Life Course Analysis

views updated


The life course approach is an interdisciplinary program of study, under development since the mid-1970s, which has been increasingly influential in demographic research. It is concerned with explaining how and when events such as leaving the parental home, starting or dissolving a union, having a child, migration, job entry and exit, and retirement are experienced. Life course analysis entails the collection of life course data together with the (statistical) analysis of the timing of events (when do they happen?),their sequencing (in which order do they happen?),and their quantum (how many events happen?). The focus of this article is on quantitative methods, although qualitative life course analysis has also been influential and is sometimes integrated with quantitative research.

In their 1998 review of methods of life course research, Janet Giele and Glen Elder identify the chief elements that shape individual lives and that are crucial for the analysis of life courses. These are: individual development; history and culture (location in time and place); and social relations (linked lives). Parallel and potentially interdependent trajectories of individual lives are the main units of analysis, with the trajectories marked by events. These elements have natural counterparts at the macro level: individual development lies behind the use of age as the primary time axis; location in time and the idea of linked lives suggest using a cohort approach to the study of social change; history and culture emphasize the importance of period and location.

Collection of Life Course Data

Quantitative life course data may be collected in surveys, using either question lists or so-called event history calendars. Retrospective collection of the timing of events has become a standard feature of most demographic surveys. Life course data can also be obtained from panel surveys or other follow-up surveys and from civil registration data.

In surveys, the timing of events is usually asked within a roster of questions for each trajectory separately and following a particular order. For instance, the timing of events concerned with the relationships within the family of origin would be asked before the timing of events on union formation and dissolution. The quantum of events is collected in the same context, while the sequencing of events is derived indirectly from the information on timing. This way of collecting life course data has been widely advocated and used in life course research. In the 1990s it was implemented in surveys such as the Demographic and Health Surveys.

Studies employing event history calendars start by collecting data on the timing of so-called landmark events through a question list. Information on the timing of other events, and on the state the individual occupied in each time unit over the reference period, is then collected by use of a graphical display of the trajectory of primary interest (either on paper or on a computer screen). The complexity of the calendar depends on the length of the reference period and on the time unit used–typically one month. Event history calendars are extensively applied in panel surveys such as the U.S. Panel Study of Income Dynamics.

Robert F. Belli, William L. Shay, and Frank P. Stafford compared the two types of data collection in 2001; their results indicate that event history calendars generally yield more accurate reports, although sometimes with increased overreporting of events.

Statistical Analysis of Life Courses

Two approaches have been followed in the statistical analysis of life course data. Event history analysis focuses on time-to-event as the key dependent variable. Sequence analysis focuses on life courses or segments of the life course as a conceptual unit.

Event history analysis. This statistical method has found major application in demography since the 1980s. Applied to individual life course data, the approach uses life table nonparametric techniques to compare the timing of life course events across space (including international comparisons) and time (mostly using the cohort as the preferred time dimension). This is the individual-level equivalent of period and cohort analysis in traditional demography.

The regression models of event history analysis contribute to the explanation of life course dynamics by linking the time-to-event variable with explanatory variables (covariates). Covariates can be external to a trajectory (as is the case for macro-level variables, comprising period effects), or internal to a trajectory (as is the case for other trajectories of the life course that are potentially influencing the time-to-event). External covariates can be of three kinds: those that are fixed during a life course or from a particular point of time (birth cohort, age-at-marriage when studying time-to-divorce); covariates whose temporal dynamics cannot be influenced by the events in the trajectory of interest (age of an individual in the case of time-to-divorce); and those located at an aggregate level of social dynamics (time period, regional economic indicators, policy indicators). Multilevel event history models, developed in the 1990s, allow accurate treatment of the case of individuals aggregated into household or regional clusters, therefore taking into account the "time and place" dimensions of life courses.

Internal covariates usually refer to other trajectories of the life course of the same individual or of linked individuals, and their use allows researchers to study complex interdependencies between trajectories. Event history analysis may take into account unobserved factors underlying these interdependencies, such as value orientations or attitudes. The relevance of time-constant unobserved factors for the analysis of parallel and potentially interdependent trajectories has been debated in the literature. The so-called causal approach of Hans-Peter Blossfeld and Götz Rohwer assumes that all factors that are relevant to the simultaneous analysis of several trajectories are observed and included in the past history of the trajectories. Other modeling approaches allow for the effects of time-constant unobserved factors. The most general applied approach is the multilevel and multiprocess modeling of life courses developed by Lee A. Lillard and Constantijn W.A. Panis.

Sequence analysis. Older life course concepts like that of the family life cycle were holistic and made more or less explicit reference to biological structures. The recent life course literature has been more analytic. Nevertheless, by focusing on specific events researchers may achieve a unitary, holistic perspective on trajectories of the life course. The most-used analytical technique for the holistic analysis of life courses is known as sequence analysis, introduced into the social sciences during the 1990s by Andrew Abbott.

In sequence analysis, each life course or trajectory in the life course is represented as a string of characters (states). This representation is analogous to the one used to code DNA molecules in the biological sciences. Indeed, one method used to analyze sequence-type data, optimal matching analysis (OMA), was originally created for the alignment of DNA sequences. The goal of OMA is to compute a matrix of dissimilarities between pairs of sequences, starting from a definition of distance between states, and of "costs" of inserting and deleting states in a sequence. The dissimilarity matrix can be used as the input for statistical techniques based on dissimilarity, such as cluster analysis or multidimensional scaling. The method has been applied to the sociology of occupations. A demographic application–to the analysis of the transition to adulthood–is discussed in Francesco C. Billari's 2001 article.

Alternative approaches to the analysis of life courses as a conceptual unit have also been explored, although they have not surfaced extensively in the literature as of the early twenty-first century. These include the use of correspondence analysis and data-mining techniques.

See also: Cohort Analysis; Event History Analysis; Family Demography; Family Life Cycle.


Abbott, Andrew, and Angela Tsay. 2000. "Sequence Analysis and Optimal Matching Methods in Sociology." Sociological Methods and Research 29:3–33.

Belli, Robert F., William L. Shay, and Frank P. Stafford. 2001. "Event History Calendars and Question List Surveys: A Direct Comparison of Inter-viewing Methods." Public Opinion Quarterly 65:45–74.

Billari, Francesco C. 2001. "The Analysis of Early Life Courses: Complex Descriptions of the Transition to Adulthood." Journal of Population Research 18: 119–142.

Blossfeld, Hans-Peter, and Götz Rohwer. 2002. Techniques of Event History Modeling. Mahwah, NJ: Erlbaum.

Dykstra, Pearl A., and Leo J. G. van Wissen. 1999. "Introduction: The Life Course Approach as an Interdisciplinary Framework for Population Studies." In Population Issues: An Interdisciplinary Focus, ed. Leo J. G. van Wissen and Pearl A. Dykstra. New York: Kluwer Academic/Plenum Publishers.

Freedman, Deborah, Arland Thornton, Donald Camburn, Duane Alwin, and Linda Young-DeMarco. 1988. "The Life History Calendar: A Technique for Collecting Retrospective Data." Sociological Methodology 18: 37–68.

Giele, Janet Z., and Glen H. Elder Jr., eds. 1998. Methods of Life Course Research: Qualitative and Quantitative Approaches. Thousand Oaks, CA:Sage.

Lillard, Lee A., and Constantijn W. A. Panis. 2000. AML Multilevel Multiprocess Statistical Software, Release 1.0. Los Angeles: Econ Ware.

Mayer, Karl Ulrich, and Nancy B. Tuma, eds. 1990. Event History Analysis in Life Course Research. Madison: University of Wisconsin Press.

Wu, Lawrence L. 2000. "Some Comments on Sequence Analysis and Optimal Matching Methods in Sociology: Review and Prospect." Sociological Methods and Research 29: 41–64.

Francesco C. Billari

About this article

Life Course Analysis

Updated About content Print Article