SEMI-MECHANISTIC MODELLING IN NONLINEAR REGRESSION: A CASE STUDY

This paper discusses the use of highly parameterized semi‐mechanistic nonlinear models with particular reference to the PARJIB crop response model of Reid (2002) [Yield response to nutrient supply across a wide range of conditions 1. Model derivation. Field Crops Research 77, 161–171]. Compared to empirical linear approaches, such models promise improved generality of application but present considerable challenges for estimation. Some success has been achieved with a fitting approach that uses a Levenberg–Marquardt algorithm starting from initial values determined by a genetic algorithm. Attention must be paid, however, to correlations between parameter estimates and an approach is described to identify these based on large simulated datasets. This work illustrates the value for the scientist in exploring the correlation structure in mechanistic or semi‐mechanistic models. Such information might be used to reappraise the structure of the model itself, especially if the experimental evidence is not strong enough to allow estimation of a parameter free of assumptions about the values of others. Thus statistical modelling and analysis can complement mechanistic studies, making more explicit what is known and what is not known about the processes being modelled and guiding further research.     

The evolution of federal employment

The first federal employees (other than Founding Fathers) were clerks. As federal employment grew in the 1820s, its ranks came to be dominated by patronage workers. After the Civil War, bureaucrats slowly displaced patronage workers at the federal level. Now, federal employments are being privatized and the bureaucracy shrunk. I explain this evolution of federal employment with one simple model of politics in a changing environment. Politicians combine into political firms to promise benefits to subsets of voters in return for election. If elected, politicians provide the benefits as efficiently as possible. Thus, politicians choose the form and size of their political firms to maximize expected political profits. Environmental changes affect the choices. The breach of the Appalachians in the 1810s, defeat of the South in the Civil War and the simultaneous rise of big and transcontinental industry, and contemporary worldwide economic inte-gration (globalization) are three environmental changes that changed the efficient organization of federal political firms.     

INTEGRATING ECONOMIC DUALISM AND LABOR MARKET SEGMENTATION:

Although the U.S. economy of the early twenty-first century is vastly different from the U.S. economy prior to the 1970s, the nature of these economic changes and their impact on U.S. workers is unclear. This article claims that despite contemporary economic shifts, differential labor and employer power continues to segment the economy, and workers' position in the labor market continues to predict their rewards, beyond the effects of gender, race, and human capital. Drawing on segmented labor market and dual economy research, we propose a four-category model of the structural factors that influence variance in work-related rewards. We examine the distribution of jobs in each of four categories between 1974 and 2000 and observe that losses and gains across categories are unevenly distributed by race and gender. While white men have experienced the greatest declines in employment and earnings, they have maintained their absolute advantage over women and nonwhites. In multivariate analyses, we find that the structural position of employment continues to be a significant determinant of wages. Although women and racial minorities have experienced sizable increases in employment in primary labor market jobs in the core of the economy, both groups remain overrepresented in low-paying jobs. Moreover women, but not nonwhite men, consistently receive significantly fewer rewards for their labor in both low-paying and high-paying jobs. Our findings suggest that structural factors continue to influence earnings inequality, especially across race and gender lines.     

The effect of the sex composition of jobs on starting wages in an organization: Findings from the NLSY

We show that individuals in a job with a higher percentage of females earn lower starting wages with an employing organization. This holds true with controls for individuals’ human capital, job demands for skill or difficult working conditions, and detailed industry. We use a measure of sex composition that applies to detailed jobs: cells in a three-digit census occupation by three-digit census industry matrix. We use pooled panel data from the 19791987 waves of the National Longitudinal Survey of Youth. The unit of analysis is the spell-the time in which a person worked for one organization. The dependent variable is the first wage in the spell. We use models with fixed-effects to control for unmeasured, unchanging individual characteristics; we also show results from OLS and weighted models for comparison. The negative effect on wages of the percentage female in one’s job is robust across procedures for black women, white women, and white men. For black men the sign is always negative but the coefficient is often nonsignificant.     

A family of models for uniform and serial dependence in repeated measurements studies

Data arising from a randomized double-masked clinical trial for multiple sclerosis have provided particularly variable longitudinal repeated measurements responses. Specific models for such data, other than those based on the multivariate normal distribution, would be a valuable addition to the applied statistician's toolbox. A useful family of multivariate distributions can be generated by substituting the integrated intensity of one distribution into a second (outer) distribution. The parameters in the second distribution are then used to create a dependence structure among observations on a unit. These may either be a form of serial dependence for longitudinal data or of uniform dependence within clusters. These are respectively analogous to the Kalman filter of state space models and to copulas, but they have the major advantage that they do not require any explicit integration. One useful outer distribution for constructing such multivariate distributions is the Pareto distribution. Certain special models based on it have previously been used in event history analysis, but those considered here have much wider application.     

Determinants of short- and long-term mobility expectations for home owners and renters

Confusion about the role of residential satisfaction vis-à-vis structural factors in the mobility process stems from the failure to examine the determinants of mobility over varying time frames and housing tenures. Using survey data for a random sample of 580 Phoenix-area households, we test models of short-term (1 year) and long-term (5 years) mobility expectations for home owners and renters. The results show that residential satisfaction mediates the effects of structural variables on mobility expectations in the short term for home owners. In the long-term model for home owners and the short-term model for renters, the role of satisfaction as an intervening force declines in relative importance. Among renters, structural variables operate directly on long-term mobility expectations.     

Infant feeding and post-neonatal mortality in Derbyshire,England, in the early twentieth century

This paper examines influences on post-neonatal mortality in Derbyshire (England) in the early twentieth century, by applying multivariate hazard analysis to a rare individual-level data set. The data allow detailed patterns of breastfeeding and weaning to be examined. The role of feeding is given special attention as a mediator between mortality and the other environmental, social, and demographic factors considered. Twins and illegitimate children were more likely to have been hand-fed, but this could explain only a small fraction of their increased vulnerability. Artificial feeding was associated with increased risks of death from diseases was predominantly the result of the greater likelihood of congenitally weak children being hand-fed. Most of the variation in post-neonatal mortality, particularly from respiratory disease, was explained by environmental influences - population density, altitude, and the presence of mining.     

Ctrl+ALT-RIGHT: reinterpreting our knowledge of white supremacy groups through the lens of street gangs

For the last 30 years, gang scholars have explicitly disregarded white youth who are active in groups with overt or implicit ties to neo-Nazi/skinhead or white power ideologies from street gang research. This article argues that this decision to keep these two groups separate stems from a misinterpretation of the realities of these groups and this decision has had a long-standing impact on how researchers, law enforcement and policymakers understand and interact with youth active in these groups. The coining of the term ‘Alt-Right,’ and the re-emergence of white power youth in the international dialogue, underscores how the lack of systematic research severely limits our knowledge about youth involved in gangs with white power leanings. Based on these concerns, this article challenges the current understanding of both skinheads and their troublesome youth groups/gangs. By orienting the limited research on skinhead gangs within key street gang domains, this article draws attention to disconnects in the literature that have influenced how researchers approach the study of skinhead youth. This study’s conclusions support the purposeful inclusion of skinhead youth in future street gang research.     

Individual- and Contextual-Level Factors Affecting the Use of Social Support Services among Older Adults

ABSTRACT

By 2060, the number of Americans aged 65 and older is expected to more than double, while the number of Americans aged 85 and older is expected to nearly triple. As the nation's aging population grows, older adults will need to rely on social support services, such as transportation and housing services, in order to remain active and lead independent lives. In this study we use data collected from the elderly supplement of the Southeastern Pennsylvania Household Health Survey (SPHHS) (n = 3,042) to explore the relationship between the availability of elderly specific social service providers and utilization of social support services among older adults. We find that while the number of elderly specific social service providers can increase use of social support services among older adults, its impact is relatively minimal. We find that individual factors, instead, are stronger predictors of service use. This is a finding that should be particularly encouraging for elder care providers who may not have the resources needed to undertake large structural changes (like building new facilities). Still, future research should explore how the availability of a broader range of elderly specific social services (than explored in this study) impacts use.