Mark Tranmer

The Application of Zone-Design Methodology in the 2001 UK Census:

This paper reviews the automated zone-design procedures adopted for the creation of 2001 Census output geography in the United Kingdom. A microsimulation approach is used for the creation of household records to populate actual postcode and enumeration district boundaries, and a series of output area design scenarios are applied to these data, allowing the effects of the new design constraints to be evaluated. The authors identify the advantages of using an intra-area correlation measure for the maximization of social homogeneity within output areas, and explore the differences between the 1991 and 2001 approaches to output geography.

Personal support networks of immigrants to Spain: A multilevel analysis

Abstract Immigrant flows to Spain have increased greatly in the last decade, but little is known about the composition and role of their personal support networks. Our research questions are: (1) Which factors are associated with ties between immigrants and ‘Spaniards’ (the more settled resident Spanish population), compared with immigrants and non-Spaniards (other immigrants)? (2) Do the support roles of Spaniards and non-Spaniards differ? We analyse personal network (ego-net) survey data. Multilevel logistic regression models are applied, in which the unit of analysis is the undirected tie between an immigrant (ego) and an alter and the dependent variable is whether this tie is to a Spaniard alter, as opposed to a non-Spaniard. We determine the characteristics that are most strongly associated with the probability of a tie between an immigrant and a Spaniard, compared with a non-Spaniard, and consider characteristics of the immigrants (ego), the alters, the relative characteristics of ego-alter, support roles, and local geographical factors. We find a tie to a Spaniard alter is more likely if the immigrant’s country of birth is Portugal or Eastern Europe; if the alter is a work colleague or neighbour; if alter is older than ego. There is geographical variation in the probability of ties to Spaniards, partly explained by the local area presence of co-nationals from the same country of origin as the immigrant. A tie to a Spaniard alter is less likely for immigrants from North Africa (Maghreb); those with no previous contact with Spain; those who are not the first of their peer group/family to immigrate; if ego and alter both work in agriculture. Material help is more likely to be exchanged with a Spaniard alter. Non-Spaniard alters are more likely to exchange help with accommodation and information. ‘Finding a job’ is equally associated with Spaniard and non-Spaniard alters. A tentative conclusion is that some combinations of these characteristics, where a tie to a Spaniard is less likely, may be associated with higher levels of prejudice. Conversely, those characteristics that are positively associated with a tie to a Spaniard may indicate situations where integration of the immigrant population with Spaniards is successfully taking place, and where prejudices are lower, or non-existent. These findings may therefore be helpful for targeting resources to reduce such prejudices. The different types of support exchanged between immigrants and Spaniards and immigrants and non-Spaniards, may indicate current shortfalls in this process, as well as indicating where this support is successfully being exchanged.

Combining Sample and Census Data in Small Area Estimates: Iterative Proportional Fitting with Standard Software*

Abstract The combination of detailed sample data with less detailed but fully enumerated marginal subtotals is the focus of a wide range of research. In this article we advocate careful modeling of sample data, followed by Iterative Proportional Fitting (IPF). The modeling aims to estimate accurately the interaction or odds ratios of complex tables, which is information not contained in the marginal subtotals. IPF ensures consistency with the subtotals. We advance this work in three practical ways. First, we show that detailed small-area estimates of both counts and proportional distributions usually gain accuracy by combining data for larger areas containing the small areas, and we illustrate the multilevel framework to achieve these estimates. Second, we find that a general classification or socioeconomic typology of the small areas is even more associated with the within-area interactions than is membership of the larger area. Third, we show how the Statistical Package for the Social Sciences (SPSS) can be used for IPF in any number of dimensions and with any structure of constraining marginal subtotals. Throughout, we use an example taken from the 1991 U.K. Census. These data allow us to evaluate various methods combining 100 percent tabulations and the Samples of Anonymised Records. *Census data are Crown Copyright; the 1991 Sample of Anonymised Records (SAR) s were purchased by the U.K. Economic and Social Research Council and Joint Information Systems Committee (JISC) for academic and other research purposes. The work for this paper and its writing were supported respectively by the U.K. Economic and Social Research Council awards R000223703, “Combining aggregate and micro-data to extend census tables for local areas” and RR000271214, “Local demography and race.” Paul Norman provided helpful comments on a draft.

Ignoring a level in a multilevel model: Evidence from UK census data

Because of the inherent multilevel nature of census data, it is often appropriate to use multilevel models to investigate relationships between census variables. For a local population, the data available from the census allow a three-level nested model to be assumed, with an individual level (level 1), an enumeration district (ED) level (level 2), and a ward level (level 3). The consequences of ignoring one of the three levels in this model are assessed here theoretically. Empirical results, based on 1991 UK Census data, are also provided, comparing the variance components estimated from the three-level model with analyses based on models where the ED or ward level are ignored. The results show how the variation that occurs at the level not included in the models is redistributed to the other levels that the models do include.

Leaving home in Spain: When, where and why?

Holdsworth C., Voas D. and Tranmer M. (2002) Leaving home in Spain: when, where and why?, Reg. Studies 36, 989–1004. Spanish transitions out of the parental home are characterized by older ages of leaving and a close association between leaving home and partnership formation, compared to countries in northern Europe. Yet throughout Spain we find important regional variation in the intensity and timing of leaving home. This paper provides new insights into the causes of this regional diversity, focusing on both economic and cultural dimensions. The analyses are based on multilevel models that control for individual and provincial characteristics. Economic factors play a vital role at the individual level, yet are less important at the provincial level. Regional variation is more closely associated with the impact of both the timing of partnership formation and the relationship between leaving home and getting married. These distinctive regional patterns of leaving home and partnership formation are discussed with reference to historical traditions of family and household formation.

Multiple-membership multiple-classification models for social network and group dependences

The social network literature on network dependences has largely ignored other sources of dependence, such as the school that a student attends, or the area in which an individual lives. The multilevel modelling literature on school and area dependences has, in turn, largely ignored social networks. To bridge this divide, a multiple-membership multiple-classification modelling approach for jointly investigating social network and group dependences is presented. This allows social network and group dependences on individual responses to be investigated and compared. The approach is used to analyse a subsample of the Adolescent Health Study data set from the USA, where the response variable of interest is individual level educational attainment, and the three individual level covariates are sex, ethnic group and age. Individual, network, school and area dependences are accounted for in the analysis. The network dependences can be accounted for by including the network as a classification in the model, using various network configurations, such as ego-nets and cliques. The results suggest that ignoring the network affects the estimates of variation for the classifications that are included in the random part of the model (school, area and individual), as well as having some influence on the point estimates and standard errors of the estimates of regression coefficients for covariates in the fixed part of the model. From a substantive perspective, this approach provides a flexible and practical way of investigating variation in an individual level response due to social network dependences, and estimating the share of variation of an individual response for network, school and area classifications.

Using the relational event model (REM) to investigate the temporal dynamics of animal social networks

Social dynamics are of fundamental importance in animal societies. Studies on nonhuman animal social systems often aggregate social interaction event data into a single network within a particular time frame. Analysis of the resulting network can provide a useful insight into the overall extent of interaction. However, through aggregation, information is lost about the order in which interactions occurred, and hence the sequences of actions over time. Many research hypotheses relate directly to the sequence of actions, such as the recency or rate of action, rather than to their overall volume or presence. Here, we demonstrate how the temporal structure of social interaction sequences can be quantified from disaggregated event data using the relational event model (REM). We first outline the REM, explaining why it is different from other models for longitudinal data, and how it can be used to model sequences of events unfolding in a network. We then discuss a case study on the European jackdaw, Corvus monedula, in which temporal patterns of persistence and reciprocity of action are of interest, and present and discuss the results of a REM analysis of these data. One of the strengths of a REM analysis is its ability to take into account different ways in which data are collected. Having explained how to take into account the way in which the data were collected for the jackdaw study, we briefly discuss the application of the model to other studies. We provide details of how the models may be fitted in the R statistical software environment and outline some recent extensions to the REM framework.

The embeddedness of organizational performance: Multiple membership multiple classification models for the analysis of multilevel networks

We present a Multiple Membership Multiple Classification (MMMC) model for analysing variation in the performance of organizational sub-units embedded in a multilevel network. The model postulates that the performance of organizational sub-units varies across network levels defined in terms of: (i) direct relations between organizational sub-units; (ii) relations between organizations containing the sub-units, and (iii) cross-level relations between sub-units and organizations. We demonstrate the empirical merits of the model in an analysis of inter-hospital patient mobility within a regional community of health care organizations. In the empirical case study we develop, organizational sub-units are departments of emergency medicine (EDs) located within hospitals (organizations). Networks within and across levels are delineated in terms of patient transfer relations between EDs (lower-level, emergency transfers), hospitals (higher-level, elective transfers), and between EDs and hospitals (cross-level, non-emergency transfers). Our main analytical objective is to examine the association of these interdependent and partially nested levels of action with variation in waiting time among EDs – one of the most commonly adopted and accepted measures of ED performance. We find evidence that variation in ED waiting time is associated with various components of the multilevel network in which the EDs are embedded. Before allowing for various characteristics of EDs and the hospitals in which they are located, we find, for the null models, that most of the network variation is at the hospital level. After adding these characteristics to the model, we find that hospital capacity and ED uncertainty are significantly associated with ED waiting time. We also find that the overall variation in ED waiting time is reduced to less than a half of its estimated value from the null models, and that a greater share of the residual network variation for these models is at the ED level and cross level, rather than the hospital level. This suggests that the covariates explain some of the network variation, and shift the relative share of residual variation away from hospital networks. We discuss further extensions to the model for more general analyses of multilevel network dependencies in variables of interest for the lower level nodes of these social structures.

Introduction to multilevel social networks

Social network research focuses on the study of social systems y conceptualizing their internal structure in terms of sets of comlex dependencies among social agents in the form of dyadic social ies. Typically, models for social networks incorporate additional eatures such as actor attributes. Models for social networks may lso be extended in various ways by considering, for example, muliplex or bipartite representations. However, incorrect inferences can be drawn from social netork analysis if the system is conceptualized in an overly simplistic ay. This can happen if crucial elements of social structure are gnored when the data are collected, or are mis-specified in he model used for the analysis. As social network researchers, e know this well, because we avoid individualistic analysis of ttributes when social structure is relevant. In one of the foundaional articles of contemporary social network analysis, Harrison hite and co-authors warned against relying on social classificaion as the sole basis for understanding social structure: “. . .largely ategorical descriptions of social structure have no solid theoretial grounding; . . . network concepts may provide the only way to onstruct a theory of social structure” (White et al., 1976, p.732). etwork researchers (but not all social scientists) have learnt that esson well. Yet, having imbibed White’s warning, we all-too-often content urselves with a relatively simple network representation, without hinking through whether our conceptualization is sufficient for he purpose of our research. We face important theoretical choices ere, and of course it is not feasible to control for every possible actor. Social systems may have important dynamic, temporal and eospatial elements, and if we regard these as central to the proesses we are studying, then they need to be incorporated into ur conceptualization of the social system and hence into the type f data we collect. Recent special issues of Social Networks have oncentrated on network research involving dynamic and spatial actors. Social systems can also contain hierarchy. In his classic paper n the “architecture of complexity,” Herbert Simon (1962) offered compelling analysis of the evolutionary mechanisms responsile of producing the hierarchical structure so frequently observed

Unravelling ecological analysis

Ecological analysis involves analysing aggregate data for groups of individuals to make inferences about relationships at the individual level. Often the results of such analyses give badly biased estimates. This paper will consider the sources of bias in linear regression analysis using aggregate data. The role of variation of the individual level relationships between groups and the consequent within-group correlations and how these are related to auxiliary variables that characterise the differences between groups is considered. A method of adjusting ecological regression for the effects of auxiliary variables is described and evaluated using data from the 1991 Australian Census.