An Asian Comparative Description of Events in the Transition to Adulthood
An Asian Comparative Description of Events in the Transition to Adulthood
Peter Xenos (East-West Center, Honolulu, Hawaii)
Sulistinah Achmad (Lembaga Demografi, University of Indonesia)
Hui-sheng Lin (Bureau for Health Promotion, Department of Health, Taiwan)
Ping-Keung Luis (The Hong Kong Polytechnic University, Hong Kong)
Chai Podhisita (Inst.for Population and Social Research, Mahidol University)
Corazon Raymundo (Population Institute, University of the Philippines)
Shyam Thapa (Family Health International, Kathmandu, Nepal)
The Life Course and the Transition to Adulthood
The life course framework deals with the whole of individual lifetimes. As a frame of reference (and in some respects a theory) it can help us to recognize broad historical transformations and connect those to changes in event sequences and interdependencies within individual lives. Both formal demography and social demography are prominent in life course analysis and can even be said to form its very backbone. Conversely, life course change lies at the heart of the demographic transition, which is a social transformation that encompassing changes in longevity and survivorship and the complete reconfiguration of childbearing. Where ever it occurs, the demographic transition inevitably involves a longer period of adolescence no matter how that is defined (or in some settings its very emergence), a much shorter period of childbearing, a somewhat shorter period of childrearing, and a very much longer period of post-reproductive independent life, much of which occurs prior to old age.
Life course studies have sought to capture these broad changes in historical terms and beyond that have sought to link personal changes with these group experiences. Specifically, there has been effort to distinguish transformations of "personal time" and "family time." Another important theme in life course studies is the causal interdependence and interpretative meaning of personal histories-defined as meaningful, self-explaining biographies. These studies generally consider the whole of the life course, or large segments of it, and track sequelae which may be played out over many years. For example, there are studies of the lifetime, long-term sequelae stemming from disadvantages suffered by female children in their years from conception up to adolescence.
The life course issues of primary concern in a study of youth center on one important transition among the many-the transition to adulthood. It is well recognized that childhood and adulthood are different, but research and theory grapple with the fact that the transition is complex, involving multiple events which for each individual may or may not occur, and in sequences that may vary among youth within a society, and which certainly do vary in configuration across societies.1
Age: Social, Subjective, Normative
Age has a factual side-duration since birth-but factual or chronological age can be redefined and re-scaled in physiological, social or other terms by the analyst-observer. There is also a subjective side to age. Subjective age identification is concerned with how young or old a person feels, with what age category heor she identifies, or how young or old he or she would wish to be. The Barak and Stern (1986) "Ages of Me" instrument, for example, assesses Feel Age, Look Age and Do Age and Interest Age (all defined as one might surmise from the labels), which are all considered to be dimensions of Cognitive Age.
Many authors (e.g.: Settersten and Mayer 1997) have criticized the use of simple chronological age, noting that this is often a poor proxy for biological age, maturation, psychological development and social membership. Birren and Cunningham, for example, try to operationalize several kinds of age including biological age, social age and psychological age. It is generally felt, at least for the West, that age is less and less useful as a marker of personal progress through life - that there is more and more variability among similarly-aged individuals (Dannefer 1987).
A useful exercise in a comparative analysis context would be to contrast youth within countries and perhaps population sub-groups within countries on types of age scale. The obvious anchors for such comparisons are birth and potential life span, the latter taken to be a fixed quantity across all countries for males, another for females. Another is life expectancy from birth for the particular age group, again different for males and females. Life expectancy at birth varies across our AYARR countries, though the life expectancies of youth (say, persons who have reached exact age 15) will vary by much less. Next we can look at physiological age, defined essentially in terms of sexual and reproductive potential. This brings in the age at menarche or puberty which is available in several of the surveys. Then there is psychological age as defined by Birren in terms of "use of adaptive capacities of memory, learning, intelligence, skills, feelings, motivations, and emotions for exercising behavioral control or self-regulation." The AYARR surveys provide no measures of this kind.
Age structuring is one of the most pervasive and revealing aspects of social organization. All societies, in varying ways and to different degrees, associate social roles, statuses and expectations to chronological age or some variation of chronological age. These social definitions are expressed as age norms or sometimes even age-specific legal enactments. Societies sometimes enact age-specific law-for example, relating to age at marriage-in order to influence age norms and, thus, age-graded behavior.
We can distinguish several kinds of age norm. Statistical age norms form around the empirical distributions of ages at which particular events occur to individuals in a society. One may be judged early or late (early to begin dating, late to marry) relative to the prevailing statistical distributions. There are also optimal age norms which are group ideas about the ideal or preferred ages for particular events to occur. That these optimal age norms typically vary by sex and other population sub-groups is only testimony that they reflect something genuine about a society. Finally there are proscriptive/prescriptive age norms which have to do with group notions of ages at which particular events should or should not occur. Whether an individual will be influenced by these norms depends, at the outset, on whether he or she is aware of them. Statistical age norms, in particular, may sometimes be known to the researcher-observer but to no one else.
The study of optimal and proscriptive/prescriptive norms has been approached in a variety of ways. For reviews see Settersten and Mayer (1997) recently and an earlier review by Hagestad (1982). Of the many approaches summarized there we can highlight Neugarten's Kansas City Study of Adult Life, with it's Timetables for Men and Women and it's Age Norm Checklist, and also Wood's Age-Appropriate Behavior instrument. We pass over a set of additional approaches focused on images of the incumbents of various ages, such as Neugarten and Peterson's Age Association Items. Some proscriptive/prescriptive age information is available from the AYARR surveys.
Collecting Data on Ages and the Transition to Adulthood
Appendix Table A gives a detailed presentation of the various kinds of event information available from the various AYARR surveys.2 Table 1 is extracted from Appendix Table A and indicates only which "ageat ...(event)" variables are available for each of the surveys. The appendix contains much detail but for proper use requires a careful reading of the General Notes and Specific Notes provided at the bottom of the table.
It is clear that analysis of these data might be at the same time frustrating and fruitful. For every kind of analysis some countries will be absent for lack of appropriate data, and no country provides complete information. Still, it is possible to produce some worthwhile results. The analysis in this paper is limited in scope, and in particular utilizes only the age at event information.
This paper reports on an initial stage of analysis of these data and will focus on the pattern of certain core events essential to or related to the transition to adulthood. Among these are socially-recognized markers of adulthood which are virtually necessary for a person to claim adult status. Puberty (menarche or it's male equivalent) is certainly one. Certain other events may be recognized as important in the transition to adulthood although not essential, and in fact not actually experienced by some proportion of each cohort. Thus, nearly all enter marriage in Asian societies, but a small proportion do not. We will also examine a further set of events that may not be markers of adulthood but which nevertheless tend to cluster during the life-stage of adolescence or young adulthood (military experience, for example). Finally, we will consider some events that are not at all common during this stage of life but which may occur to some persons with important implications when they do. All these events will be identified and variations in their occurrence described both within and between countries and for each of the sexes. In so doing, it will be necessary to deal with the fact that our youth age range provides truncated event-distributions--an incomplete picture of lifetime experience-for certain events. We deal with this problem by limiting our attention to those in the 20-24 age range, and giving our attention primarily to medians and related non-parametric measures which are influenced only marginally by the fact that some events may have yet to occur.
The next stage of analysis (incomplete at this time) will consider sub-sets of events taken in combination. Here the goal is to identify the most common and the most important events-combinations within our samples, again looking for variations both within and between countries and by sex. We will seek to identify types of life course if that proves possible, and also to characterize the life-segments or durations between events. An important methodological problem here is the fact that not all youth will experience all events, making comparison of mean ages and durations and the recognition of patterns more difficult.
After these preliminary analyses we will turn in subsequent work to a more detailed and analytic exploration of selected combinations of events and their causal connections. This will be guided as much data limitations as by policy priorities or theory. Thus far we have looked at the connections between age at school-leaving and age at first sex (c.f. Research Brief No. 9). Other reports involve other events. See, for example Choe et. al. (2001) for analysis of determinants of age at marriage.
The Shape and Tempo of Transition
Some Basic Data
Much of the discussion in this section draws upon a statistical summary of life-events information that has been prepared for each of the six countries and for males and females separately in each. The basic information is provided in Table 2 and depicted in boxplots in Figures 1 and 2. Figure 1 has 12 parts for males and females in each of the six countries (Figure 1.Indonesia.Male, etc.). Figure 2 presents the same information, this time organized by event. Each part is for one of the 12 events and combines information for variables reflecting opinions, ideals, preferences or understandings, rather than actual events. These are included when there seems to be a direct relevance to life course issues. Finally, current status or lifetime status ("ever experienced") information is indicated when this seems relevant. Most information is available for the entire youth cohort, but when a variable is available only for specific demographic groups this is indicated (by M: male; F: female, etc.). A more detailed tabulation, available on request, indicates where in the data files (under what field names) each piece of information can be found, or what transformations are required to generate each of the variables. for each country and sex. For example Figure 2.Puberty presents boxplots for that event, first for males by country, and then for females. Next to the name of each event is the percentage of the population group (country and sex) reportedly experiencing the event.
Comparisons are assisted by the fact that all parts of each figure have the same age scale and the same arrangement of events. The age range is from birth through age 24, and the events are arranged top to bottom in the most commonly observed chronological sequence of their medians. Nevertheless, it must be recognized that comparisons among these countries and between males and females are constrained because a somewhat different subset of events is available for each country and sometimes for each of the sexes within a country. The presentation is designed to make clear which events are or are not available for each country or sex.
Box-and-whisker plots (or boxplots) provide a non-parametric summary of the distributions of events across ages. Following common practice, the shaded box for each event indicates the median age of occurrence (the median value is indicated) and is bounded by the first and third quartiles of the distribution.
That is, each box indicates the range of ages encompassing the central half of the distribution. The line extending in either direction from the shaded box, the whiskers, provide an image indicating the full range of reported ages, from the youngest to the oldest (Tukey 1977: chapter 2).
A few observations may help to clarify how information is conveyed by these Figures. In Figure 1.Nepal.Female, 13 of the 14 events are shown (not collected was age at first drug use). Across all the data, some events (for example, puberty-- menarche for girls, first wet dream for boys-- are concentrated in a relatively narrow age range, while other events, for example home-leaving, occur across a broad range of ages. This is best seen by noting the width of the shaded box. The whiskers often extend across a very wide range of ages and often depict extreme patterns or occasionally even unreasonable or implausible responses, such as puberty at age 23. But recall that these are rare, the most extreme responses which may well be accurate for the individuals providing them. Virtually none of these extreme values can be rejected out of hand, though we presume that some of them do represent measurement error.
How Long is Adolescence?
For the purpose of an initial foray into the data, we define adolescence as the interval of life between the puberty and marriage events.3 The initial question therefore becomes, what is the mean or median or typical duration from puberty to marriage? This is a more difficult question than it may seem, because many youth, even those already at ages 20-24, are still in their adolescence as we are defining it here-they have not married. We have the complete adolescence experience, by our definition, for the proportion of youth ages 20-24 who are married (the heavily shaded areas in Figure 3). For the unmarried, we must look at the duration from puberty to the survey interview (the lightly shaded areas in Figure 3) to see the adolescence experience of these young people, thus far. Their adolescence experience is incomplete.
These durations are in historical terms rather remarkable in length. Among those who are already married the duration from puberty to marriage averages as long as 6.6-7.4 years in Hong Kong and Taiwan, and averages 4.6-5.6 years in the Philippines and Thailand. And among those who are were yet married at the time of interview, young women had spent 7.4-8.8 years post-puberty and unmarried, while young men had spent 5.9-6.7 years post-puberty and unmarried.
The data for three of the countries (Hong Kong, Philippines, Taiwan) permit comparison of male and female adolescences, and reveal a fact that is not often noted - that male adolescence is markedly shorter than female adolescence. We saw earlier in Figures 1 and 2, portraying the timing of the two defining events, that male puberty occurs later than female. Male entrance to marriage has also bin later than female, but that gap has been steadily closing. These survey data tell us that in modern times male adolescence isshorter than female because young men and women in these three countries are now marrying at similar ages, and continue to have their puberty events substantially later than for females.
Events Densities in the Teen Years
It is a basic tenant about the adulthood transition that this is a period of intense change, marked by the occurrence of many important events, with a frequency or density that is not seen at any other period of life (Rindfuss 1991). That event-densities reach very high levels during the ages of youth is shown emphatically with the AYARR event information in Figure 4 (in parts A-F for each of the AYARR countries) and Figure 5 (in parts A-D for each type in a classification of event-types plus all events combined). We have distinguished three categories of events: five Reproductive Events (onset of puberty, first union, first sex, first pregnancy, first birth), three Substance Risk Events (first smoking, first drinking, first drug use), and four Institutional Shift Events (school-leaving, home-leaving, onset of work outside the home, first boyfriend/girlfriend). The latter set of events all involve a significant change in social environment-in the way that time is spent, or whom one lives with. An implicit underlying dimension here is the degree of connection to family versus to the rest of the social environment. All these events tend to move a young person into a non-familial institutional context. It will be noted that this classification of events is not entirely rigorous in theoretical terms, and especially, that the marriage event might just as well have been classified as an institutional shift.
These measures are termed "life event indexes" to highlight two aspects of their calculation. First, each is the unweighted total of its constituent events. Second, Some rescaling has sometimes been necessary to adjust for the fact that not all events are available empirically for each population group. Thus, only four of the five Reproductive Events are available for Indonesia females (age at menarche was not included in the data collection), and the data for Indonesia females have been rescaled upward accordingly.4 The data for Indonesian males illustrate another issue that we have had to resolve in order to proceed. Many of the events we are examining have the potential for occurring to both males and females, and in all of the countries. But among the Reproductive Events are two-first pregnancy and first birth-which by conventional treatment do not occur to males. We have scaled our indexes on the assumption that there are only three Reproductive Events for males. It must be noted, though, that while pregnancy and birth are unambiguously female events, these events could be-though they typically are not-- attributed to male partners and fathers. It is a matter of choice and convention that data are not collected and organized in that way. The implication here is that male rates of experience of Reproductive Events are lower than rates for females, as a matter of assumption rather than measurement.
At first glance, these charts certainly do support the notion that the teen and young adult years are a time of concentrated change and transformation. For every type of event, and every country and sex, the index rises steadily to peak somewhere around age 19 (later for Taiwan). For the Reproductive Events for females there is a secondary peak around age 12-13, because menarche is so temporally concentrated in those ages.
The female indexes are generally higher and peak at a higher level than the male indexes, primarily because of the Reproductive Events. Males report a greater contribution from events reflecting the onset of substance use of various kinds.
Remaining at a descriptive level we have outlined some important general facts including male-female and national differences in the pattern of life events during adolescence, defined for the present purpose as the period of life from puberty to marital union. We have seen some clear documentation of the notion that event densities rise to a crescendo toward the end of the teen years. It is interesting that there is such similarity among the countries in this regard. Both the commonalities and the sex and country differences are important background to many of the AYARR research reports, many of which focus on one or another aspect of risk-taking in relation to other events during the transition to adulthood. Laguna (2001), forexample, looks at home-leaving and work in relation to pre-marital sex, Choe et. al. (2001) look at the initiation of smoking, and Podhisita et. al. (2001) look at family factors including home-leaving in relation to pre-marital sex.
An important agenda of research into specific transition events remains to be completed. In particular, we have not carried out a detailed analysis of event sequences and types of transition defined on that basis, though a dissertation in progress by Jittinee Juntarodjana is doing this for Hong Kong and Thailand. There are a number of specific causal connections that remain to be explored. One, for example, will relate puberty and school leaving, to indicate the breakdown of youth who are pre and post-puberty in and out of school. Doing this by enrollment (grade) level will provide guidance for the preparation of classroom materials specific to each grade and sex. There are AYARR reports looking into another important issue, the factors underlying early versus late marriage (Choe et. al. 2001; Xenos et. al. 2001), but more can be done to clarify interactions between sexual and marital onset. One of the most immediate policy-relevant investigations not yet completed is to relate rates of risk-taking to various stages of the transition to adulthood and to various transition sequences.
Barak, B. and Stern, B. Subjective age correlates: a research note. Gerontologist. 1986; 26(5):571-78.
Buchmann, Marlis. The Script of Life in Modern Society: Entry into Adulthood in a Changing World. Chicago: University of Chicago Press; 1989.
Choe, Minja et. al. 2001. The Youth Tobacco Epidemic in Asia. East-West Center Working Papers No. 108-17.
Choe, Minja et. al. 2001. Sex and Marriage: How Closely are they Related in the Philippines, Taiwan and Thailand. East-West Center Working Papers No. 108-14.
Dannefer, D. Aging as intracohort deferentiation: accentuation, the Mathew effect, and the life course. Sociological Forces. 1987; 2:211-36. University Press; 1985.
Elder, Glen H. Jr. Adolescence in Historical Perspective. in: J. Adelson, ed. Handbook of Adolescent Psychology. New York: Wiley; 1980; pp. 3-46.
George, Linda K. Sociological Perspectives on Life Transitions. Annual Review of Sociology. 1993; 19:353-73.
Greene, A. L. Great expectations: constructions of the life course during adosescence. Journal of Youth and Adolescence. 1990; 19:289-306.
Laguna, Elma P. 2001. On Their Own: Effect of Home- and School-leaving on Filipino Adolescents' Sexual Initiation. East-West Center Working Papers No. 108-18
Paige, Karen Ericksen and Paige, Jeffery M. The Politics of Reproductive Ritual. Berkeley: University of California Press; 1981.
Podhisita, Chai et. al. 2001. The Risk of Pre-marital Sex Among Thai Youth: Individual and Family Influences. East-West Center Working Papers No. 108-5
Rindfuss, Ronald R. The Young Adult Years: Diversity, Structural Change, and Fertility. Demography. 1991; 28(4):493-512.
Settersten, Richard A. Jr. and Karl Ulrich Mayer. The Measurement of Age, Age Structuring, and the Life Course. Annual Review of Sociology. 1997; 23:233-61.
Tukey, John W. Exploratory Data Analysis. Reading Mass.: Addison-Wesley Publishing Co. 1977