i A Hea a aaa HE a Biot ak deepal renege st Petts Cogn ihe nea en ca A Theory of Generativity and Its Assessment Through Self-Report, Behavioral Acts, and Narrative Themes in Autobiography Dan P. McAdams and Ed de St. Aubin Department of Human Development and Social Policy Northwestern Universi Two studies describe the development _genertvity features of concern. acti and narration, A seléreport scale of generative concern — asseSSmentSTaeg (0p i esigne the Loyola Generativity Sale (LGS)-~exhibited good internal consistency and retest rehability and showed strong postive associations with reports of actual generative acts eg. teaching a skill) and themes of generauvity in narrative accounts of important autobiographical episodes, In 1 ‘sample of adults between the ages of 19 nd 68. LGSscorevof fathers were higher than thse af men ‘who ad never had children, ‘The purpose of this article is to provide a conceptual and methodological framework for the scientific study of generax tivity Over 40 years ago, Erik Erikson (1950) introduced the concept of generativty in the context of a life-span theory of personality d “is primarily I 2220 nesters moses te pot tivity versus stagnation is the psychosocial center- piece of the seventh of eight major developmental stages. the stage loosely associated with the middle-adulthood years. In the ideal Eriksonian scenario, the adult approaches the issue of| sgenerativityafterresolvingearlier developmental issuesofadult- hood: identity versus role confusion (Stage S)and intimacy ver- In generativity the adult the next generation while generating life products and ou ‘comes that benefit the social system and promote its continuity from one generation to the next, ata collected in Study 2are included ina master’ thesis by Fd ‘Aubin for the Department af Psychology. Loyola University of Chicago. The preparation of the report was aided by a grant from The Spencer Foundation to Dan P. McAdams at Northwestern Univers ‘We thank the graduate students at Loyola University of Chicago who helped design and validate the Loyola Generatviy Scale: Rachel Al- brech. Barry Hoffman, Tom Nestor, and Dinesh Sharma. Thanks are also extended 10 Gina Logan. Karen Rambo, Janet Shlaes, Carol Stowe Joe Vero and three anonymous reviewers for their comments ‘onan early drat ofthe arte, Correspondence concerning thisartile should be adresed to Dan McAdams, Department of Human Development and Social Policy, Northwestern University, 2003 Sheridan Road, Evanston, tIlinois 0208) From Eton pot of vie sn ESD in that parents are ately in ences oimed ther own offspring. But not ll parent are especially peers tive, Erikson maintained. and generatvity is by no means lim- ited to the domain of parenthood. One may be generative ina yme of Eriksons most compelling cxam- nls of genera sppageipipachobceaphical explore ton ot lag ier (Erikson, 1958) and Mahatm: Er “- their most ther than in the private realm of friends and family, Despite Erikson’ provocative analyses of generativny in case studies (see also Erikson, 1975, 1976, 1980, 1982) and despite the general upsurge of interest during the past two decades ‘among social scientists in adult development, research and theorizing on generatvity have been scattered, sparse, and un. systematic. Browning (1975), Kotre (1984), McAdams (1985), and Peterson and Stewart (1990) have made theoretical state- ‘ments about generatvity that appear to expand on and depart from certain Eriksonian notions, but no attempt to organize these statements into a framework has been made, and no sys- tematic theory of generativity has been offered. A few studies have supported the general idea that generativity s indeed salient preoccupation for many American adults (Ryff & Heincke. 1983: Ry#f& Migdal, 984: Vaillant & Milofsky. 1980). Others have assessed the relations between generativity and cer- tain personality dispositions (McAdams, Ruetzel, & Foley 1986: Van de Water & McAdams, 1989). For instance, Me- ‘Adams et al. (1986) showed that generativty ratings of adult interviews were positively correlated with the sum of power and. intimacy motivation assessed on the Thematic Apperception. Test ‘ery litte attention has been paid to the problem of messur-1004 ing individual differences in generativity. To assess generativity directly, previous studies have used global clinical ratings that were based on the researchers’ reading of Erikson (Snarey, Kuehne, Son, Hauser, & Vaillant, 1987), simple self-raings (Ry & Heincke, 1983), or standardized personality scales de- signed to assess traits that seem to be components of generativ= ity such as dominance, nurturance, and breadth of interests (Ryff & Migdat, 1984), Ochse and Plug (1986) reported a 10- item self-report scale for generativity embedded in a large per- sonality inventory purporting to assess each of Erikson’ fist seven stages. A similar measure has been developed by Hawley (1985), embedded in an assessment of all eight Eriksonian stages. Neither ofthese two short scales was designed with at tention to problems of discriminant and convergent vali ‘Thus, in both cases, scores on generatvity are highly correlated with scores on many other stage scales in their measures— scales that purportedly measure very different constructs. Fur thermore, neither scale has been used in a systematic program of research on generativity designed to validate the measure and the construct. A Theory of Generativity Itisour belie thatasystematic research program centeredon ‘complex construct like generativity must begin with a clear and integrative theory: What follows is a brief and schematic ouline of the theoretical framework that guides our research nd partly informs the development of initial assessment pro- cedures. The theory draws from Erikson’ scattered writingson sgeneratvity, but it also departs significantly from Erikson on some key points. Inaddition, it brings together descriptivestate- rents about generativty offered in the writings of Becker (1973), Browning (1975). Kotte (1984), McAdams (1985, in press), and Peterson and Stewart (1990). Generativity has been variously described as QING as eencouchedin ems HEE {@ drive to reproduce onesel), in normal growth).and: adult into a productive ni identified with behav- jor (Child rearing), with motives and values (concern for pre- serving what is good and making other things bette), and with a general attitude toward life and the world (having a broad perspective and understanding one’s place in the sequence of generations). It is our belief that, unlike personality traits such as extraversion (Eysenck), developmental stages such as formal ‘operations (Piaget), and social processes such as causal attribu- tion (Heiden, generativty is not readily construed as a single, structured concept located “within” the individual. Its rather more like the construct of attachment (Bowlby) and certain ‘other relational and multiply contextualized constructs that re- {uire the scientist to operate simultaneously on a number of different levels and to take into consideration the particular relation or fit e attachment id exists in a psychosocial space that subsumes person and environment, ‘We view generativity as configuration of seven psychosocial DAN P. MCADAMS AND ED pe ST. AUBIN features constellated around the personal (individual) and cul tural (societal jing for the next generation. Figure 1 displays t and their proposed interrelations. ‘tne fet i ; Ohare ened ieee ures ity, an bine fo promote. in the adult years, a forthe next generation, With the support ‘ofa belief (4) in the concern may stimulat nature of an adult generat iprocally influence be- liefand concern, as wel. If demand and desire refer to funds- ‘mental motivational sources for generativity then concern, be- lief, and commitment refer to the resultant thoughts and plans about generativity that the individual formulates in the adult re aio cnet eas the arrows in Figure | indicate. But ‘the most personally efficacious, psychologically fulfilling, and socially valuable behavioral expressions of generativity are of: ten the produetsof the adults reasoned commitments o genera- tive endeavors and goals. Thus, ideally generative action steras | comment Le >| [7] pevonite ory forme ese pentgencion| | deinons NR DESIRE symbolic teen (communion) (iezivational (Thoughts, plans) (ebavior) eect Preure 1 ie i or her generative inclinations. The feature of cultural demand in generativity therefore, encompasses a wide spectrum of factors and forces external to the individual. and ‘much ofthis domain remains unexplored in contemporary psy~ chological research ‘A second major source of generativity is inner desire, Gener ty is frequently described in motivational c instinct, or dive, producing a desire or want = ‘defined gen: et ate as the “desi ub: stance in forms of life and work that will outlive the self” (p ei 22 ees Beer 57 describing generativty as partly the er 4 ‘defining lezacy that may be off ing generations asa gift, As uct (GoASIRICE eles that lve on. In addition, the generative Seven features of generatwity adult expressesa felt"need to be needed” (Stewart etal 1988, p. 56), a desire to nurture, assist, or be of some important use 10 ‘other people. These two desires would appear to ere in human Tves,, (ioe so aa Finer ee ee Me ‘ams 198; Wiis & Broughton 9SSETHE GAR fc mortality would appear to be one manifestation of agency, asa tendency to assert, expand, and develop the self in a powerful and independent way The desire ta be needed by others would appear to be one expression of communion, as the general ten- dency to relate to others in loving, caring, and imimate wavs, ‘even to be at one with others, ‘Cultural demand and internal desire combine to promote in lod ID: nes seein: Ts development out making a contribution to the next generation and inner desires for agential immortality and communal nurturance come together in adulthood to pro- mote the necessity or accident” (1964, p.131). Adults may translate their1006 DAN P. McADAMS AND ED be ST. AUBIN F mn has called a Erikson, 1963, p. 267; Van de Water & McAdams, 1989). This is a basic and general belie in the fundamental goodness and worthwhileness of human life specifically as en- visioned for the future. To believe in the (human) species isto ‘hood, McAdams suggests that idemtity development isthe ma- {or psychosocial issue forthe preponderance of one’ adult life- ‘ime, and generativity is incorporated within itas one of many different and important aspects. In the context of an evolving personal myth, an adult constructs and secks to live out a gener tivity tivity sriptis into his or her own personal veri ely and the social world he or ‘and deprivation. When such a beliefs she inhabits, and, in some extraordinary cases, into society’s lacking, the adult may find it difficult to make a strong com- ‘mitment to generative action, because it may appear that a gen- crative effort may not be very useful anyway. ‘Guided by commitment, which itself isa product of desire, concern, and belief, generative in any of three loosely related guises Franz. and Layton (1988) identify productivity as one of four main themes in generative content of personal documents, and McAdams (1985) emphasizes that generatvity, unlike simple altruism or general prosocial behavior, involves the creation of, ‘product or legacy * i n rorecting home, or church) that link generations and assure continuity over time (Browning, 1975: Erikson, 1982). Finally, generative behavior Becker, 1973, McAdams, 1985) tive father is both 2 self ageran ddiaing creator and a self-sacrificing giver. Biologically and so- cially, he creates a child in his own image, working hard and long to promote the development of that child and to nurture all that 1s good and desirable inthe child. But he must eventu- ally rant the child hisor her own autonomy, letting o when the time is right, letting the child develop his or her own identity, ‘make his oF her own decisions and commitments, and ulti= mately create those offerings of generativty that will distin= uish that child as someone who was"given birth to” in order to “give birth to” ‘The last feature of generativity We conceive of agenerativity within the larger cont “Adams's (1985, 1987, 1990, in press) life-story theory of adult identity Accord- i process of identity develop- the gradual consteuction and ‘successive reconstruction ofa personal myth integrating onc’ perceived past, present, and anticipated future while specifying ‘ways in which the individual fits into and distinguishes him- oF herself the social world. Rather than viewing identity s part of a psychosocial stage for ate adolescence and young adult- ‘own encompassing history. The generativity script, which may change markedly over the life course, addresses the narrative need in identity for a “sense of an ending” (Charme. 1984; Ricoeur, 1984), satisfying vision or plan concerning how, even ‘though One’ life will eventually end, some aspect of the self will live on through one’s generative efforts. The generativty script enables the personal myth to assume the form of “giving birth to" As Erikson wets in midlife and algae ingly likely to define him- or herself (0968, p.141), This article focuses on three of the seven features inthe pro- posed theory of generativity—generative concern, action, and ‘arration. We describe the development and use of(a)a 20-tem self-report scale. the Loyola Generativity Scale (LGS), which primarily assesses individual differences in generative concern: () a behavior checklist that includes a number of everyday actions that are suggestive of generativity; and (¢) a thematic assessment of critical autobiographical experiences that taps {into generative narration in one’s self-defining life story: The article thus provides a theoretical and assessment framework forthe study of generatvity and lays the groundwork for subse- ‘quent explorations of generativity from the multiple perspec: ‘ives encompassed in al seven features of the construct (Me: ‘Adams, Logan, & de St. Aubin, 1992), Study I: Construction of a Scale ‘The construction and validation of the LGS followed the general sequential procedure for developing self-report scales for personality constructs adopted by Jackson (1971: Jackson & Paunonen. 1980) and others (¢g., Wiggins, 1973). According 19 this procedure, a scale is developed with an eye toward both theoretical and empirical criteria, tems are rationally derived from theory; the item pool is then reduced and refined throug various empirical procedures that maximize internal consi tency and convergent and discriminant validity while minimi2~ ing the influence of response styles. Two samples were used in the current study—one containing adults ranging in age from 19 1068 years and the other containing a much more restricted range of college undergraduates. Results were also obtained ‘concerning (@) age and sex differences in generativity and (b) relations between generativity scores on the LGS and ones pa rental status, Method Subjects. The fst sample adult sample) consisted of 149 adults (6 rmenand83 women) ranging nage from 19 to68 years(M = 32.7 yea aGENERATIVITY 1007 ee eee crue sneer Fee of 1988 and were n0t Paid for their participation, All were Tesapen tesa weed amie of cial eens sce oo ence aa ioesecadeomen secaenaoe. amen pers 18. nd 20) 4s tens conc eR HEE illhavealasting impact and will proximately 60 orginal items generated by a team of researchers who ‘were thoroughly familia withthe following sources on generatvity: Erikson (1963, 1969); Kote (984): McAdams (1985); McAdams etal, (4986); Stewart et al. 1988) and Van de Water and McAdams (1989), ‘We sought to covera wide range of generative content associated with an individual’ concern forthe next generation with ems that were structurally simple and easy ro understand. A substantial number of, items were phrase in the negative (low score would indicate high ‘eneratvity) {© minimize the influence of social acquiescence re- sponse set Results GS. Daatiom bth smpleson the 3-tm version ofthe GS were aed delineate the best 20 ses o compose he fal vesion of the LGS Fist, desciptvestatsison cach oF the) tems on the LOS were shamined. Those inter ohh the subjects showed Title sara sn responding were ‘dropped from the final version of the scale. Second. each item scores cored wit te al seore forthe 3:tem UGS {Tow leme owing ow and sisal nonueean pa tol corltions were roped from the na eon oe UGS Corelations foreach of te LOS Heme andthe tl tore on the tw ote generat measures and om soil esaby wets a hed am were chosen th showed rately high coneltions wih ether or boh generat tales ignerally above. 30) and lowandnonsincant core tons with soca dearbity gene Below 20 The same proce for determining equality often va loved in Sothsamples The we sls reeled sing similar Pat tens of nercrraions in his ead “ihe Append ins the 20 bet tes rom te above ana ‘Thee composed the cre verso ofthe LGS sed ut stcy Ino cole and te alu sapien cach of ese cn showed eave de ariaiy neyo ih Corratons withthe tol LOS soe (high cerelaionseh Cxteral messes sgn Comeret ran) dee oped by Ochs snd Pg (98) and by Hest 98), and) tow corclaons wth he response se of soci ty (criminal). Cronbach ap cocice the 20 feme mth al LS veron wee caleuate br oth sa ple For the aduksample apa", rte college ape, Sip = 8 These ales sgt at he sel has igh ie The LGS tems cover many of he mos sem ideas inthe thera ie Porte tems ect concern Bo KAGE, on others, especially to the next generation (Items 1.3.12, and 19), pd 14). Two RARBE ier sa anctoy Mnreectn nese ass items, the original set of 39 items contained a fairly large num- ber emphasizing caring for other people (eg, nurturance). But ‘most of these were too highly correlated with social desirability to be included in the inal version of the scale. One result was1008 ‘that virtually all tems referring tocaringforchildren have been climinated (the one possible exception being Item I! on adopt- ing children). Thus, the scale cannot be said to “discriminate against” adults who do not have children, even though child rearing receives considerable attention in theoretical writings on generatvity ‘A factor analysis of the 20 LGS items was performed, sepa- rately, for each of the two samples. In both samples, @ first general factor of Positive Generativity (with loadings from a ‘number of positively worded items) accounted for 26% (adult) and 29% (college) of the variance, whereas a second factor of | Generative Doubts (with loadings from such negatively phrased items as 2, 14. and 15) accounted for an additional 10% (adult) and 93% (college) of the variance in LGS scores. Adult sample: Sex, marriage, and children. Mean scores for the 20.item LGS did not differ asa function of sex (men: M = 40.8, SD = 7.9: women: M = 42.0. SD = 7.0), Neither were sex differences observed for the generatvity scales of Ochse and Plug (1986) and Hawley (1985). Women did score significantly higher on social desirability however (147)= 3.36, p<.01. The GS was stconaly correlated with scores on the other two mea- sures of generativity rs(147) = .66 and 67, ps <.001; and only modestly correlated with social desirability (147) = 21, p < 0S. This isnot surprising in that the 20 items composing the final version of the LGS were explicitly chosen, in part, for their strong correlation with the generativty scales and theit ‘weak correlation with social desirability. Generativity concern was uncorrelated with age forthe entre sample and within the female subsample. However, among men, there wasa nonsigni- ficant trend, 768) = .23, p<.10, for generatvity tobe positively associated with age. ‘To examine the relations among sex, marriage, having chil- ‘ren, and generativity. a series of two-way analyses of variance (ANOVAs) was performed. Inthe fist set of ANOVAS, the vari- ables of sex (men vs, women) and marital status (married vs single; the data from 12 subjects who were either divorced or widowed were dropped from this analysis in that they were ‘currently not married, though they had at one time been So} served as independent categorical predictors of the continuous variables of generativity (as measured on each of the three scales) and social desirability. These analyses produced no sige nificant effects beyond the aforementioned main effect (posi- tive relationship) of sex on social desirability (women scoring, highen, In. second set of two-way ANOVAS, the variables of sex and having children (yes vs. no) were paired as independent predic- tors. A yes referred to being or having once been a parent of 2 child. These analyses yielded striking findings for the LGS (but not for any of the other measures) especially among men. The ANOVA yielded a significant main effect for having children, Figure 2 displays the important interaction effect. As can be clearly seen, the relationship between having children and gen- erative concern as assessed on the LGS is markedly strong among men but only mildly evident among women. Men who are or who have been fathers showed a mean generativity score 0f 45.2, ascompared with men without children, who showed a ‘mean score of 38.2, The comparable means among women are DAN P. McADAMS AND ED De ST. AUBIN 42.2 and 41.5, respectively These differences cannot be attrib- ‘uted to marital status, in that marital status was unrelated to LGS scores for both sexes. Rather, it is whether a man has ever bbeen a father to a child that seems to make a difference in predicting his generativty score on the LGS, The result isespe- cially noteworthy in that no items on the LGS explicitly deal with being a father and raising children, College sample: Sex differences. For the college sample, the intercorrelations between LGS and the other three selEreport scales mirror those obtained for the adult sample, Again, LGS was strongly correlated withthe generativity scales designed by ‘Ochse and Plug (1986) and by Hawley (1985). The correlation between LGS and SD was 10 (ng. Unlike the adult sample, lear sex differences emerged in the college sample on al four scales, as exhibited in Table . Thus, college women score signif cantly higher than college men on all three scales of generativ- ity including the LGS. Data from the adult sample and the college sample were ‘combined 10 assess differences between the two groups. A tW0- way ANOVA with sex and sample fadult vs. college) as the 1¥0 independent predictorsshowed significant main effect forsam- ple and for sex on the LGS. As can be seen in Figure 3, adult women scored the highest on generatvity, followed by adult ‘men and college women (whose scores were virtually identical), and finally by college men (who showed the lowest scores on LGS). The mean LGS score for the college men (37.2) was slightly lower than the LGS score for those men in the adult sample who did not have children (38.2), though this diference ‘was not statistically significant, ‘When the data from the two samples were combined, LOS scores proved 10 be modestly but significantly associated with age, 312) = .17, p <.05:and wit social desirability 312)~ 17, p<.05, Discussion Data from 149 adults between the ages of 19 and 68nd from, 165 college students provide inital evidence forthe viability of ‘a now self-report measure of individual differences in genera~ tivity concern, the Loyola Generativty Scale. The 20-item sale shows good internal consistency and has | though it ‘was unrelated to marital status per se. T ne een tad never exe ae open uecoreciea eor utss e Tigh pareve eon among mes'sspreianrar acon quence. or both, of parental status. Still, the results suggest the provocative possibility that having children is more intimately Intel wih mark genre mere ea eats wos Clery more esearcron tis tones ended Aon nec aes arr genre RG res US tre ames were conned ad Mstmploasa whic codsgnitoniy geen ne saty tun dt coluse sani The LOS pores tose promaing esse gens concern. Concern for the next generation may lead to gener szGENERATIVITY 1009 Have children Figure 2 Do Not Have children “Mean Loyola Generativty Seale (LGS) scores a a function of gender and whether the subject has (or has ever had) children (adult sample. sve commitments in an adult life and the eventual display of eneratve behavior. It would be expected, therefore. that LGS cares should be modestly but significantly associated with re- ‘ors of generative behavior. Study 2 assessed generative behav- vor dvectly through simple counts of real-life generative ats. In audition. themes of generativity were sampled in aviobiographi- al narranves to explore the extent to which the facets of con- fein, ation, and narration are interrelated in the broad do- ‘main of generativity. ‘Study 2: Behavioral Acts and Autobiographical Themes In Study 2, we obtained test-retest data on the LGS and examined how LGS scores for generative concern relate to n= Table Means and Standard Demations for Four Seales the Calete Sample by Sex Men Women Seale wv sD M sD + Loyola Generativiy Scale 37.2 BR UA 7524S" re & Pg (1986) Soe Desirably subscale 256 60 287 57 33008 Generatviy subscale 188-40 DSBS 43D Hawley (1985) Generatvity “subscale 886713 +p <5 "pel, ** p00 exes of generative behavior and ‘Simple counts of concrete behavioral acts are ‘useful in personality research, as both a direct assessment of some personality dispositions eg, Buss & Craik, 1984) and as external validation evidence for measures of internal behav- ‘oral dispositions. Although in the present study we clearly do rot adopt the view that generativity can be reduced to alist of discrete behavioral acts, we do believe that act counts ate one potentially fruitful approach wo assessing the action component ‘of generativty. and they can provide further evidence of the predictive validity of the LGS. The focus on autobiographical events affords a second. albeit subjective, perspective on action and allows one to assess generative narration through the con tentanalysisof personal experiences. Theassessmentof person alty themes in fantasy (McClelland, 1961; Winter, 1973) and autobiographical (Comis, 1947; McAdams. 1982, 1985) narra- tives has enjoyed some success in the history of the social Sciences. though few previous studies have examined themes of jgemeratvity in such narratives (but see Stewart eta, 1988). Method Subjects total of 23 male and 56 female subjects, ranging in age from 25 10 74 years (M~ 45, SD ~ 9-4, participated inthe study. The subjecis were obtained in two different ways, Approximately 25% of the subjects volunteered to participate an the sud by responding to ‘employee natices placed on bulletin hoards at two busineses(anarchi- ‘eetural firm and an insorance agency) in the greater Atlanta, Georgia1010 Generativity «0 (18s) 39 38 Figure 3 (adult ws. colege and gender area, We were able o draw on personal contacts fo recruit subjects from these firms, Participation was purely voluntary and subjects were not paid. The remdining 75% of the subyects were obtained through an urban midwestern unversty Students in introductory psy cology classes earned credits by partepating in research proyects or by obisining the parncipation oftheir parents. In ths study, parentsof students served a subjects. Pracedure Atte intial contact. subjects were asked to complete a small packet of measures that ncladed() the 20tem LGS described in Study) (b)a behavior checklist, and (6 autobiographical recllec- ‘ions (fn addon, thesebjects completed two measures of ego devel- lopment and personality tras, but these data are not included inthe present article) Pe subjects were asked to complete the measures in ‘het spate time and mail them back tothe researchers man enclosed stamped envelope, The subjects were then recontacied 3 weeks ater thee packets were ceturned and were asked to complete the LOS a second time. to obtein an estimate of test-retest eliabili “The behavior checklist consisted of 65 items phrased as behavioral ‘ets. OF the toa. 49 acts were chasen to suggest generate behaviors generative actscovereda wide spectrum andi ow bee A 1p and lage each at core sponded to one ofthe tree main behavioral manifestationsof genera: {iy creating, maintaining or offenng. Examples of acs purportedly unrelated to generativity included "begana det to lose weight” "reada nonfiction book,” "went to 2 musical concert” and "sent somebody flowers” (On the behavior checklist, the subject responded to each act by wate DAN P. McADAMS AND ED pe ST. AUBIN 2 oN ad o Mean Loyola Generativity Scale (LOS) scores asa function of sample specifving how often during the previous 2 months he or she had et formed the given ac, The subject marked a iP the act had not bee” performed during the previous 2 months, 1 ifthe at had been De formed once during that perio, anda2ifthe act had been performed more than ance during the previous 2 months. The researchers o& tained indvadual tem scores as well as composite scores ofthe see" tive acs umming across the 48) the act irrelevant vo genera") (summing across the 16), an total aes 168). ote thatthe Bith experience des 8" the subjets past but rather deserbes anevent that might happen sometime inthe future) For cath eP the subject was asked to desribe the episode in at last 8 WE? paragraph and to address the following questions: What happened ° the episode? When did it happen? Where did it happen? Who * involved? What were you thinking and feeling? What might this 6" sode say about who you ae, who you were, who you might be. oF POY ou have developed overtime Verbatim instructions or he five OF lections may be obtained from us) ‘We developed content anasissystem forcoding themes of ae tivity inthe autobiographical episodes. Fr each episode he Pres (score I) or absence core 0) ofeach of five generatvty themes *™ Biv themes are reference ro the subjects creating new product siatng projects. or generating new ideas, or desiring todos. ExaPhe> inciude “I wanted 10 ereate something that. . “sx copies © newly published book arrived. and". . bund a successful co pany” “MRR renew snes pai ft tonaRPaEsETTRG an Onging product, projec or tain ia GENERATIVITY eanplsot woezy,improrcoce, of cation of a Spence Beampls ude “T rs wor thing aon projet my wifeand Thad undertaken onacondo sen he pee cnc thetraon ut amy i am ssa Chota eso ne este to enege i ch Eng na nese monet _ Examples include “I wanted to provide her with comfort Se FE ey pail bat | sed any meceatons and sus {ion wat anything eae Day il any reference toa purposive and positive imerac- “ or individuals in 3 younger generation. Exam- seine mysoer els taee rie hiliren made picnic. . “Sand “lasked two of my it and our #0 oad any reference to leaving legacy having an 5 ing behind products that wil outive ones Poe they will be taking care ofthe planet longafter we ae ll, so ec clicve thatmsBook wil Become a pat ofthat istry eine peo lt my tie are penent odes, lind tal oer information about the se eed te autbrprapal epsoes fr themes of ener or ere summed ats thee and epaoes foreach ae ve ata total generative theme score foreach subject (Individual tear cpsode sores were ai stdin subeqen dita shi) then Say aclu aa crréaon between the tl Ince to ere The covelaon was, spgsing ih ‘Sonera gone Results total of 71 of the 79 subjects completed the LGS for the joncek retest Test-retest reliability of the LOS over the 3-week period was.73, p <.00}. suggesting moderately high temporal Stability, Mean scores on the LGS were not significantly differ. cnt for the twoadministrations(M = 39,83, SD=8,67 at Time and Mf» 40.17, SD = 8.65 at Time 2; ns. “To create a generatvity score from the behavior checklist, scores were summed ac70ss the 49 generative acts for each sub- ject. These gener 1p < 001. Corzelations were also caleulated for each of the 49 stemsas they related to LGS scores. Of the 49 individual ems assumed to suggest generativty 24 showed statistically signif cant correlations (p < 0S) with the LGS, and I| individual stems were significant at the p <.01 level. Correlations for the 11 items most strongly associated with the LGS are shown in Table 2. These 11 items may be viewed asa shortlist of genera the behaviors that are endorsed frequently enough to show a reasonable amount of variability sn the sample and that mamt- fest the strongest bivariate relations with scores on the LGS. ‘The correlation between the total score summed across these || items and LGS was extremely high, 174) =.75, p <,001 ‘The behavior checklist also yielded scores on 16 acts that were considered to be unrelated to generatwity. Summing across these 16 acts foreach subject, total scores ranged from | 1017 (M= 5.73, SD » 3.48). Total scores on unrelated acts were Positively associated with the total scores on the 49 generative acs, (96) = 38, p<.001., However, the correlation between the 1011 Table 2 Eleven Acts Showing Highest Correlations With LGS Scores Act CCorelation with LOS Drew upon my past expenences to help a person adjust toa situation. 560 2. Taught somebody a sil ‘ase 5. Served as role model fora young peson. ae 4. Told somebody about my own childhoos. a 5, Provided constructive entcusm about somebody's performance. a 6, Restored or rehabed a house, part of a house, piece of furniture, et 36" Did something that other people sonsidered to be unique and Important, 36 Made a decision that infuenced ‘many people 9, Performed a community serwce, 10 Produced a plan for an ‘organization or group ouside my ‘own fami so 11. Seed oF mended 2 garment ot other objects 20 Nate. LGS = Lovola Generatwity Scale, Spel,“ p<00 total score on acts unrelated to generativity on the one hand and LGS on the other was nonsignificant, (74) = .18, In addition, only 1 of the 16 individual items for acts unrelated to generat ity was significantly associated with the LGS. The item “Took ‘an out-of-state vacation” correlated with LGS, r(74) = .23, p <.0S. Thus. i would appear that the strong association be- ‘ween generatwity scores on the LGS and on the behavior checklist is not simply due to any tendency for the subjects Scoring high on the LGS to endorse more activities overall on the behavior checklist. More generative people are not simply Of the 79 subjects sampled initially, only 64 provided com- plete accounts fo all five ofthe autobiographical episodes re- quested. Therefore. 15 sul east one ofthe fiveexperi- ences blank. Response rates inthis regard ranged from a high of 73 complete responses for nadir experiences to a low of 66 complete responses for future experiences, Intercorrelations ‘among the five enerativity themes yielded 3 (out of 10) signifie cant correlations: 61} = 41, p <.01 between thematic catego- ries of oferimg and next generation. N61) = 40, p <.01 between ‘mainiaimng and symbolic immorality: and n61) =.25, p-<.0S between offering and symbolic ommortality: Total generative ‘theme scores were modestly positively correlated across the five episodes. although only 2 of the 10 intercorzelations reached Saustical significance (generativity as expressed in peak and nadir experiences—r[65] = 28, p <.0S—and generativity ex- pressed in peak and future experiences—r[63] = .25, p <,05). Total scores on generative themes summed across the five autobiographical episodes ranged from 41018 (M=8.75, I 2.93). The total theme scores were significantly associated with1012 oth LGS, 61) = 40, p <.01; and the sum of 49 generative acts, (61) = 45, p <.001; suggesting substamtial convergence among the three methodologically distinct assessments of gen- erativity. The correlation between generative themes and the short index of|| generative acts yas ith respect to particular autobiographical ‘episodes, total generatvity theme scores on nadir experiences showed significant associations with both the LGS and genera- tiveacts, Of the four other episodes, goal experiences correlated. significantly with LGS, whereas generative theme scores on peak, nadir, and future experiences all correlated significantly ‘with generative acts. Few sex differences were observed in the data, LGS and be havior checklist scores did not differ by sex. With respect to generativity vhemes in autobiographical recollections, men showed a slight but nonsignificant trend to score higher on ‘creating than did women (Ms = 2.4 and I.8. and SDs~ .87 and 1.26, respectively. (61) = 1.82. p <_.10, No significart age ef fects were observed. The LGS. checklist, and theme scores were all unrelated to age. A large percentage of the subjects in the study were parents (90%), making it impossible to examine the relations between generatvity and parental status. Genera- tivity scores on all measures were unrelated 10 the number of children in the subyeet’s family Discussion Data from 79 aduls between the ages of 25 and 74 years showed how three different measures of generawity appear to converge on the construct. Assessments ofa general generate concern (the LGS). of real-life generative acts displayed during a 2-month period (the behavior checklist), and of generative themes contained in narratives of autobiagraphtcal episodes showed statistically significant intercorrelations at .40 and Table 3 Cortelanions of Generatwity Themes in autobiographical Episodes With LGS Scores and With Generative Acts “Theme and type of episode Les Geenerauve at Theme Creating 2s Maintaining 3 Offering aie a Next generation 10 W Symbolic immorality 2M 2 Epuode Peak 20 sor Nadie 38° are Commitment 2 Goal ov Future Is oe Total 0 see Nite. LGS > Loyola Generativity Scale oye. “pe 00 DAN P. McADAMS AND ED p& ST. AUBIN higher. All three measures are themselves aggregates in which test items, generative acts, of theme and episode scares are ‘pooled to provide scores with maximal reliability and represen. tativeness (Epstein, 1979). Disaggregating the act and narrative measures revealed that particular single acts and certain theme and episode scores showed stronger relationships withthe LGS than did others. In addition, the LGS itself showed modesty tig eset cable eae oa ea eae. tendency for women to score higher than men on the LGS and for LGS scores to increase slightly withage. Study ? provided no support for these tends with any ofthe three assessment strate siesused. The assessment procedures ofStudy 2 did notsuggest a simple developmental picture for generativity. The study did snot provide any clear support for the common developmental notion, found first in Erikson. that generatvity becomes an increasingly salient psychosocial concern as one moves from ‘young to middle adulthood, One should be cautious in intet= preting this result. however. in thatthe sample for Study 2 w2s fot obtained in an ideal way for testing age or cohort differ- ences. The design emphasis was on the validation of measures rather than on the testing of age—stage hypotheses. No attempt was made. for instance, to sample representatively across the adult life span. Future esearch using larger and more carefully drawn samples i clearly needed. including. for instance. stu ies with random samples stratified by age (McAdams et a 1992), General Discussion ‘This article provides a conceptual and an assessment context fort 5 sco sess individual differences in generative concern. AmOae 1% especially. having been a parent was postvey associated wi scores on the LGS. fn Study 2, LGS scores signficanty dicted indexes of generative acts on a behavior checklist and themes of generativity in autobiographical narration. Thus. 38° sessments of generativity from the standpoints of concern. 3€ tion, and narration would appear to converge on the genet construct oF generativity from three diferent but related a> ales. ‘The assessment of generative narration using autobionraph cal episodes provides a personologically rich sampling of the meaning of generatvity in adult lives, If adults provide thet lives with a sense of identity through the autobiograpbics narratives they create and tell (McAdams. 1985, 1990). ther ‘would appear that generatvity runs through some life stories dominant anderen hematin eres 8°27GENERATIVITY ide eacnein ce, Th nih ae ae vied by «wo women inthe study The fist isa 29-year-old ov amen hth The fs ye sie meinen cei asp reins ie ace ee spenence On May — Lei ith tomy third cil who Pek meee son TR pair “ubor wa my mot dif Bboy pearl one nectar ree cut ater ad ue to complains wasn oxaen cat ns red once bor and ws prety 2m ald doctor and atoms wee preset 20 Mya esa and tis bor ced ae ei ale nd hy two cle Iwan hat 2a ei rey a, hen fm we bora et fet out bat second. My busband wat much mor excel ast 3 pon thouphtbe wold bee, my hsand, began hovng avec war hop Holiness tiny erm cress reaty amt hgs whith ately too tremendous fo write 98 and ct important becae fe he Benning of ie Taso and esponsiity mich wll ite, smother oft By contrast, 8 45-year-old mother of two describes a very diferent peak experience. This second woman scores relatively {bm on both the LGS and the generatwity behavior checklist Hor life-story high point concerns renewed love and intimacy, put no generatvity themes emerge: ‘eat experience About theee years ag0 Twas feeling dows, We fad just nosed toa bigevent it my life However, fet hat tmp husband. Band Twere defiing apart. was going o work in the city every day and 1 had developed a frendship wath La unger man at work-—oothing mote than a trendship. But 1 et that Band I wete drifting forther and forter apart and | did't Low wy even expressed the feling that B didn love me to my best fiend. C.Allshe could say wa that she was sure Bstil loved tne very much but I thought she was just trving 10 make me fel Beret Later that day {got home snd B was home-—nd our chi {ren werent, We started otal (which s unusual because Bis not fnuch of take) He tld me that be had talked to C earlier. I Scems he was jealous of my relationship with Land was Feeling Serv insecure T can express the happiness felt when I learned that Brill loved me—in fact he was ealousofL He thought was “native enough to ave this younger man intrested n me—and tne with him. l was elated natorall Tassured B hat I loved hitn 4nd him only but [was certamly on cloud mine. My husband of twenty years Was actually ealous—and he til loved me a much {sever We nentoutthateveningfortne most romanticdinacr ve erhad Tn this account of @ nadir experience, a 50-year old woman turns the death of her husband into a story of rebirth and renewed ‘generative commitment, including themes of ezeating, main- ‘Gaining, and next generation. Her account shows that genera tivity 1s as much a matter of creating products and outcomes that benefit the self as itis of creating things that will benefit other Nadir experience: On my husband dred a age $6 of a heart attack. [fet angry, frightened and git for being angrs My bus band had worked hard all his life and a die before retirement seemed to me like such 4 waste, Before my husband died, Thad worked parttime sometimeand mostly no time qtside the home. After about three months of dsiusionment and confusion, I de- 1013 ‘cided that I had no choice but 0 get of my dif and do something to help myself. Before my husband ded. he had purchased several piecesof property for income. There was noincome because alo the property needed repairing. Sol began a project. My two Sons, iy daughter. and myself have been repairing the buildings, One ‘uilding is eady tobe rented. I received a promotion on my ob. My daughter is back in college. My sons are working full me repairing the other two buildings. Weare not outof the woods yet, tout we can seea light atthe end of the tunnel “The conceptual scheme for generativity integrates ideas found in the writings of Erikson, Kotre, Browning, Stewart, McAdams. and others. It rejects the concept ofa discrete stage of generativity im adulthood and substitutes the notion that gen: ‘erativity may become a salient issue for adults as they grow ‘older because of increasing cultural demand. It contends that generativity cannot be understood from. a single personal oF Social standpoint, but that it must instead be viewed as a pS)= chosocial patterning of demand. desire, concern, belief. com- ‘mitment, action. and narration. While the seven facets of gen- erativity may shade into one another, they suggest multiple per spectives from which the complex construct of generativity may be viewed. No particular perspective is prototypical or essen- tial. Adults are generative in different ways, sometimes through their beliefs and concerns, sometimes through their commit. ‘ments and actions. and so forth, Ta comprehend the nature of generauivity in a given adult life, the unique patterning of de- ‘mand, desire, concern. belief. commitment, action, and narra- tion within that person’ life, situated in a particular social and historical context, must be assessed. interpreted, and evalu- ated, The two studies reported herein lay the groundwork for ini- tial assessment, interpretation, and evaluation by providing data from three different measurement strategies—selfreport ‘of generative concern, behavioral acts, and narrative themes in ‘autobiographical episodes. The three strategies would appear two hold promise fr further research applications asin studies of the manifestations and origins of individual differences in ‘enerativity. Nonetheless. other measurement strategies are needed if generativityistocommand the broad and deepempit- ical serutiny that this rich theoretical construct deserves, Links between generative desires (for immortality and the need to be needed) and other aspects of generativity (action, belief. etc} hhave never been explored in the empirical literature, The ways inwhich cultural demandscan promoteand hinder the develop- ‘ment of generative concern, belief. commitment, and action also await systematic empirical examination. With the excep tion of Kotre (1984), researchers have not attempted to under- siand how mature adults create meaning in generativity and how they engage in personal narrations of generatvity that as- ‘sume important positions within their own evolving stories of self and identity By thematically analyzing autobiographical ‘episodes. the current study makes a modest move in the direc- tion of Kotre's (1984) rich qualitative analysis. But more “thick description” of generativity is desperately needed in the scien- tific bterature What kinds of life stories do especially genera- tive men and women construct? What life trajectories do they specify? What formative experiences do they appropriate wai their seldefimng life stories co explain how itis they ‘came to adopt especially generative roles and values in life?1014 Future studies of generativity might do well toadopt both quali- tative and quantitative methodologies to explore questions such as these. ‘And there ae, of course, important developmental questions to address. Whether one adopts Erikson’ discrete-stage posi- tion or our contention that generativity arses gradually in adult- hood as a result of increasing cultural demand, most scholars would agree that generativity is a manifestly adult issue in the life cycle. Whereas children may be altruistic and prosocial, only adults are generative, we would contend. Only adults are entrusted with the responsibilities of caring for the next genera- tion, and only adults are aware oftheir place in the sequence of generations as producers (Stewart eta, 1988), creative rtual- ists (Browning, 1975), and keepers of the meaning (Vaillant & Milofsks, 1980). One could even argue that the desire for sym= bolic immortality (Kotre, 1984) and the need to be needed (Erikson, 1963)—the twin motivational desires in generativity —are not experienced by most children and adolescents with the kind of urgency and salience that theorists claim adults report. The currentaticle provides some conceptual and meth odological tools for examining more closely developmental hy- ppotheses about generativity: Although the current findings offer little support for any particular developmental viewpoint, a framework is offered for further eapioration of age and cohort elfects in generativity and of the development of different fea- tures of generativity over the adult lifespan. References Bakan.D (1966) Thedbaltraf man existence Holanomandcommae non wn Western man. Boston: Beacon Press, Becker. E. (1973), The denial of death New York: Free Pres. Browning, D.S. (1975). Generaine man Psychoanahte perspectives ‘New York: Dell Buss, DM. & Craik. K.H.(1984), Acts, dispositions. and persona. TnB. A. Maher & WB, Maher Eds, Progessimexperimental person ality research (Nol. 13, pp. 241-301). Sam Diego, CA: Academic Press (Charme.S.1 (1984). Meaning and mut nthe study ofles' A Sarvean ‘perspective Phladelphsa: Unversity of Pennsslvanta Pres ‘Combs, A, W 1947), A comparaivestudy of motivations as revealed in “Thematic Apperception stories and autobiography. Jowrmal of Clim al Picholngy. 3, 65-78, Epstein 5 (1979), The stability of behavior: |, On predicting most of ‘the people mich ofthe time. Journal of Personals and Social Ps chology 37 1097-1126 Erakson, EH. 1980), Childhood and covery New York: Norton. Erikson. EH. (1988), Young man Luther 4 studvon psvchoanalsteand Insior New York: Norton, Erikson, EH. (1963). Childhood and soviet (2nd ed). New York: Nor Erikson. EH. (1968), sight and responsi: New York: Norton Erikson, EH. (1968), kenny. Youth and cnsis New York: Norton Erikson, EH, (1969) Gandhtserah On the origins of milan noo lence, New York: Norton, Erikson, E. H, (1975), Life sore and the historical moment. New ‘York: Norton Erikson, EH. (1976), Reflections on Dr. Borg’ hfe cycle. fn E. H Enkson (Ed). Adulthood (pp. I-31), New York: Norton. Enkson.E,H (980). Themes of adulthood in the Freud-lung corre spondence: tn N. J Smelser & EH. Erikson (Eds), Themes of wrk DAN P. McADAMS AND ED pe ST. AUBIN «and love inadulthood pp. 43-74), Cambridge, MA: Har sity Pres. Erikson, EH. (1982). The fe cycle completed New York: Noro, Flavell. 1H. (982), Structures, stages, and sequences in cognitive de- velopment In W.A. Collins (Ed), The concept of development (pp 1-28), Hillsdale, Ni Erbaum. Hawley G. A, (1985) Construction and validation of an Eriksonian ‘measure of psychosocial development. (Doctoral dissertation, Uni versity of North Carolina at Chapel Hill, 1984) Dissenation ab sracs International, 45, 28084. Helson, R Mitchell, ¥, & Moane, G (1988), Personality and patterns ‘of adherence tothe sci clock. Journal of Personality and Social Paychologs 45, 1079-1096, Jackson, DLN. (1971). The dynamics of strvewred personality tests Paychologial Reve; 78, 229-248 Jackson, DN, & Paunonen, S. V (1980), Personality structure and assessment. dnnual Review of Psychology 31, $08-852. Kowre. J 988). Outlng these Generauty and the tepreation of live: Baltimore: Johns Hopkins University res McAdams D (1982) Experiences of itimacy and power: Relation: ‘ships between social motives and autobiographical memory Jowea! ‘of Persona and Social Psychology, 42. 292-302, Meadams.D-P (1983). Power inumacy and the ife stor. Person cal nguines mio ident: New York: Guilford Press. MeAdams, D P0987), A lifestory modelof identity: In R. Hogan W. Sones (Eds), Perspectres im personality (No.2, pp. 15-80). Green wich, CT JAL Press. Meadams. D P.(1988). Personal needs and personal relationships. In 5 Duek (Ed), Handbook of personal relationships: Theor: earch land interventions (pp. 7-22), New York: Wiles MeAdams, DP (1990), Unity and purpose in human lives The emes- sence ofidentny asa life sory In A. Rabin, R.A. Zucker. R.A Emmons, & 5. Frank Eds), Studion persons and ives (pp. 148-200), New York: Springer. Meadams,D.P (in pres). Personal mths and human Ines: How we find meaning ard purpose though stones. New York: Willi Morrow. McAdams, DP. Logan, G. & de St. Aubin. E_ (1992). Gener ‘among young. middle-aged, and older adulis Unpublished mam- McAdams, DP, Ruetel, K. & Foley . M. (1986). Compiexity and feneratvty at mid-life: A study of biographical script forthe Oi tute Journal of Personal and Social Psychology 0. 80-807. MeCleliand, D.C:(1961) The achieve society: New York: Free Press MeCrac. RR. & Costa. PT, Je (1990), Fesonaliym adalihood New ‘York: Guilford Pres. [Neugarten, BL. de Hagestad, G.0,(1976), Ageand the ie course 8 RH. Binstock & E.Shanas Eds), Handbook of aging and he social sevences (pp. 38-55) New York: Van Nostrand Reinhold cise, R & Plug, C: (1986). Cross-cultural investigation ofthe valid styof Erikson theory of personality development. Journala/ Pers aly and Socal Prvchologs: $0. 1240-1282 Peterson. BE, & Stewart, A. 11990), Using personal and fctionat documents toassess psychosocial development: A casestudyof Vem Britain gencrativity Psychology and Aging. §, 400-A1 Ricoeur (1984). Time and narraive. Chicag: University of ChicaBo Press. RyIEC.D. & Heincke.S (1983). Subjective organization of persons! vyinadulthood and aging Journal of Personaiyand Socal Psycho ony. 44 807-816 Ryfl,C. D, & Migdal S. (1984), Intimacy and generativny Sele ceived transitions. Signs. 9. 470-481 Snarey, 5, Kuehne, WS. Son, L, Hauser, S, Vail rd Unive. 1G. 6.0989) iGENERATIVITY lois xc of parenting in mes psychosocial development longi etsy dle ssa mee nea gi peril Pcl 23, S93-60, en. Fane C.& Layton (988). The changing self Ung Cr ——— Pete (Eds), Psvehobiography and Ie naraives (pp. 41~74). Dur- PER, INC Dake Univers Pres. ulin, GE, Milfs E0980), The natura history of mae atoll health 1X. Empirical evidence for Eiksons model af Bence eran Jounal ef rch 137134838, Van de Water, D.& McAdams, DP (1989), Generatvity and Erikson “belief in the species" Journal of Research in Persona: 28. 435~ 389, Wiggins, 1S. (1973), Personality and prediction Reading. MA: Ad: dison.Weses Wiggins, JS, & Broughton, R. (1985). The interpersonal circle: A structural mode for the integration of personality research. In Hogan and W Jones (Eds), Perspecves i personality (No. J. pp. 1-47), Greenwich, CT. JAI Press, Winter. DG. (1973). The power mote: New York: Free Pres. Appendix Loyola Generativity Scale | ty to passalong the knowledge Ihave gained through my exper {donot fel tha other people need me, [think would like the work of teaches. Teel as though Ihave made a difference to many people 1 do not volunteer to work fora charity. have made and crested things tht have had an impact on other people. ity tobe creative in most things that 1 do, [thank that Uw be remembered fora long time after Ide. "pelieve that society cannot be responsible for providing food and sheiter frail homeless people 10, Others wold say that have made unique contributions tosaciety 11. Wl were unable to have children of my om, 1 would ike 10 adopt children 12 Ihave important skills that] ty 10 teach others 13, fee that T have done nothing tha will survive ater die 1. in general, my actions donot have a positive effet on others 15. fel as though Ihave done nothing of worth to canribute to others 16, Tnave made many commitments to many diferent kindof peo- le, groups. and activities in my fe 17. Other people say tat lam a very productive person 18, Ihavea responsiblity tomprove the neighborhood in which ive 19. People come to me for advice. 20. feel as though my contributons will estate die, Received March 25.1991 Revision received September 6, 1991 ‘Accepted December 17,1991