My mission in this chapter is to examine the process of  innovation in the life course. Recognition of innovation is important because it helps society distinguish master patterns that have yet to be institutionalized from deviance that is to be punished and cultural variation that is to be accommodated with special programs. In the case of women’s changing roles, for example, the pioneer women who developed dual work and family careers during the 1960s and 1970s were treated as deviants and were subjected to resistance and discrimination. By the 1990s, however, it was widely recognized that women’s new life patterns were here to stay and that major social policies should take these shifts into account. Welfare rules began to require that indigent mothers eventually return to work rather than stay at home, and the Family and Medical Leave act of 1993 asked employers to provide temporary leave and job protection for workers who wished to take time off for family caregiving.

One of the first questions about innovation in the typical life course is how to recognize it when it occurs. How do the new life course patterns come about? How do wars, depressions, or longer life expectancy set the stage for innovation? How do myriad small changes in the lives of individuals suddenly add up to a sea change? Recognition of innovation can guide individuals as they chart new paths to adulthood, midlife, or old age. In addition, understanding that innovation has occurred can potentially illuminate some of the most pressing policy issues of our day such as how to live through old age happily or how to help dual earner families fulfill their parenting responsibilities. By studying social change and life pattern innovation quite deliberately, life course scientists can clarify which innovations are likely to become the norm and then document the changes in other institutions that will be required.

True innovation in the life course, according to Martin Kohli (1986), comes about when the temporal ordering of life events or the sum total of an individual’s roles not only depart from tradition but at the same time prefigure the institutionalization of that way of life so that it becomes the standard. The type of life course innovation considered in this chapter is distinguished from the biography of a pioneering individual by its eventual widespread adoption. A normative change in life course pattern is evident when the new pattern becomes institutionalized so that it is reproduced through the socialization process and cultural expectations. In the lives of women such major changes have been rather few and far between. Beginning around 1840, the ideal of true womanhood gained importance among middle class women who gradually withdrew from economic production on farms and in the new regional towns to lead the life of the "lady" (Welter, 1976). For nearly a century most middle class women preferred not to be employed or involved in public life but to manage a household, care for children, and represent the family in the community through volunteer work. After 1950, however, the spread of a service economy, longer lives, and lower fertility introduced millions of women to what historian Joyce Antler (1981) has termed "feminism as life process," the independence and potential for recognition made possible by expanded education, smaller families, and the promise of careers.

The question for life course scientists is how such momentous and long-lasting changes came about. What made it possible for isolated individuals to depart from the cultural script and chart a new path without knowing exactly where it would lead? And how did it happen that so many individuals made the change independently without knowing that they were part of a wider trend?

In recent years, the life course field has made new advances in recording massive changes in individual and family life and their connections to particular economic and political circumstances. But the findings have yet to be codified in ways that illuminate how innovation occurs. This chapter identifies methodological strategies for identifying and describing innovation as well as some of the main theories that have been offered to explain it.

ðØHISTORY OF THE QUESTION

It is first necessary to describe the background and significance of studying life course innovation -- what the major issues have been and how they are currently understood in relation to historical change and human development. Two of the best known examples of change in the normative life course are retirement age among males and rising labor force participation among females. Studies on age of retirement have made clear a continuing and profound interaction between institutional change and life course change.

As Kohli (1986) has shown, the institution of age 70 for pension eligibility in Germany in the 1880s and of age 65 in the United States (with the institution of Social Security in 1935) has reshaped the typical life course of men over this century. The decrease in older men’s labor force participation has been dramatic, dropping in Germany from 47 percent of men aged 70 and over in 1895 to only 5 percent in 1980. In the United States the pattern was similar, with a drop in the employment rate of men above age 65 from 68 percent in 1890 to only 19 percent in 1980. Moreover, the numbers of men ages 60-64 who were still employed also fell and between 1970 and 1988 dropped from 80 percent to 55 percent in the United States and even more in France, Germany, and the Netherlands (Kohli, Rein, Guillemard, & Van Gunsteren, 1991).

Employers who currently want to downsize their companies and move older people out of the work force use the workers’ own internalized expectations about retirement age and the pension and social security systems to accomplish their ends. Life has become compartmentalized into what Riley (1978) termed "three boxes", the first one at the beginning of life for preparation and training, the second in midlife for work and family, and the third at the end of life for leisure. Together these changes have solidified not only what Kohli (1987) has termed a "work society" but also a "tripartite" life course. In other words, the macro-structural rules of pension eligibility have induced changes in life planning while at the same time longer life, improvements in health, and job insecurity have reshaped the use of the pension system (Mayer & Müller, 1986). The new institutionalized life course thus represents two sides of the same coin.

This work on men’s retirement patterns is instructive for suggesting how to codify knowledge about women’s changing life course. A first step is to identify the key changes that have occurred and then to describe and explain them.

Up until the end of World War II, change in women’s lives was conceived primarily as winning the vote. Although woman suffrage had been proposed as early as the "Declaration of Sentiments" at Seneca Falls in 1848, it was not until 1920 with the passage of the Twentieth Amendment that women could vote in every state. A variety of factors produced the change: women’s higher education, women’s reform work, the need to represent the interests of families and children, and recognition of women’s patriotism and contributions during World War I (Giele, 1995).

Nevertheless, despite women’s having won the vote, the ideal economic role of woman as lady and homemaker had not deviated very much from the rule enunciated by John Stuart Mill ([1869] 1909, p. 75) that it was "...not...a desirable custom that the wife should contribute by her labour to the income of a family." Mill reasoned that children were better off and the management of the household superior when a mother did not have to divide her effort between gaining bread and providing for the comfort of her family.

Beginning in the 1940s, however, the picture changed markedly. Although in 1948, only 35 percent of women ages 25-54 were in the labor force, this proportion had almost doubled in a mere 25 years to 67 percent in 1983. The trends were even more dramatic for married women whose participation rose from 21 to 52 percent and for mothers of children under 6 whose rate increased from 11 to 50 percent during the same period (Giele, 1988, p. 300). By 1988 over half of all mothers with a child under a year old were employed at least part-time or for part of a year (Moen, 1992, p. 15). Komarovsky (1982), comparing the results from two studies of freshmen at Barnard College found that four times as many expected to have careers in 1979 as in 1943 (48 percent vs. 12 percent) and of those, all but 2 percent expected to be married and all but 14 percent expected to be mothers.

Working class women and black women had of course long had "multiple roles", in the sense of combining paid work with family responsibilities (Bell, 1974; Jones, 1985). What was different for the white middle class majority after World War II was a new legitimacy for paid work that allied it as much or more with women’s satisfaction and self-fulfillment as with economic necessity. Among black women too, work histories began to change. Gilfus and I found in our 1982 survey of three colleges that even though more graduates of the black women’s college had mothers who had been domestic workers, more of them had gained a postgraduate education and held jobs at high professional rank than their age peers at another women’s college and a coeducational college (Giele & Gilfus, 1990).

Corresponding with the changing employment rates and career expectations of women with family responsibilities was a changing age profile in women’s labor force participation. The profile was gradually changing its shape from that of an M-shaped curve (indicating women’s departure from employment and later reentry at midlife) to that of an inverted U-shaped curve similar to men’s that indicated more continuous employment through the middle years.

To pinpoint this change and to capture the fact that the change was not just a matter of rising employment but of combining women’s work and family roles, researchers and commentators began to speak of a rise in women’s "multiple roles." Women were also continuing their education beyond their late teens and early twenties. Davis and Bumpass (1976) found in the 1970 National Fertility Survey that over one-fifth of all women had attended high school or college since marriage, and the proportion was higher among younger women. Felmlee (1988) reported that the number of women over age 22 who were enrolled in college had more than doubled between 1972 and 1982. In 1980 at least one-third of all women college students were over age 25 (Rossi, 1980) .

After 1980 the description and explanation of these rising multiple roles became the main device for tracing change in the typical life course of women across historical periods, different age groups, and even different societies. Myrdal and Klein (1956) as early as the 1950s had written of Women’s Two Roles and the tension of combining responsibilities to home and work. Likewise using the two-role rubric, Moen (1992) updated the analysis in light of all the changes that took place from the 1960s to 1980s. A focus on "multiple roles" made it possible to link the impact of changing economic and family structure to women’s changing life patterns. One could also observe that the trend toward combining work and family roles created a demand for displaced homemaker legislation in the 1970s and more extensive provision of child care outside the home in the 1970s and 1980s (Burstein, Bricher, & Einwohner, 1995). Tough new work requirements for welfare mothers surfaced in the Family Support Act of 1988, and supports for parental leaves received legislative endorsement in the Family and Medical leave Act of 1993 (Wexler, 1997).

At first, in the 1950s and 1960s, these massive changes in women’s roles were analyzed primarily as social problems that created role conflict and role strain for the women who were trying to combine "male" and "female" roles. The best view of the role conflict perspective was provided by its critics. In The Feminine Mystique Betty Friedan (1963) roundly criticized the psychoanalysts and psychologists who assumed women’s innate need to look after families and to be subservient in the world of politics and economics. She ridiculed the "functionalist freeze" of the anthropologists like Margaret Mead (1949) and sociologists like Parsons and Bales (1955) who saw the division of labor between male and female roles as needed by society and therefore unlikely to change.

By the late 1970s, however, several scholars were beginning to see role combination in positive terms. Marks (1977), Sieber (1974), and Thoits (1983) began a theoretical revolution that identified the advantages of holding several roles simultaneously. Multiple roles could protect a role incumbent by preventing her from putting all her emotional eggs in one basket. The presence of alternative roles could also increase a sense of choice and satisfaction. Rather than being torn between the two roles of work and family, both women and men might actually benefit from being able to limit their commitments to "greedy roles". As Kanter (1977) argued in Work and Family in the United States, the experiences gained in the world of the family were useful to the world of work and vice versa. This reorientation among theorists then led to women’s simultaneous work and family involvements being recast as beneficial. In a major analysis of policy directions for women’s roles that was done for the Ford Foundation, I showed that a new paradigm of greater role crossover and flexibility was beginning to permeate American society (Giele, 1978). Women were gaining access to achievements and leadership that had once been reserved for men and men were being encouraged to take on some of the affiliative and caregiving functions that had traditionally been the province of women. Baruch, Barnett, and Rivers (1983) studied a sample of Boston-area women and discovered that women who were able to satisfy both their need for a sense of competence through work and a need for pleasure through family ties were better off psychologically (less depressed and feeling a greater sense of personal control) than those women who were limited to only work or family roles.

By the early 1980s the attention given to the dramatic changes in women’s lives came under a new rubric of "adult development" and "the middle years." The new focus was needed because human development theories (like that of Kohlberg [1969] on moral development) tended to stop with adolescence and leave unspecified what happened in adulthood. No one could really say whether the rapid and unprecedented changes in women’s roles were healthy or not. Levinson (1978) elaborated Erikson’s (1950) "Eight Stages of Man" and Sheehy (1976) popularized his work and produced several generalizations about the "midlife crisis" -- that it could be provoked by the loss of a parent, or of health and vigor, or a sudden feeling that life was finite and must be lived more intensely with perhaps a new marriage or a career change before it was gone.

Efforts to apply these insights to women revealed the limitations of stage theories and of the "middle years" as a conceptual category. The Social Science Research Council, under the auspices of its Committee on Work and Personality in the Middle Years, sponsored a scholarly year-long seminar during 1977-78 of experts on women’s lives at Brandeis University. One of the clear conclusions of the seminar was that stage theories were inadequate to the task of mapping all the changes in women’s lives and their consequences for society, families, and women themselves. The chief problem was that women’s changing roles were not only tasks given by the self as a project for adult development, but were being driven by a tremendous economic demand for women’s work, by the growth of higher education for women, and by increasing divorce and uncertainty about future family life (Giele, 1982a).

Demographers and sociologists began to suggest that a preoccupation with the middle years was "cohort-centric" in the sense that it concentrated on the problems of a particular generation born in the 1920s who happened to reach middle age in the 1970s. If the scheme was truly generalizable, why had midlife not been raised as a problem by earlier generations, and would it disappear for younger groups who were coming along? There was a distinct possibility that the midlife crisis was a transitional phenomenon that had come to that cohort of persons who were living through midlife in a society that was different from the one that they had expected (Rossi, 1980).

With the eclipse of role conflict and midlife stage theories, the foundation was laid for the ascendance of a new conceptual framework, the life course perspective. By the mid-1980s, changes in women’s lives had begun to take on a new meaning. They seemed to be a massive adaptation by women to profound trends in the larger society -- the rise of a service economy, increased numbers of female-headed families, and women’s own need to be ready in case the traditional age-gender system did not work for them. Coinciding with these actual social changes were developments within the field of psychology. The new life-span development psychology focused on positive adaptation to the aging process (Baltes, 1993). Thus the pioneers of women’s changing roles could be understood to have hit upon new more adaptive patterns for their particular circumstances.

A reinterpretation of age-related patterns was also occurring within demography and sociology. Historical period, social experiences of a birth cohort, and chronological age of the individual at the time of a significant life event were all seen as highly correlated but distinguishable dimensions that helped to explain changing social roles. The new sociology of age stratification suggested that novel women’s roles would be found among particular age groups who were located in distinctive social and historical environments and who had adapted to their situation by developing more effective or satisfying life patterns (Ryder, 1965; Elder, 1975; Riley, 1978; Swidler, 1986; Uhlenberg, 1979).

This new life course perspective was able to illuminate a much more complex dynamic of change in women’s lives than had been suggested either by role-conflict or adult development theories. The emergence of women’s multiple roles could now be seen as the product of changes in demography, economy, and family as well as individual life history.

Oppenheimer (1970, 1979, 1982) demonstrated the interconnection between historical period and cohort size in the conditions favoring innovation. Women flooded into the labor market in the 1950s because of rising demand in the female-labeled occupations of teaching, nursing, clerical work, and other service occupations. Yet the traditional supply of young unmarried women with the requisite education was unusually small due to the low birth rate of the 1930s. The upshot was a redefinition of who could be hired that included older married women who might reenter the labor force.

But within the eligible age groups to be hired, there were some who were ready and interested, others who were not. The difference had partly to do with their individual life experiences prior to adulthood. Elder’s (1974) work on Children of the Great Depression illuminated some of the relevant family dynamics that connected historical events and chronological age of the innovators. Those women who had grown up in the most economically deprived families, where their mothers worked and they had to do many of the household chores as teenagers in the 1930s, were more likely to become full-time homemakers than the girls whose families had not been deprived and whose mothers had not worked. Age at time of the Depression was also crucial: girls born several years later whose families also experienced economic deprivation did not show such marked effects in their career choices (Bennett & Elder, 1979).

Still there were women who married after the Second World War or who graduated from college during the 1950s who as young adults expected to follow the traditional homemaker path. But even these women had revised their course by the time they had reached their fifties in the mid-1980s. An unexpected divorce, a change in goals as a result of further education, or the desire to have other interests in addition to one’s family brought many women back into the labor force or back to school and caused them each independently to create their own idiosyncratic dual career patterns (Gerson, 1985). In a meta-analysis of fifteen longitudinal studies of college graduates collected by Hulbert and Schuster (1993) as well as of work histories of less well educated women in the U. S. National Longitudinal Surveys, I found that multiple roles had become a new norm, especially among those cohorts of women who were born in the 1930s and 1940s and who had together experienced the economic expansion, higher educational levels, and family insecurity of the intervening years (Giele, 1993). As shown in Figure 10.1, this interpretation traced the ways that the three dimensions of age were interwoven. The table designates historical period by the different study waves in the columns, birth cohorts in the rows, and the age of the subjects at time of each study in the shaded cells along the diagonal. Goldin (1995) using the National Longitudinal Surveys came to similar though less detailed conclusions about the increasing readiness of younger cohorts to enter the labor market.

[ Figure 10.1 Studies of Educated Women by Birth Cohort ]

Of course, one may wonder whether these trends toward multiple careers can be reversed. Although Easterlin (1978) hypothesized that "the good news about 1984" would be women’s return to the home in a burgeoning economy that made their employment less necessary, no turnaround has yet appeared. Since 1985 women’s employment has continued to rise to nearly 70 % in 1995 and over 80 % among college educated women ages 25-54. Thus women’s multiple roles appear to be truly an innovation in the normative life course of women.

• METHODS FOR STUDYING INNOVATION IN THE LIFE COURSE

How then should innovation be studied? To identify innovation requires solving three different methodological problems. First, one has to know innovation when one sees it. This requires a research design that compares successive age cohorts so that one can observe the start of an innovation and then trace its growth into a trend. Second, one has to collect comparable data from different age groups. This is more of a challenge than meets the eye because it is more difficult to get the same type of facts about life at age 35 from a 65-year old than from a 35 year-old. Having compared age groups and collected comparable data, the third step is to look for patterns in the life course. These patterns are diverse and complex and the nature of major trends is not always self-evident. Although there are still relatively few studies of life course innovation, it is nevertheless possible to articulate some general principles of research design, data collection, and pattern identification in the study of women’s changing lives.

Locating innovation in the life course usually requires some sort of comparison by age groups. Since change is the result of a tension between the programming by society of the life course and the individual’s own biography that expresses choice and agency (Kohli, 1986), one way to understand the connections between micro-level aging over the life course and macro-level institutional change is to observe the changing life patterns of successive cohorts. The study of cohort differences provides clues as to why change is occurring (Riley, 1987). As Ryder (1992, p. 230) remarks, "Cohort analysis is peculiarly appropriate for the study of long-term normative change, whether in reproductive institutions [the family] or elsewhere in the social structure.... Each new cohort is simultaneously a threat to social stability and an opportunity for social transformation." Inter-cohort comparisons permit one to contrast life events across age groups and identify the nature and extent of innovation. Intra-cohort comparisons make possible the analysis of why certain individuals in an age group become pioneers of change while others do not. These insights can then be extended to understanding why some age groups more than others are the originators of change.

Comparisons across age groups. Dennis Hogan's (1981) study of U. S. men born between 1907 and 1952 compared how men of different class backgrounds and ages experienced depression and war. The World Ward II soldier generation from blue collar and rural backgrounds tended to follow new patterns of getting married before getting a job or finishing their education. Featherman and Sørensen (1983) found a similar pattern of growing complexity in the major life events of young Norwegian men born in 1921, 1931, and 1941. The younger men, by combining more different roles simultaneously, were adapting to the complex demands of a modern society.

The comparison of different cohorts of women college graduates provides an analogous survey of change in women’s lives. Among Wellesley College alumnae who graduated between 1911 and 1960, I observed that the rise of a multiple role pattern began with the classes of the 1930s. They were the first group who at the time of the alumnae census in 1962 had combined marriage, motherhood, employment, and some post-graduate education over their life course. By contrast the most common life pattern among the classes of 1911-1915 was remaining single. Fully 25 percent of these early graduates did not marry but gained a post-graduate education and pursued a career as compared with only 20 percent who had been employed (if only briefly), gained further education, married, and had children as shown in Figure10.2 (Giele, 1982a; Perun & Giele, 1982).

[ Figure 10.2 Distribution of Major Life Patterns, Wellesley College Alumnae ]

The 1962 cross-sectional survey did not, however, make it possible to pinpoint when innovation occurred in the organization of the adult years because one could not analyze event histories by age and year of event. Event history analysis did not become possible until the Life Pattern Study of 1982. A mailed survey asked 3000 women alumnae from three colleges who graduated between 1934 and 1979 to report retrospective life histories of major events. Comparison across age cohorts revealed that a simultaneous combination of marriage, motherhood, and employment at age 35 together with having gained some post-graduate education only began to accelerate with the graduation classes of the 1950s (born in the 1930s). The timing of their marriages, childbearing, and post-graduate educational events was well ahead of the peak of the new woman’s movement in 1970, suggesting that the innovations of this age group may have actually helped to spawn the revival of feminism. Only 6 percent of the class of 1934 had a "multiple role pattern" at age 35, but this proportion had risen to 36 percent among graduates of the class of 1969 (Giele, 1987; Giele & Gilfus, 1990).

Life course comparison within age groups. Age cohort comparisons do not, however, explain the precise mechanisms that link history with individual development or change (Elder & O’Rand, 1995). For that we have to turn to within-cohort studies. With biographies or other life history materials, it is possible to compare women’s lives so that the connections between family background and subsequent life paths become apparent. Using archival materials in a comparison of 49 temperance and 49 suffrage leaders who were active between the 1940s and the 1920s, I described in Two Paths to Women’s Equality how the two pioneer groups were different from the norms for their day as well as from each other. All of the women were leaders and thus shared such background characteristics as coming from secure middle class roots and having an advanced education. But differences in their personal tragedies and frustrations helped to explain why they chose different reforms -- the temperance leaders more often having experienced a family loss; the suffrage leaders, a denial of professional opportunity (Giele, 1995).

Among investigations on contemporary women, a prime example of an intra-cohort study of pioneer and traditional behaviors is the qualitative analysis by Kathleen Gerson (1985) in Hard Choices. Gerson interviewed 63 women ages 35-44 who revealed the factors in their choice between a homemaking career or paid work. Significantly, Gerson did not use a cohort comparison to identify innovation. Instead, having identified growth in married women’s employment as a key trend, she held cohort, age, and historical period constant by choosing women of similar age and focused on events in their own personal histories that gave them reason to follow a domestic or non-domestic career.

Comparative and cross-national research. Just as intra-cohort comparisons can illuminate the individual differences that help to account for a particular life pattern, so cross-national comparisons of similar age groups can illuminate the macro-economic, social, and cultural conditions that foster or retard a particular innovation. Kohli, Rein, and others (1991) have used this methodology to compare advanced industrial societies in average age of men’s retirement. Women’s labor force participation and multiple roles can similarly be compared across countries to discover which social institutions and cultural factors facilitate women’s multiple roles.

The principal challenge to such comparative work is to find data sources on women’s lives that are set up in such a way that it is possible to compare changes in life patterns across similar cohorts. Fortunately, several national longitudinal panel studies can supply this need. For example, the German Socio-Economic Panel Study that began in 1984 provides biographical information on all subjects up to 1983 or the last year before they began to take part in the survey. When these data are arrayed to compare successive age cohorts in their frequency of occupying multiple or single roles at ages 25, 30, 35, 40, etc., the results are very similar to the American alumnae results and National Longitudinal Surveys in the United States. As shown in Figure 10.3, older birth cohorts were more likely to occupy single roles (marriage, or motherhood, or school attendance, or employment) whereas younger cohorts were more likely to combine roles across the public-private divide (working and/or education with marriage and/or motherhood) (Giele & Pischner, 1994).

[ Figure 10.3 Public-Private Role Combination by Age, West German Women ]

When contemporary women’s life patterns are compared in Germany and the United States using the German Socio-Economic Panel Study (GSOEP) for 1984-90 and the U. S. Panel Study of Income Dynamics (PSID) for 1983-89, it becomes apparent that the "static" factors in individual women’s employment patterns work in similar ways in each country. The employed women in both places are likely to have more education, fewer children, and lower initial income than the non-employed women. But the fact that the German women have a 10-15 percent lower labor force participation rate than the American women leads to a search for the dynamic factors that push their overall rate of participation upward. Analysis of differences at the macro level between societies reveals that German women’s lower employment may be due partly to more conservative gender roles in Germany as a whole. In 1989-90, only 11 percent of employed women in Germany had more than one child in the household compared with 28 percent in the United States. In addition, the Germans appear to have a broader safety net as revealed by the fact that as many as 67 percent received some sort of government benefit in 1987 (such as pensions, housing subsidies, etc.) compared with only about 16 percent of Americans in 1986 (Giele & Holst, 1997).

Having developed a research design that brings out period, cohort, and cultural differences, the next key challenge to the longitudinal study of pioneering behavior is the question of how to collect relevant data and how to organize the analysis so that some pattern is revealed. As shown earlier in this volume by Karweit and Kertzer, life course research treats life events as the basic building blocks for descriptive and explanatory analyses. Calendar dates and the person’s age are the markers attached to each event that help the researcher piece together connections between biographical threads and the historical landscape. Individual life processes can then be dynamically linked to social structure and changes in the social order. To study innovation, life history data should thus always include major life events, timing markers, and a construction of career paths or life histories that pieces together these discrete observations.

Nature of events. In study after study of women's lives, the types of events that are coded turn out to coincide with certain key role domains: (1) education, (2) employment, (3) marriage, and (4) parenting (Elder, 1974; Young, 1978; Marini, 1984; Otto & Call, 1990). A fifth domain that might be called public service takes different forms according to gender. In the case of men it is military service; in the case of women it is volunteer work and community service. Another dimension of these role activities is level of achievement or role attainment such as number of years of education, number of children, occupational status, and so on. In addition to the person's role activities, a record of major life events includes changes in household composition (Kertzer, 1986) and a place for unexpected events such as accidents or illness. Various devices can be used to record these events. A good example is Freedman's Life History Calendar that is reproduced in Scott and Alwin's chapter in this volume. The 1982 Life Patterns Study of women alumnae of three colleges used the retrospective Life Events Chart shown in Figure 10.4.

[ Figure 10.4 Age, Period, and Cohort Operationalized on the Life Events Chart ]

Timing markers. Chapter 1 of this book showed how timing of events is the common medium for linking human agency, social relationships, and geographical and historical location to account for the shape of the individual life course. Thus along with events and role transitions, the life history record must capture the timing of role entries and exits as expressed in two ways--person’s age and calendar month or year at the time of beginning or ending an event; entry into the event is coded as T1 and exit as T2. With these data points one can reconstruct not only the sequence and duration of the events but also the amount of role concurrence or simultaneity in the numbers of roles or events being combined at any one time.

Constructing "career paths" and life histories. Once the events data with timing markers have been collected, the next step is to construct career paths for an individual in each of the principal role domains and then to assemble these into a life history for each person. Because individuals of different ages are likely to have differing numbers of educational, employment or other events, data management using the conventional rectangular form with the same numbers of rows and columns for each respondent is an inefficient use of storage space. To respond to this problem two major types of data reduction technique have recently been proposed that take the form of either a "matrix" or a "chain." Together these recent developments represent a welcome ferment in the search for usable data reduction strategies that go well beyond the coding techniques and conceptual frameworks reported by Dex (1984). It is still too soon, however, to summarize their strengths and limitations.

The matrix form of life history is well suited to following a number of roles at once. Karweit and Kertzer (1986) use their CASA system to handle variable length files in each major role domain such as employment, education, childbearing, or coresidence. One can then construct a "sample path" of events for each person in each major role domain and examine co-events and antecedents and consequences. Otto and Call (1990) develop a special algorithm that constructs a unique number in each year for different combinations of multiple roles. Elder, Pavalko, and Clipp (1993), using examples from the Berkeley Growth and Guidance Studies and the Terman data, demonstrate how to construct life paths in parents’ fortunes, health, work, and other role domains and then to link these event histories to each other.

The chain form of data reduction focuses on the chronological ordering of events to emerge with several major alternative pathways. Abbott and Hrycak (1990) developed techniques for "optimal matching" of careers of 18th century German musicians to several major types of sequences in career events. Blair-Loy (1996) has used these techniques to construct five major career sequences for women who have reached senior executive positions in finance. Using data from the Wisconsin Longitudinal Study, Carr, Ryff, Singer, and Magee (1995) use Boolean sets to characterize the different chronological representations of life course histories that are associated with depressed, healthy, vulnerable, and resilient mental health outcomes.

The third methodological challenge in studying innovation is to identify it when it has occurred. The patterns seen by any given researcher are probably related to his or her particular analytic questions and investigative techniques. A researcher who traces interlocking events across many domains of a person’s life will probably be interested in how innovation is expressed as a different combination of roles. The researcher who codes for sequence is likely to be looking for new sequential patterns.

Modernity and multiple roles. Thomas and Znaniecki ([1918-20] 1927), the authors of some of the earliest attempts to write modern life history, focused on the changes in life experienced by the peasant who left rural Poland and was thrust into a modern urban society. They found a man named Wladek, on whom they focused a whole volume of description, a number of thematic trends: individualization, a search for efficiency, less traditional and fatalist ways of relating to occupation, to the division of labor between the sexes, to social class, and to neighborhood and community. Bauer, Inkeles, and Kluckhohn (1956), in their study of changing life patterns within the Soviet Union, and Inkeles and Smith (1974) in a study of modernity in several developing countries noted a new set of experiences and attitudes among the more modern: they were more likely to read newspapers, participate in egalitarian relationships with women, communicate with a larger world than the local community, and believe in flexibility, self-direction, and choice.

The comparable theme in women’s lives is less acceptance of the traditional sex-role system and an interest in equality and "feminism" that plays out in assuming part of the breadwinner role. Thus modernity in women’s roles has largely been studied through the adoption of multiple roles, meaning the combination of employment with marriage and motherhood. "Modern" women will usually have limited their fertility, gained further education, and delayed marriage as well as pursued employment beyond early adulthood. Adoption of multiple roles in this sense has been copiously documented on an aggregate level for the United States by Oppenheimer (1970), Smith (1979), and Spain and Bianchi (1996). Cross-national studies show broad similarities in the impact of modernization on women’s lives--greater longevity, fewer children, more education, and more paid employment (Giele & Smock, 1977; Kahne & Giele, 1992). Studies of college alumnae pinpoint the pioneer cohorts. Rather than focus on the coming of modern attitudes and personal traits, this research takes the adoption of multiple roles as the mark of innovation in the life course patterns of 20th century women (Gerson, 1985; Giele 1993; Giele & Pischner, 1994; Goldin, 1995).

New sequential patterns. Use of sequence to identify innovative patterns has been less prominent than efforts to link deviance from the normative life course to unfavorable outcomes (Hogan, 1978); but a decade of work suggests that the negative impact of a disordered life sequence is not at all clear (Marini et al., 1989). If anything, the normative trend may be toward a less clearly ordered sequence of life events overall. Modell, Furstenberg, and Hershberg (1976) in their study of Philadelphia youth in the 19th century discovered that a typical sequence of finishing education, gaining employment, and then getting married became less common in younger groups. Hogan (1981) found that such sequencing also became less common among veterans of World War II, especially among those from rural and blue collar origins, because they delayed employment and gained further education after having married. On the other hand, Blair-Loy (1996), in her research on American women financial executives, discovered that pathways to success became more rigidly ordered among younger age cohorts; it is now extremely rare for senior women to have worked part-time or to have spent time outside the labor force.

• EXPLANATIONS FOR INNOVATION IN THE LIFE COURSE

Whatever the patterns of innovation that are uncovered, the theoretical challenge is to explain why they occurred in one period rather than another, and in some groups and not others. The explanations proposed generally fall into three classes that relate to the basic paradigm of the life course presented in Chapter 1. The first type of explanation focuses on human agency and the motivations that characterize the pioneers. A second type notes the social linkages between a person’s intimate world and her distinctive life course adaptation. A third interpretation uncovers an association between cultural and temporal location and its impact on the life course. Life course patterns are seen as adaptations to circumstances, or what Thomas and Znaniecki ([1918-20] 1927) termed "the solution of a situation," in which the person reacted to objective conditions with pre-existing attitudes and defined the situation accordingly. Or in the words of Clausen (1991, p. 805), "The life course is shaped by the interaction of cultural and social structural features with physical and psychological attitudes of the individual and by the commitments and purposive efforts of the individual." Elder (1985b, p. 42) sums up the process as one of adaptation: "Crises may arise when claims are elevated well beyond control of outcomes. Novel adaptations to crisis situations are ways of dealing with resources and options in order to achieve control over the environment." Such efforts to regain control are prompted not only by the individual’s inability to produce life outcomes that are in line with expectations but also by large-scale historical events and processes such as depression, war, or migration.

At least four types of motivation are commonly given as explanations for why pioneers invent new life patterns: (1) a change in moral outlook or consciousness; (2) an attempt to meet economic needs and new threats to survival; (3) a conscious or unconscious search for improvement in quality of life; and (4) a need to cope with uncertainty. All of these factors interact and feed back on each other to reinforce traditional roles or stimulate new ones.

Changing moral outlook and consciousness. Beliefs are usually the most prominent reason given for why reformers mount their social movements. In the case of the woman’s movement of the past century, the largest single contingent were the temperance women who were drawn to a cause that would protect men and families from the ravages of alcohol. Woman’s temperance leaders were shaped by the moral outlook of small midwestern towns and the frontier Methodist, Baptist, and Presbyterian churches and reported more experience of personal family loss such as the early death of a child or husband. Suffrage leaders, on the other hand, were more likely to be from eastern urban centers or the west, from established Congregational, Unitarian, or Episcopal churches or the Quakers, and more often reported educational or career frustration. These differences were consistent with the ideologies and organizational structures and strategies of the two movements (Giele, 1995).

In research on contemporary women’s life course innovation, an egalitarian sex-role ideology (believing that wives can have careers and that mothers can work) appears to be an important foundation for a woman herself to be positively oriented toward a career (Bielby, 1978). Women’s greater numbers in the labor force have been explained as a grass-roots movement that began "in the hearts and minds of women" (Uchitelle, 1994) and that "broke the feminine mystique" (Nasar, 1992). In Carolyn Heilbrun’s (1986, p. 60) eloquent terms, writing about women poets, "Only in the last third of the twentieth century have women broken through to a realization of the narratives that have been controlling their lives. Women poets of one generation -- those born between 1923 and 1932 -- can now be seen to have transformed the autobiographies of women’s lives, to have expressed, and suffered for expressing, what women had not earlier been allowed to say" (p. 60).

Economic and survival needs. Probably the most prevalent reason given for women’s life course innovation is the economic necessity of work. Women of the 1850s-1870s sought women’s rights to control their own earnings in the face of an irresponsible husband or father who might squander their wages. During the Great Depression of the 1930s, women went into the labor force not only to supplement their income but to meet heightened standards of consumption; it was much more legitimate for them to work than for their children (Wandersee, 1981). Today women work for many kinds of economic reasons -- as a result of divorce, to put children through college, or to reach a higher standard of living (Oppenheimer, 1979).

Economists since the 1960s have understood women’s rising employment as the product of rising demand in women's occupations as well as women's increased education and higher wages whereby women's leisure became more costly and they chose paid employment instead. Women were more likely to be employed if they had higher education, less family responsibility, and less household income. Conversely, they were more likely to be out of the labor force if they had less than twelve years of education, were married, had more than one child, and an income above the mean (Mincer, 1962; Cain , 1966). Recent analyses have elaborated a dynamic dimension to the understanding of women's employment decisions. Change in marital status, taxes, and transfers result in a complex interaction of internal and contextual factors in the employment decision. Women who are already working are surprisingly insensitive to changes in the wage rate, suggesting that women work not only for income but for important intrinsic and non-pecuniary reasons as well (Mroz, 1987; Van der Klaauw, 1996).

Well-being, satisfaction, and personal control. An increasingly common explanation for women’s dual careers is the search for a better quality of life, by which is meant a greater sense of well-being, satisfaction, freedom from depression, and personal control. These motivations repeatedly crop up in the accounts of those women who have adopted multiple roles. Helson and Picano (1990) found that the Mills College graduates with multiple roles had better health and greater satisfaction. In a study of the college class of 1961 surveyed by the National Opinion Research Center at graduation and seven years later in 1968 when the respondents were roughly 28 years old, Baker and Sween (1982) discovered that the young women were deciding about how to coordinate family with career. Their action was based not only on earlier family/career choices but also on the extent to which their career activities had been rewarding. Since these changes in women’s life patterns were occurring before the new women's movement peaked in 1970, they suggest that a search for more satisfying life patterns may be part of the reason why women have gravitated toward the new multiple roles.

Numerous studies have documented links between multiple roles, greater sense of personal control, and better mental health. Barnett and Rivers (1996) find that the happiest women feel both mastery and pleasure. A sense of mastery is reinforced by competence and autonomy at work; pleasure, by strong emotional ties to family. Barnett and others (1991) conclude that multiple-role involvement nourishes women’s subjective well-being, because each role provides an independent source of positive experience, the benefit of which accrues regardless of the nature or quality of the other roles a woman occupies. In their follow-up of women studied in 1956 and 1986, Moen, Dempster-McClain, and Williams (1992) also found that occupying multiple roles as well as doing volunteer work was positively related to health and social integration. Heilbrun (1986, p. 17), in her study of women writers, touches on the issue of multiple roles in noting the traditional ideal of selflessness but also importance of activities in the public as well as private domain: "The hardest part of all for women to admit and to defend is that women’s selfhood, the right to her own story, depends upon her ability to act in the public domain" (p. 17).

Coping with uncertainty. The newest explanation for life course innovation in contemporary society is adaptation to the need for contingency planning and uncertainty itself. The trend away from traditional patterns is apparent on many different fronts--education, work, marriage, and childbearing. In the U. S. even rather traditional husbands encourage their wives and daughters to have an education and a profession as an "insurance policy" in the event that something unforeseen happens to the principal breadwinner. Women themselves can compare their situation and the society around them with the lives of their own mothers and see that conditions have drastically changed. Less time is needed to perform necessary household tasks. Nor can household production produce much cash income if there is no market worker in the family. Moreover, numbers of children have declined and time requiring a mother in the home has shrunk just as women's life expectancy has increased. In addition, it is becoming increasingly evident that time out of the labor force worsens chances for later reentry. These trends suggest that the multiple role pattern represents a positive adaptation to longer life expectancy and economic uncertainty, and not just the "need for two incomes." Thoits (1987) observed that uncertainty in the form of uncontrollable or involuntary events such as natural catastrophes, family crises, or plant shutdowns are likely facilitators of role bargaining because they destabilize established patterns and sometimes alter power relationships, thus making room for renegotiation and innovation. These new behaviors, when routinized, then begin to take on a normative power of their own. But the possibilities for role negotiation and innovation are more limited among those who lack financial power, education, and complex social networks.

Ott (1992) theorizes that uncertain conditions make household partners avoid a very high degree of specialization by gender because it is risky if they become separated or their contractual obligations to each other are weakened. If partners think ahead to time 2 while still in time 1, they avoid rigid specialization and opt instead for greater flexibility. Women’s multiple roles, to the extent that they allow both men and women to respond more flexibly to unforeseen events in work and family life, can thus be interpreted as one of the major adaptations to uncertainty in modern life. Or in the words of Thomas and Znaniecki ([1918-20] 1927, Vol. 2, p. 1906), "The individual must be trained not for conformity, but for efficiency, not for stability, but for creative evolution."

Throughout the new research on life course patterns as adaptation is a dual theme of diversity and generalization. Diversity most often arises in connection with black women’s roles. Many observers have noted that black women have always combined family activity with working for a living. Thus the new feminism associated with white women’s emancipation from the home hardly seems that "new"; in fact, black women seem to have pointed the way (Bell, 1974; Giele & Gilfus, 1990). Heilbrun (1986, p. 61) notes that black women writers escaped the dilemma of achievement vs. domesticity. She quotes Toni Morrison:

Social scientists like Patricia Hill Collins (1990) and Elizabeth Higginbotham and Lynn Weber (1997) have explained the different stance of black women toward multiple roles as the result of the legacy of slavery and racial discrimination that have always made the position of black men insecure. Thus black women do not expect marriage to support them. Instead they rely on their own work to be self-supporting and think of marriage as a separate undertaking. Latino women, on the other hand, are more likely to be working within the context of a traditional patriarchal culture that has expected women to be family caregivers and men to be heads of the family (Segura, 1997). When Mexican immigrant women go to work, they do so because their husbands cannot support the family. When Cuban immigrant women work, it is to help their husbands establish a business that eventually permits the wife to stay at home once again (Fernandez-Kelly, 1997). Similarly, Asian women immigrants work as domestics or service workers in the context of a traditional patriarchal family system (Glenn, 1986; Chow, 1997). Unlike African-American women who seem more independent in job-holding vis à vis their men, the Latin-American and Asian-American women appear to be following a trajectory of modernization in women’s roles that leads from menial work on behalf of the family to "lady of the house" when middle class status is first achieved and then perhaps to professional status in the third generation.

All of these differences among minority and majority women’s role adaptations are a signal that society comprises various subpopulations who accommodate to their situations in alternative ways. Thus, while the modal pattern may shift, a variety of other patterns can also be found. It is an open question which of these will turn out to be innovations that become institutionalized and which are transitional adaptations to limited resources and temporary conditions.

Another side of the coin, an opposite trend toward a more abstract sense of self and greater generalization of idealized life patterns, has also been noted. Meyer (1986) hypothesizes that the reason life satisfaction has become such a general goal in modern society is that the institutionalized life course that is demanded by the modern state is not necessarily connected to the subjective self. There is less psychological virtue attached to identification with family, local community, ethnic group, or religious denomination and the upshot is that individuals are more likely to identify with state, nation, or their work. The principle of the universalized subjective self becomes an integrative mechanism in a complex society that frees the individual from ascribed boundaries of family, ethnicity, religion, or even gender.

In 1996 and 1987, in the context of the new welfare law that imposes time limits on benefits and work requirements even for mothers of young children, the institutionalized life course works against diversity of adaptation and generalizes one pattern to fit all. A challenge to future life course research is thus to assess what may be the negative effects of such life course generalization by investigating overwork, underemployment, and the strains of dealing with contingency work and chronic uncertainty.

The effects on the life course of living in one cultural or temporal context rather than another are tested in different ways. Cultural impact is seen by comparing the reactions of people from different backgrounds to similar social and institutional settings. The impact of a historical period, however, is best understood through the adaptations of persons from similar cultural background who differ by age and cohort.

Cultural location. Cultural influence is evident in cross-national surveys of sex-role attitudes and women’s labor force behavior. A 1988-89 survey of attitudes toward women’s work in Germany, Great Britain, and the United States found that a majority in all these countries approved so long as the care of children was not an issue. But when asked their opinion about whether a woman should stay home or work if there was a pre-school child, the answers varied by country. In Germany 74 percent of women answered that the woman should stay home compared with 66 percent in the U. K. and 50 percent in the U. S. (Alwin, Braun, & Scott, 1992). Women’s labor force behavior shows similar variation in the twelve countries of the European Union. Although roughly 90 percent of all women with a graduate degrees are employed throughout the EU, there is much more variation when it comes to women who have only a compulsory level of education and a child under 14 living at home. Over 50 percent of such women in Denmark, Portugal, Belgium, and the U. K. are employed as compared with less than 40 percent in Spain, Greece, the Netherlands, and Ireland (European Commission, 1995). With Holst, I found in comparisons of American and West German women’s levels of employment that within the two countries the reasons why women were employed were similar. But between the two countries the 10-15 percent lower rate of German women seemed to be explained by differences in general cultural and institutional conditions that provided more public benefits to families and were less favorable to employment of mothers (Giele & Holst, 1997).

Temporal location. The same historical period and socio-economic climate can also have different effects depending on whether it is experienced in childhood, youth, or early adulthood (Ryder, 1965; Stewart & Healy, 1989). Events during childhood shape basic values and attitudes about gender role which derive primarily from the family of origin. Thus a girl growing up in a traditional family with her mother in a traditional homemaker role is likely to value that role and want to follow in her mother's footsteps. In youth when the person is constructing an individual identity, norms define the appropriate work, marital, and civic duties which are associated with manhood and femininity. These norms vary with the historical period and have a kind of moral force (Stewart and Healy, 1989). In adulthood economic conditions and employment opportunities shape a woman's actual marital, maternal, or work behavior. In good times as well as hardship, job opportunities can make a woman override her values derived from childhood about the traditional role of woman and encourage her to follow a new path that departs from that of her mother (Giele, 1993).

The evolving life patterns of college women display these principles. Each succeeding age group of women confronted the new life pattern of women at a younger stage of the life course with different effects on the model that they adopted. Thus women born in the 1910s and 1920s did not feel the impact of the woman's movement of the 1960s until they were already past forty, and so continued the homemaker pattern. Women born in the 1930s and 1940s were still in their twenties and thirties when they discovered women's changing roles, and so it was possible for them to forge a new dual career pattern. But for women born after 1950 the new woman's role was a reality by the time they left high school; they generally delayed marriage and childbearing to establish their careers.

• SUMMARY AND CONCLUSIONS

The importance of life course innovation is that it creates uncertainty and establishes new milestones for individuals living in changing times. At the same time it demands the adjustment or creation of social policies that will accommodate and institutionalize the new life patterns that have evolved. Although innovation in the life course has not been a clearly recognized object for research, there is considerable knowledge available about changes that have occurred in retirement age, the transition to adulthood, and the roles of women.

Two major conceptual advances have resulted from the analysis of women’s changing roles. First, the phenomenon when reviewed from a life course perspective began to be defined as innovation rather than deviance. So long as the issue of wives and mothers working was conceived as a problem of role conflict or of a mid-life stage, there was little progress in understanding because the problem was located in the women themselves--their role choices and their development. Only when women’s changing life patterns were reconceived as a positive adaptation to several interlocking circumstances (their own needs, the realities of the larger social and economic setting, and the demands of their immediate jobs and families) could the more complex analysis occur that would uncover the reasons that some became pioneers. Second, a specific indicator (the adoption of multiple roles) helped to pinpoint when and in what groups innovation had occurred. The dramatic change was concentrated in the period from the 1940s to the 1980s when women’s labor force participation almost doubled from 35 percent to 67 percent. The innovating pioneers were concentrated in a specific cohort born between the 1930s and 1940s; they invented their new life patterns during a particular historical period from the 1960s to 1970s and were in their early and middle adulthood at the time.

Using the life course perspective has brought new analytic methods to the study of inventive life patterns. The most useful research design has been age cohort analysis. Comparison across cohorts has helped to pinpoint the pioneers. Comparison within cohorts between traditionals and innovators has explicated the processes by which innovation comes about. The most illuminating data have been detailed life histories or longitudinal records of major life events. By constructing life histories from matrices or chains of significant life events that contain markers by age or calendar time, the researcher is in a position to analyze the factors in innovation that relate to age of the individual, historical conditions, and the cohort norms that have shaped the person’s response.

Out of the various interpretations of why life patterns have changed has come a clear understanding of the ways individual motivation, social structure, and cultural and temporal location shape a person’s life. There are many rich and persuasive accounts of the role that moral consciousness, economic need, dissatisfaction, or uncertainty have had in motivating individuals to devise new ways to live. However, even while norms shift to reflect new ways, there is simultaneous recognition that not everyone takes the same path and that there are important variants among different racial, ethnic, or socioeconomic groups. At the same time, coupled with diversity in modern society, there is a trend toward generalization and institutionalization of some tracks more than others, so that, for example, in the scheduling of women’s lives, it is becoming ever more common to expect mothers to work and to frame social policy accordingly.

Whether an innovation "takes" by becoming widespread in a given society or group is dependent on cultural and temporal location. The issue is the degree to which a change in the statistical norm will be converted into a change in normative expectations. Cultural location can either facilitate or retard the conversion. Even highly educated women with children who are very likely to work in some societies are less likely to in others because of more conservative sex-role ideology or the presence of alternative institutional structures such as widespread public benefits that supplement the male breadwinner role. The individuals’ temporal location (age at time of exposure to the innovation-producing conditions) can also hasten or retard adoption of new life patterns. Exposure in adolescence or early adulthood to the ferment that produces change is likely to breed a pioneer whereas for an older person the formative period will already have passed and, for a younger one, not yet come.

In the end, one asks how the study of life course innovation can benefit society and social policy so that people can live happier and more productive lives even under conditions of constant change. The research agenda to answer these questions is only now being formulated. What one learns from the remarkable recent changes in women’s lives is that multiple roles have brought better health and satisfaction to many. But we have yet to explore fully what may also be some of the less favorable correlates: overwork, less time for family and community, and continuing uncertainty about the future. We have also to expand the study of role innovation to other groups by bringing similar methods and analysis to men’s changing lives and the evolving life patterns of children, adolescents, and elders.

• REFERENCES