Featured Manuscripts

Nonlinear dynamics in births to teens in Texas

Patti Hamilton, Texas Woman’s University; Bruce J. West, Army Research Office

Background

Rates of birth to teens of all ages and racial/ethnic groups have been dropping since 1990 (Ventura, Curtin & Matthews, 1999). However, the size of the problem in Texas remains significant. In 1996, in Texas there were 80,490 pregnancies and 52,273 births to girls 15-19 years old (National Campaign to Prevent Teen Pregnancy, 1999). In 1996 the U.S. rate of pregnancy for young women 15 to 19 years old was 97 per 1000 girls of that age. In that same year the rate in Texas was 113 per 1000 girls of that age. The mean age for teens giving birth was 17.62 years. Aside from the personal costs borne by individuals, there are economic costs to the state. These costs include loss of skilled and educated workers and increased public assistance for both mother and child when young mothers drop out of school. To plan strategically to alleviate these problems, the most complete and accurate information possible is necessary. The methods discussed in this paper have strong potential for obtaining this information.

An interdisciplinary team made up of nurses, physicists and other health professionals has been established within the Center for Nonlinear Science at Texas Woman’s University (http://www.twu.edu/cns/). The goals of the team are 1) to determine whether nonlinear analysis techniques reveal new patterns in daily counts of teen births; 2) to model the complex process determining teen births in Texas; 3) to identify "real world" factors in the model above; and 4) to appropriately transfer nonlinear analysis techniques from physics into sociodemographic research.

The team carried out research efforts in two phases. Phase One entailed the search for chaotic dynamics in the data. Phase Two, currently under way, is a feasibility study of two promising nonlinear analysis techniques for application to demographic data, namely 1) relative dispersion and 2) recurrence analysis. Data for the study were obtained from computerized birth certificate files of the Texas Department of Health.

Figure 1. Daily counts of births to teens from 1964 through 1998.

It is worthy of note that all the analyses presented here were carried out on daily counts of births. Rates of births per 1000 girls in the population are informative but not feasible when the unit of measurement is a daily count. The size of the population is available from census estimates at five-year intervals. Thus, to create a rate on a daily basis was deemed inappropriate. The time series appears in Figure 1.

During the study period (1964 to 1998) the year with the fewest births was 1964. The year with the greatest number of births was 1982. The fewest teen births over the study period occurred in the second quarter of each year. The third quarter of every year (July, August and September) had significantly more births to teens than any other quarter. There was one notable exception, however, in 1982. In that year only, more births to teens occurred in the first quarter than in any other quarter.