Fitting Constrained Poisson Regression Models to Interurban Migration Flows

This paper demonstrates the effects of fitting singly and doubly constrained spatial interaction models using the Poisson regression approach. A large data set containing migration flows between labor market areas in Great Britain in 1970–71 is used. The results of fitting unconstrained, singly constrained, and doubly constrained models are compared with respect to goodness of fit and the interpretability of parameter estimates. The addition of other explanatory variables to the model is also explored.     

Modeling Migration Careers,Using Data from a British Survey

Development of more sophisticated techniques for modeling longitudinal data has implications for improving our understanding of migration. This paper uses longitudinal data from the British Social Change and Economic Life initiative to disentangle the effects of population heterogeneity, progress through the life cycle and secular change on observed migration differentials. The data consist of retrospective life histories from people sampled in several contrasting localities in Great Britain, in which residential moves can be linked to changes in occupation and household structure. We present a framework for analysis of data of this type using a generalized linear modeling approach, together with results concerning variations in the probability of migration with age, gender, and changes in household and occupational circumstances. Of particular note is the evidence of substantial duration-of-residence effects and an unexpected later-career increase in migration propensity.     

After Augustine,after Markus: the problem of the secular at the end of antiquity

This article revisits Robert Markus's account of the de-secularization of the Latin West between Augustine and Gregory the Great. It uses letters of advice for rulers written by early sixth-century clerics to contest his narrative of a ‘grand simplification’ of Christian thought. Multiple overlapping conceptions of the secular were still in play after the fall of Rome, articulated, not in the absence of, but in dialogue with robust political institutions. By uncoupling Christian ideas of secularity from the actual degree of religious pluralism or tolerance in a given society, historians can better capture the continued complexity of early medieval secularities.     

A Framework for Inserting Visually Supported Inferences into Geographical Analysis Workflow: Application to Road Safety Research

Road safety research is a data-rich field with large social impacts. Like in medical research, the ambition is to build knowledge around risk factors that can save lives. Unlike medical research, road safety research generates empirical findings from messy observational datasets. Records of road crashes contain numerous intersecting categorical variables, dominating patterns that are complicated by confounding and, when conditioning on data to make inferences net of this, observed effects that are subject to uncertainty due to diminishing sample sizes. We demonstrate how visual data analysis approaches can inject rigor into exploratory analysis of such datasets. A framework is presented whereby graphics are used to expose, model and evaluate spatial patterns in observational data, as well as protect against false discovery. Evidence for the framework is presented through an applied data analysis of national crash patterns recorded in STATS19, the main source of road crash information in Great Britain. Our framework moves beyond typical depictions of exploratory data analysis and transfers to complex data analysis decision spaces characteristic of modern geographical analysis.     

Representing Structural Change in U.S. Migration Patterns

This paper focuses on the identification of three sets of relationships exhibited by a time series of age-specific national migration rates and multiregional transition probability matrices: the relationship across time for the same age group, as defined by period (temporal) transformation coefficients or matrices; the relationships across age at a particular moment in time, as defined by age transformation coefficients or matrices; and the relationships across time and age, as defined by cohort transformation coefficients or matrices. Given any two of these relationships one can solve for the third. This feature of age-period-cohort relationships has a practical application in the development of improved methods for representing structural change in migration patterns over time and for indirectly estimating migration patterns using inadequate data.