Practical, example-driven introduction to maximum likelihood for the social sciences. Emphasizes computation in R, model selection and interpretation.

Since 1986, the British government, faced with dwindling budgets and growing calls for public accountability, has sought to assess the value of scholarly work in the nation’s universities. Administrators have periodically evaluated the research of most full-time academics employed in British universities, seeking to distribute increasingly scarce funding to those who use it best. How do such attempts to quantify the worth of knowledge change the nature of scholarship? Juan Pablo Pardo-Guerra examines the effects of quantitative research evaluations on British social scientists, arguing that the mission to measure academic excellence resulted in less diversity and more disciplinary conformity. Combining interviews and original computational analyses, The Quantified Scholar provides a compelling account of how scores, metrics, and standardized research evaluations altered the incentives of scientists and administrators by rewarding forms of scholarship that were closer to established disciplinary canons. In doing so, research evaluations amplified publication hierarchies and long-standing forms of academic prestige to the detriment of diversity. Slowly but surely, they reshaped academic departments, the interests of scholars, the organization of disciplines, and the employment conditions of researchers. Critiquing the effects of quantification on the workplace, this book also presents alternatives to existing forms of evaluation, calling for new forms of vocational solidarity that can challenge entrenched inequality in academia.

This book is designed to introduce doctoral and graduate students to the process of conducting scientific research in the social sciences, business, education, public health, and related disciplines. It is a one-stop, comprehensive, and compact source for foundational concepts in behavioral research, and can serve as a stand-alone text or as a supplement to research readings in any doctoral seminar or research methods class. This book is currently used as a research text at universities on six continents and will shortly be available in nine different languages.

Time series, or longitudinal, data are ubiquitous in the social sciences. Unfortunately, analysts often treat the time series properties of their data as a nuisance rather than a substantively meaningful dynamic process to be modeled and interpreted. Time Series Analysis for the Social Sciences provides accessible, up-to-date instruction and examples of the core methods in time series econometrics. Janet M. Box-Steffensmeier, John R. Freeman, Jon C. Pevehouse and Matthew P. Hitt cover a wide range of topics including ARIMA models, time series regression, unit-root diagnosis, vector autoregressive models, error-correction models, intervention models, fractional integration, ARCH models, structural breaks, and forecasting. This book is aimed at researchers and graduate students who have taken at least one course in multivariate regression. Examples are drawn from several areas of social science, including political behavior, elections, international conflict, criminology, and comparative political economy.