Katrin Golsch

Applied panel data analysis for economic and social surveys

Introduction: Benefits and challenges of the panel design.- Outline of the book.- Audience and prerequisites.- Acknowledgements.- Managing panel data: The nature of panel data.- The basics of panel data management.- Three case studies on poverty in Germany.- How to represent a population with panel data?- Conclusion and further reading.- Describing and modeling panel data: Some basic terminology.- Measurements over time are not independent.- Describing the dependent variable.- Explaining the dependent variable over time: typical explanatory variables.- Modeling panel data.- Estimating models for panel data.- Overview of subsequent chapters.- Panel analysis of continuous variables: Modeling the level of Y.- Modeling the change of Y.- Conclusion and further reading.- Panel analysis of categorical variables: Modeling the level of Y: Discrete response models for panel data.- Modeling the change of Y: Discrete-time event history models for panel data.- Conclusion and further reading.- Conclusion: How to do your own panel analysis.- A: Functions of random variables.- B: Estimation and testing: Ordinary least squares.- How to compute a regression model fitting the data?- Sampling and sampling errors.- How to choose between different estimation methods?- How to estimate the parameters of an unknown population with a sample of data?- How to test parameters of an unknown population with a sample of data?- Maximum likelihood .- C: Web site of the textbook.- Bibliography.- Index.

Employment Flexibility in Spain and its Impact on Transitions to Adulthood

In contemporary societies, the growing need for flexibility and its impact on labour markets are highly debated. The deregulation of labour markets has led to a growing proportion of workers in various atypical employment relationships, such as temporary jobs. In this respect Spain stands out, from a European perspective, as a country that has experienced a fast and intense shift from one of the most rigid employment protection systems to a highly flexible labour market with an extraordinarily high proportion of temporary employees. The Spanish labour market is highly segmented with ‘insiders’ in permanent positions, and ‘outsiders’ (unemployed and inactive) and work-insecure employees in temporary posts, particularly among youth. This article attempts to gauge the impact of increasing flexibility on transitions to adulthood in Spain. Using the Spanish European Community Household Panel, the analysis concentrates on three interrelated transitions: labour market entry, first marriage and first parenthood. Previous contributions have well elaborated the repercussions of non-employment. What makes this study distinct is the particular focus on precarious employment relationships and individuals’ perception of job security. The results give a first indication that the type of employment relationship and the perception of job security seem to matter in the studied transition processes.

La flessibilità come principio guida: conseguenze sull’inserimento lavorativo, i piani di vita individuali e le decisioni legate alla formazione di una famiglia da parte dei giovani in Gran Bretagna

Questo articolo sintetizza le implicazioni che l’incremento del livello di deregolazione del mercato del lavoro e della flessibilita hanno avuto sui giovani in Gran Bretagna negli ultimi decenni. La sintesi dei dati offerti si sofferma sulla transizione scuola-lavoro e la qualita del primo impiego ottenuto, la percezione dell’insicurezza del lavoro, il salario ottenuto e le prospettive di promozione, i rischi di disoccupazione e le opportunita di uscita da questi stato, i rischi di mobilita discendente, la formazione di una unione di coppia e la transizione alla paternita/maternita.

How to Do Your Own Panel Analysis

The textbook is intended as a motivation for social scientists and other researchers to pursue their own research projects using panel data. To this end, this chapter provides an overview over existing panel studies that are available for the scientific community. Moreover, it points to the more applied literature that introduces interested readers into the practicalities of the panel design. It also directs the reader to the more specialized panel literature in the various social sciences disciplines.

Useful Background Information

The last chapter includes useful background information (e.g., the book’s website with example data sets and syntax files) and reviews some basic statistical terms (random variables, estimation, testing) that are helpful to understand the statistical discussions in the text.

Panel Analysis of Continuous Variables

The chapter discusses statistical techniques to specify and estimate regression models for continuous dependent variables. They focus either on the level or the change of the dependent variable. While most techniques explicitly model the panel design, some only control for the serial correlations, e.g., by using robust standard errors. Ordinary or generalized least squares is used for model estimation. The discussion includes first differences as well as fixed and random effects estimation and their variants.

Panel Analysis of Categorical Variables

The chapter discusses statistical techniques to specify and estimate regression models for categorical dependent variables. Again, they focus either on the level or the change of the dependent variable. As an example of the former type, the logistic regression model with fixed and random effects is introduced. Discrete-time event history models are used as examples of the second type. All models are estimated with maximum likelihood.

Describing and Modeling Panel Data

Since the panel design observes the same units repeatedly over time, observations at time point t will not be independent of observations made before t. This serial correlation contradicts the assumption of independent observations, which forms the basis of many classical methods of statistical inference. While the independence assumption is easily defended in case of cross-section data, its failure in case of panel data has to be explicitly dealt with. The chapter illustrates these statistical dependencies with simple examples using continuous and categorical variables that have been measured repeatedly over time. Statistical techniques are introduced to describe not only these serial correlations, but also the trend and the change of these variables. These descriptive analyses are followed by a discussion of the sources of these statistical dependencies and how cross-sectional regression models have to be extended to deal with them. The chapter concludes with a discussion of some important specification errors when applying these models to panel data.

Managing Panel Data

Panel data are usually more complex than cross-section data. Hence, before starting a statistical analysis, it is necessary to extract the relevant information from a larger data base often consisting of several data files. Panel data can be stored for statistical analysis either in wide or in long format. The chapter introduces the main data management operations to build such analysis files from a larger panel data base, such as the German Socio-Economic Panel (SOEP). In three case studies it is shown how software commands that are available in all major statistical program packages (such as merge and append) can be applied to SOEP data. The chapter concludes with a discussion of balanced and unbalanced panel data and possible countermeasures against such unit non-response.