The book belongs to the series Computational Social Sciences and presents a modern methodology, techniques, and estimations of demography and social processes on the example of New Zealand longitudinal censuses data spanning 1981–2006. The monograph is organized in 11 chapters and appendix, containing multiple tables and references to the most recent sources.
Chapter 1 of “Introduction” discusses expansion of internet and power of modern computers in computational sociology via statistical modeling for demographical dynamics of birth, death, marriage, family size, longevity, immigration, etc. Simulation techniques, particularly, the agent-based modeling (ABM) can be applied for description and prediction of social changes. Chapter 2 of “Conceptual and Analytical Foundations” describes this project as performed by the model called SociaLab for studying social change and its key drivers. Social phenomena can be seen as aggregation of actions taken by individual human beings. Considering life trajectories by microdata in statistical and computational form in various counterfactual “what if” scenarios permits to predict sociological changes. Chapter 3 considers “SociaLab: A Dynamic Microsimulation Model” in ABM approach completed by the discrete-time dynamic adjustment. The key census-based social and demographic indicators are used in the life-course statistical analysis in age/period/cohort reconstruction for shaping societal change in 5-yearly census data collections. Chapter 4 of “Tracking Societal Change: Its Major Components” is focused on the “life-course” concept which connects individual biographies to different life aspects, for instance, family, work, and institutions, political and economic reforms and innovations. All these first chapters give a review of various methodologies selected from the works of other authors for implementation in the SociaLab modeling. The statistical description starts from Chapter 5 of “Data Preparation” by census microdata linked over time for constructing simulation models by a hundred thousand and more observations by demographic, household, educational, income, employment, welfare, housing tenure, immigration, and other characteristics. To fill the missing values MICE software is applied.
Chapter 6 of “Statistical Analysis” presents estimations of probability of transitions between each two censuses starting with 1981–1986 and finishing with 2001–2006, performed by regression modeling with various indicators. In the “Main Module,” an outcome, or dependent variable is presented by one of the changing over time characteristics, the predictors are all possible lagged variables preceding in causal sequence. Each outcome in its turn becomes a predictor variable in the next models. For example, living arrangements (living alone, or with children) could depend on demographic indicators (age, ethnicity, etc.) of the previous moments in time, and have impact on material (employment, income, welfare) and non-material (education level, religion) assets. All time-varying variables are analyzed as outcomes, with logistic regression for binary outcomes, multinomial model for multi-category, and multiple linear model for continuous dependent variables. SAS hpgenselect procedure is used for building generalized linear model with stepwise variable selection by AIC criterion. The results of the analysis is a series of lagged dependent variables (LDV) in the form of predictive equations describing the functioning of social mechanism depending on demographics. The models are considered separately by age groups as a proxy for life stage in the life course. In the “Population Dynamics Module,” additional models are considered to account for the effects of demographic processes, such as birth and death, immigration and emigration, with projection of these changes in population numbers into the near future.
Chapter 7 of “Simulation” deals with the problem of testing different potential policies in theoretical scenarios with help of changing data and departing from the base model for transitional probabilities. For example, a logistic regression of the partnership status (partnered or not) is used in the base time period, and then by the predicted probability a binary random variable is generated and stochastically assigned to the following models where this variable serves as a predictor. This simulation is repeated twenty times, the results are averaged, and such approach is applied to estimate impact of population dynamics. Questions of calibration, alignment, and validation are discussed, and a customized program for microsimulation called Simario is described.
Chapter 8 of “The ‘Seven Ages’: A Framework for Social and Policy Issues,” considers aggregate patterns of social and demographic changes going by several stages of an individual’s life-course. These seven “ages of man” (by W. Shakespeare) and their core features are as following: 0–4 years old, early childhood—health and thriving; 5–14 years, childhood and early adolescence—education and readiness; 15–20 years, youth and emergent adulthood—gaining and keeping employment; 21–29 years, transition to adulthood—setting into stable partnership; 30–54 years, middle adulthood—successfully raising families; 55–74 years, older life—retirement and successful ageing; 75+ years old, later life—the risks of dependency. Transitions between the stages are related to the government resources to the corresponding concerns, and society regulates them via established systems and institutions.
Chapter 9 of “Tracking Societal Change: Descriptive Results” presents a few tables of numerical output for demographic, for instance, in age groups and ethnicity comparisons. It reveals, that in three decades from 1986 to 2006 the population of New Zealand has been ageing and becoming more ethnically diverse, with more highly educated workforce, and other socio-economic differences. Chapter 10 of “‘What If?’: Counterfactual Modelling with SociaLab,” is devoted to creating numerous series of alternative histories by different assumptions, and projecting them to estimating possible future developments in various scenarios for countless aspects of the society transformations in time. Some examples are considered and presented in tables, for instance, answering to the question “what if policies governing inward migration had not been liberalized”? Or “what if the government decisions of 1999–2008 about paid parental leave, 20 hours early childhood education, a comprehensive family support, and employment incentive package had not been enacted”? These and other examples are given on counterfactual modeling and prediction of plausible society features in such parallel imaginary worlds, with outcomes of computer simulations and statistical estimations presented in tables. Chapter 11 of “Conclusion” summarizes the key achievements on the testing policy options in a realistic and reliable way for finding feasible and even optimal courses of actions via virtual experiments by varying parameters in the models of social changes. Directions of the further upgrading the models are discussed as well. Finally, the last third part of the book contains an appendix with multiple tables of the modeling and prediction outcomes in different scenarios.
A special software Simario for this project had been developed in R language, with Shiny application providing a user-friendly R web interface with a table builder. It is available in the public domain https://compassnz.shinyapps.io/SocialabShiny which permits to make modeling and forecasting of the developments for the entire societies. The code is stored at: https://github.com/kcha193/SocialabShiny. The Simario code for the models is available at: https://github.com/kcha193/simarioV2. More detail on SociaLab on the project pages at: https://www.compass.auckland.ac.nz. The monograph and corresponding software can serve as a valuable source in demographical, socio-economic and political studies for researchers and professionals, graduate students and academicians interested in the scientific innovations and practical estimations of governmental decisions and countries development in the contemporary civilization.
Minneapolis