The course consists of three parts, which may be taken individually or as a whole:
I Windows Pre-session: The pre-session is intended to make the SAS sessions accessible to those without previous experience with Windows on personal computers. Only the bare essentials of Windows will be covered; those familiar with Windows need not attend.
II Basic Introduction: Sessions One and Two provide an overview of SAS and its underlying logic; an explanation of the use of the Display Manager System to run a SAS job; an introduction to the SAS Data step for reading, transforming, and storing data; and a demonstration of how statistical analyses may be performed in SAS Proc (procedure) steps.
III Intermediate Topics: Sessions Three and Four will concentrate on SAS programming techniques to modify data and enhance SAS output. As well, more statistical procedures will be introduced.
This course consists of three parts, which may be taken individually or as a whole:
I Windows Pre-session: The pre-session is intended to make the SPSS sessions accessible to those without previous experience with Windows on personal computers. Only the bare essentials of Microsoft Windows will be covered.
II Basic Introduction: Session One is an elementary introduction to statistical computer programs, computing concepts, and the essentials of SPSS. At the end of the first session, participants should be able to run very simple programs, including some basic descriptive statistical procedures. Session Two will cover first-session topics in greater detail, concentrating on data definition facilities and various ways of formatting data.
III Intermediate Topics: Sessions Three and Four will introduce data modification, transformations, and functions. Session Five will cover the use of SPSS system files.
This course provides an introduction to the theory, methods, and empirical applications of CFA within the "LISREL" framework.
The course will cover the specification of: 1) classical test theory models; 2) the multitrait-multimethod model; 3) the second-order factor model; 4) longitudinal factor analysis; and 5) multi-sample analysis including the estimation of latent means. The course will also address estimation problems (improper solutions), and the assessment of model fit.
The course will be of interest to those who are currently using EFA and find that their research problems are more appropriately analyzed with CFA, and to those who are interested in the general structural-equation ("LISREL") model.
Familiarity with elementary matrix algebra will be useful, though not essential, for understanding LISREL syntax.
In a "policy-capture" approach to creating a policy for a large firm, factor analysis is sometimes used to help define job factors. Multiple regression is used to determine the weights to attach to each job factor in computing the value of each job.
While these statistical tools can provide powerful insights they can also, when used without sufficient understanding, produce results worse than those that might have been obtained without statistical methods.
This course will explore some of the policy pitfalls in the use of statistics for pay equity. The thesis of the course is that statistics has a powerful but circumscribed role in pay equity. It is important to appreciate the limitations of statistical techniques so that decisions with important consequences not be entrusted to a statistical procedure that is inappropriate for the task expected of it.
Some of the topics we will consider are:
Researchers in a variety of fields use clustering techniques to search for structure in (often) large multidimensional data sets. The goal is to uncover homogeneous or similar subgroups where similarity is often measured by distance between observations within groups. A principal drawback of traditional approaches to clustering lies in testing hypotheses about the number of clusters present in the data. This course will review some of this literature and point to how problems arise. It will also illustrate a new regression-based approach to clustering where it is possible to test hypotheses about the number of clusters.