the 'definitely ot at all' at the 
        crucial  stage  of  the study where  the  respondent  decided  on 
        whether  or not he or she had been mentored.   Fortunately,  most 
        respondents  underwent  the same Gestalt  interpretation  as  the 
        pilot proofreaders (perhaps being used to Likert scales); most of 
        those  who did follow the instruction literally  gave  subsequent 
        answers  which  explained where they stood, but  this  error  was 
        certainly unhelpful and led to discards of questionnaires.
           Certain areas of study were discontinued.  Data to be used  to 
        examine changes in mentoring over time and age-related  qualities 
        of  mentoring  relationships were rendered  inaccessible  by  the 
        database  files  becoming corrupt; the researcher  had  made  the 
        mistake  of  loading a software update  immediately  it  arrived, 
        rather than trying it out on a fresh piece of work.
           This  technical  problem  also led to an  inability  to  count 
        elements arising from qualitative responses (e.g. how many formal 
        mentors were seen as useful in the realm of induction).
        Analysis problems
           Strictly  speaking,  significance  levels  should  have   been 
        determined  before statistical tests were conducted, rather  than 
        using the readout from a computer program.
           Given more time, the researcher would have consulted about the 
        usage,   interpretation  and  presentation  of  factor   analysis 
        results.   No explanation was given for the choice of an  oblique 
        rather  than  an orthogonal rotation  technique.   Similarly,  no 
        sensible  rationale  lay  behind the  choice  of  structure  over 
        pattern matrices.   
           Worse  was the switch to a pattern matrice when analysing  the 
        female sample, making comparisons with other topologies a  rather 
        dubious undertaking.  The idea of doing this for closer adherence 
        to the original statistics defeats the object of the rotation, to 
        give derived solutions.
           The smallness of the mentored sample of 'blue collar' workers, 
        coupled  with the reliance on post office workers,  firefighters,  
        and  police  officers, may have biased the results in  favour  of 
        workers in large (and generally uniformed) organisations.    More 
        generally,  the overall sample may well have been  representative 
        of  the city's working population, but the  convenience  sampling 
        did  not  give  random access to  different  levels  of  employee 
           As  noted  earlier  in the study, Norwich  was  not  an  ideal 
        population  for  hoping  to extrapolate to  other  areas  of  the 
        country, let alone elsewhere.
           Whilst  the  method of data collection coincided with  a  good 
        response  rate  for  questionnaires  without  follow-up  (Bailey, 
        1978), the use of both personal distribution and first class mail 
        for completed questionnaires makes it hard to ascertain which was 
        the  effective  variable.  Also, the distribution  process  could 
        only be replicated fully by the researcher.

           Mentoring is related to outcomes; this result appears to  have 
        generalised  from American studies primarily of formal  mentoring 
        to informal mentoring in Britain.  Similarly, Kram's (1985) model 
        of the nature of mentoring relationships is supported.  
           The domain into which her model does not generalise - the blue 
        collar occupational category - provides useful evidence about the 
        more  general  outcomes.  The blue collar sample did  not  differ 
        significantly  in general outcomes, only in  factorial  responses 
        about  the  mentoring  process,  where  the  distinction  between 
        career-oriented  and  psychosocial factors  seemed  non-existent.   
        Similar  correlations  existing  between  mentoring  and  outcome 
        measures,   without   the  confounding   variable   of   academic 
        qualifications, are supportive of a theory of causal direction in 
        which  mentoring  influences  progress (as  opposed  to  being  a 
        product   of  other  forces,  merely  accelerating  an   existing 
           Gender findings replicate those of other studies.  Differences 
        are insignificant, except in the expected areas of income and, in 
        this  study,  job satisfaction.  Male differences  were  probably 
        attributable  to phenomena within the blue collar  sample,  where 
        men predominated.
           In  terms  of  mentor  characteristics,  the  pre-eminence  of 
        hierarchical  superiority  is  in  line  with  Levinson's  (1978) 
        conception of the mentoring role.
           Formal mentoring appears well suited to induction, its current 
        usage in Britain.  This study suggests a relationship between the 
        up-take of informal mentoring, with its broader implications, and 
        the  receptivity of the potential protege; social orientation  at 
        work was the measure used.
           Whilst  respondents tended to be scathing about the notion  of 
        formal  mentors as serious influences on their careers, they  did 
        seem  to appreciate their value in learning the ropes.   As  this 
        system works, it shouldn't be changed.
           Given  the apparent efficacy of informal  mentoring,  however, 
        encouragement  of  this would probably be of value  to  employees 
        (and arguably other parties).  If formal methods are  introduced, 
        then  mutual choice of persons in the relationship may  eradicate 
        some  of the negativity of responses to current  formal  systems.  
        Perhaps more realistic would be an ethos of mentoring; London and 
        Stumpf  (1986), for example, recommend that people aged  55-75  - 
        'young elders' - should be viewed as valuable resources, who  may 
        be  encouraged  to become mentors, also  providing  expertise  in 
        those  areas  in  which they  are  not  rendered  technologically 

           The  point about receptivity to mentoring  suggests,  however, 
        that a more general education about the nature of work (here, the 
        possible  extrinsic  benefits  of social  interaction)  would  be 
        valuable.   Perhaps such an intervention would best be made at  a 
        level  of  pre-vocational careers work.   An  accompanying  ethos 
        would  need  to  be  in  existence  at  workplaces,  however,  if 
        disillusionment is not to be the primary outcome.

           Such  practical suggestions assume that the  current  research 
        undertaken  is reliable, with a broad base of external  validity.  
        Areas of further research could include similar research in other 
        areas of the country and elsewhere, preferably with a larger  and 
        more  representative survey of blue collar workers.  It would  be 
        particularly   desirable   to  know  whether  or  not   the   one 
        generalisation failure in Kram's model of mentoring processes  is 
        embedded  in  the British (or even  Norwich)  context.   Analysis 
        could also examine alternative existing models of mentoring (e.g. 
        Schein, 1978) applied to this occupational category; new  factors 
        may have to be sought, however.
           Also,  research should address the question of whether or  not 
        social  orientation is the key to receptivity to  mentoring.   An 
        alternative  to, for example, factor analysis, may be  to  devote 
        the  bulk of a questionnaire to a variety of attributes and  work 
        attitudes which may correlate with mentor take-up.

           Clearly,  the work could not have been undertaken without  the 
        participation  of a large number of people totally unknown to  me 
        personally,  workers in Norwich of various kinds.  During a  time 
        when  my confidence was low, I was particularly heartened by  the 
        interest, assistance and encouragement of Bill Edmonds,  Training 
        Officer with Norwich City Council.
           As   my  references  to  her  in  the  text  suggest,  I   was 
        particularly  helped by the ideas of Judith Coles, Proprietor  of 
        DeskToPs, Prince of Wales Road, Norwich.
           My tutor, Dr. Jenny Kidd, put up with the vagary of trying  to 
        timetable the equivalent of almost 18 months research into about  6 

           Bailey, K. D. (1978)  Methods of Social Research.  Free Press: 
        New York.
           Berlew,  D.  E. and Hall, D. T. (1966)  The  socialisation  of 
        managers: effects of expectations on performance,  Administrative 
        Science Quarterly, 11, 207-223.
           Child, D. (1990)  The Essentials of Factor Analysis.   London: 
           Clawson, J. (1980)  Mentoring in managerial careers.  In Derr, 
        C. B. (ed.), Work, Family and the Career: New Frontiers in Theory 
        and Research.  New York: Praeger.
           Collin,  A.   (1979) Notes on Some  Typologies  of  Managerial 
        Development  and the Role of Mentor in the Process of  Adaptation 
        of the Individual to the Organisation.  Personnel Review, 4,  10-
           Dreher,  G.  F. and Ash, R. A. (1990) A Comparative  Study  of 
        Mentoring  Among Men and Women in Managerial,  Professional,  and 
        Technical Positions.  Journal of Applied Psychology, 75, 5,  539-
           Erikson,   E.   H.  (1959)   Identity  and   the   life-style.  
        Psychological Issues, 1, 1-171.

           Fagenson,  E.  A.  (1989)  The  Mentor  advantage:   perceived 
        career/job experiences of proteges versus non-proteges.   Journal 
        of Organizational Behavior, 10, 309-320.
           Fagenson, E. A. (1992) Mentoring - Who Needs It?  A Comparison 
        of  Proteges'  and  Nonproteges' Needs  for  Power,  Achievement, 
        Affiliation,  and Autonomy.  Journal of Vocational Behavior,  41, 
           Hofstede,  G.  (1980)  Culture's  Consequences:  International 
        Differences in Work-Related Values.  Beverley Hills, Calif: Sage.
           Iles,  P. and Mabey, C. (1993)  Managerial Career  Development 
        Programmes:   Effectiveness,  Availability   and   Acceptability.  
        British Journal of Management, 4, 103-118.
           Kram, K. E. (1985)  Mentoring at Work.  Glenview, Ill.: Scott, 
           Levinson, D. J., Darrow, C. N., Klein, E. B., Levinson, M. H., 
        and  McKee,  B. (1978)  The seasons of a man's life.   New  York: 
           London,   M.  and  Stumpf,  S.  A.  (1986)    Individual   and 
        Organizational  Career Development in Changing Times.  In  D.  T. 
        Hall and Associates, Career Development in Organizations, Jossey-
        Bass: London.
           Noe,  R.  A. (1988) An Investigation of  the  Determinants  of 
        Successful    Assigned   Mentoring   Relationships.     Personnel 
        Psychology, 41, 457-479.
           Olian, J. D., Carroll, S. J., Giannantonio, C. M., and  Feren, 
        D. B. (1988)  What Do Proteges Look for in a Mentor?  Journal  of 
        Vocational Behavior, 33, 15-37.
           Roche, G. R. (1979)  Much ado about mentors.  Harvard Business 
        Review, 57, 14-28.
           Scandura,  T.  A. (1992)  Mentorship and career  mobility:  An 
        empirical investigation.  Journal of Organizational Behavior, 13, 
           Schein, E. H. (1978) Career Dynamics: matching individual  and 
        organizational needs.  London: Addison-Wesley.
           Whiteley,  W.,  Dougherty,  T. W. and  Dreher,  G.  F.  (1992) 
        Correlates of career-oriented mentoring for early career managers 
        and professionals.  Journal of Organizational Behavior, 13,  141-


CareerSteer – careers test for career choice