Judith Singer | Harvard University (original) (raw)

Papers by Judith Singer

Abstract In this paper, we present an exploratory methodology for detecting and documenting an in... more Abstract In this paper, we present an exploratory methodology for detecting and documenting an influence on the duration of teacher employment which has heretofore eluded empirical quantification—the involuntary layoff. Using data on the lengths of employment of more than 14,000 teachers hired between 1969 and 1981 in the St.

Abstract Since the implementation in 1977 of the Education for All Handicapped Children Act (Publ... more Abstract Since the implementation in 1977 of the Education for All Handicapped Children Act (Public Law 94-142), public school systems have provided special education and related services to students with a wide range of handicapping conditions, including some children served previously in hospitals or other institutions.

Researchers often examine how individual change over time depends on selected predictors by fitti... more Researchers often examine how individual change over time depends on selected predictors by fitting a multilevel modelfor change. Generations of behavioral scientists have been interested in measuring and investigating individual change, but for decades, the prevailing view was that it was impossible to do well (Cronbach and Furby, 1970).

Abstract A survey was made of the parents of 380 children whose mobility impairments require the ... more Abstract A survey was made of the parents of 380 children whose mobility impairments require the use of a wheelchair, walker, or braces. They were asked about equipment, health services, related services, and family support services used during the previous year.

Psychologists studying whether and when events occur face unique design and analytic difficulties... more Psychologists studying whether and when events occur face unique design and analytic difficulties. The fundamental problem is how to handle censored observations, the people for whom the target event does not occur before data collection ends. The methods of survival analysis overcome these difficulties and allow researchers to describe patterns of occurrence, compare these patterns among groups, and build statistical models of the risk of occurrence over time.

Abstract This paper analyzes changes in the teacher salary schedules of Michigan school districts... more Abstract This paper analyzes changes in the teacher salary schedules of Michigan school districts between 1970 and 1980. We find that starting salaries, expressed in 1970 dollars, decreased by an average of 20% over the decade. Real maximum salaries decreased by 15%. The between-district variability of starting salaries increased markedly over the decade, making the average starting salary a much poorer estimate of the starting salary a particular teacher earned in 1980 than was the case in 1970.

How do students learn? What facilitates or impedes learning? Of the myriad potentially important ... more How do students learn? What facilitates or impedes learning? Of the myriad potentially important factors, which ones are the most important? These questions, at once deceptively simple and perilously complex, lie at the heart of much educational research. A universal answer has eluded scholars and philosophers for centuries and remains, for some, the holy grail of educational research today. Does the persistence of these questions suggest that we, as a community of scholars, have learned nothing in our decades of study?

Page 1. Singer & Willett, 2003 October 13, 2003 © Harvard University 1 Doing Data Analysis in the... more Page 1. Singer & Willett, 2003 October 13, 2003 © Harvard University 1 Doing Data Analysis in the Multilevel Model for Change Doing Data Analysis in the Multilevel Model for Change Judy Singer & John Willett Harvard University Graduate School of Education What we will cover? What we will cover?

When no dedicated weighted least-squares procedure is available, misapplying a case-weighting str... more When no dedicated weighted least-squares procedure is available, misapplying a case-weighting strategy as a method of implementing a weighted least-squares regression analysis can lead to gross inaccuracies. The magnitudes of many of the obtained estimates depend strongly on the absolute magnitudes of the weights applied during fitting and, in addition, several of the crucial regression estimates are incorrect.

Service courses in applied statistics abound in college and graduate school curricula around the ... more Service courses in applied statistics abound in college and graduate school curricula around the world [34]. In a recent survey of United States graduate programs in health professions, for example, Cockerill and Fried [12] found that all programs required their students to take at least one statistics and research methods course. In a similar survey of psychology departments, Aiken, West, Sechrest and Reno [1] found that the same was true for nine out of every ten doctoral programs.

ABSTRACT Many questions arising in drug abuse prevention and intervention studies focus on whethe... more ABSTRACT Many questions arising in drug abuse prevention and intervention studies focus on whether and, if so, when events occur. When do adolescents start using drugs? Does participation in a drug prevention program at school decrease the risk that high school students will initiate drug use? Does failure to participate in a relapse prevention program at a community health center increase the risk that newly abstinent ex-abusers will start using drugs again?

Bereits im 19. Jahrhundert formulierte John S. Mill eine einfache aber unerlässliche Voraussetzun... more Bereits im 19. Jahrhundert formulierte John S. Mill eine einfache aber unerlässliche Voraussetzung für kausale Beziehungen: "the cause has to precede the effect in time". Für eine aufgeklärte Wissenschaft war damit der Königsweg für die Analyse kausaler Beziehungen vorgezeichnet, der allein dadurch beschritten w erden kann, dass im Experiment die durch einen bewusst gesetzten Stimulus behandelte Versuchgruppe eine Veränderung erfährt und sich darum -nachdem eine gewisse Zeit verstrichen ist -von der Kontrollgruppe unterscheidet. Kausale Analyse im experimentellen Design impliziert eine Längsschnittperspektive, weil Messungen der abhängigen Variable sowohl vor als auch nach Verabreichung des experimentellen Stimulus durchgeführt werden und sich bei der zweiten Messung ceteris paribus die Unterschiede zwischen den im Idealfall zufällig zusammengesetzten Gruppen logisch zwingend auf den Stimulus zurückführen lassen. So überzeugend die Rationalität der experimentellen Erkenntnismethode auch ist -in der deutschen Soziologie stellt sie eher eine Randerscheinung dar und man greift zumeist auf quasi-experimentelle Designs zurück. 1 Allerdings ermöglicht eine Längsschnittperspektive auch in diesen Studien kausale Analysen, die, wenngleich nicht ohne Theorie, wesentlich näher an der Empirie orientiert sein können, als Querschnittsanalysen. Analysen über die Zeit ermöglichen zeitveränderliche Messungen der abhängigen und unabhängigen Variablen. Auf diese Weise nähert sich die häufig auf Umfragedaten basierende Kausalanalyse dem experimentellen Design wieder an, da untersucht werden kann, inwieweit der Prozess der Veränderung der abhängigen Variable aus dem Prozess der Veränderung von unabhängigen Variablen resultiert. Daher ist es zu begrüßen, wenn in Lehrbüchern systematisch und verständlich in statistische Verfahren der Längsschnittanalyse eingeführt wird. Das im Jahre 2003 erschienene Buch "Applied longitudinal analysis" von Singer und Willet erfüllt diese Aufgabe in sehr innovativer Weise, da erstmals zwei unterschiedliche Zugänge zur Längsschnittanalyse ausführlich dargestellt und miteinander verglichen werden. Es handelt sich dabei einerseits um Messwiederholungsmodelle, die als "multilevel model for change" bezeichnet werden, andererseits um Hazardmodelle für diskrete und stetige Zeit, in der deutschsprachigen Soziologie eher bekannt unter der Bezeichnung "Ereignisanalyse". Im Prinzip ist ein derartiges Lehrbuch schon lange überfällig, da die beiden methodischen Ansätze des Messwiederholungsdesigns (auch bekannt als Panelanalyse) und der Ereignisanalyse zumeist in unterschiedlichen Forschungszweigen Anwendung finden und leider manchmal Unkenntnis herrscht über die Angemessenheit oder über mögliche Vorteile des jeweils anderen methodischen Zugangs. 1 Vgl. zu sozialwissenschaftlichen und medizinischen Experimenten insbesondere David M. : Design and Analysis of group randomized trials. Oxford: University Press.

Journal of Educational and Behavioral Statistics, 1998

Abstract 1. PROC MIXED is a flexible computer program suitable for fitting individual growth mode... more Abstract 1. PROC MIXED is a flexible computer program suitable for fitting individual growth models to data. Its position as an integrated program within the SAS statistical package makes it an ideal choice for empirical researchers seeking to do data reduction, management, and analysis within a single program. Because PROC MIXED was developed from the perspective of a&amp;amp;quot; mixed&amp;amp;quot; statistical model with random and fixed effects, its syntax and programming logic may appear unfamiliar to users who express individual growth ...

Journal of Educational and Behavioral Statistics, 1995

American Statistician, 1988

… of Research Synthesis, Harris Cooper and …, 1994

A large proportion of meta-analyses in education and the social sciences, including many that we ... more A large proportion of meta-analyses in education and the social sciences, including many that we consider excellent, pose a surprising challenge. The reader must struggle to figure out exactly what the researchers found! How are the results of meta-analyses currently presented? Typically, and especially in the social sciences and education, readers are offered two types of summary table. The first table gives a lengthy list of the actual studies that were included in the meta-analysis, together with their date of publication, sample ...

American Statistician, 1990

Review of Educational Research, 1991