11.02.2013 Views

Here - Tilburg University

Here - Tilburg University

Here - Tilburg University

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

The ASA Spring Methodology Conference<br />

Organized in Europe by the Department of Methodology and Statistics at <strong>Tilburg</strong><br />

<strong>University</strong>, the Netherlands.<br />

SESSION: Applications of Latent Variable Models<br />

SUMMARY<br />

For many constructs of interest in the social sciences, educational<br />

measurement, psychology, biology, and economics, no direct method exists for<br />

measurement. Nonetheless, examples of such constructs are legion, for instance<br />

think of political attitudes, abilities, personal traits, or product preferences. To<br />

get a hold of such constructs, researchers gather observable variables (hereafter<br />

called indicators, manifest variables, or items) which they hope will provide<br />

indirect evidence for the constructs of interest. Latent variable models are<br />

statistical models built to quantify and help objectify this type of inference by<br />

deriving a small set of latent unobserved variables that is underlying to the set<br />

of manifest variables and should reflect the constructs of interest. Well-known<br />

instances of this type of modeling approach are factor analysis (Thurstone,<br />

1947), latent class and latent profile analysis (Lazarsfeld & Henry, 1968), and<br />

item response theory (Lord & Novick, 1968).<br />

Although the foundations of latent variable models were laid several years<br />

ago, it has taken quite some time before they became widely applied. This is<br />

mainly due to the statistical nature of these models as well as the sometimes<br />

complicated computations and algorithms needed to estimate latent variable<br />

models. With the recent advances in computation speed and optimization<br />

algorithms and the availability of general purpose software able to fit latent<br />

variable models, this has become less of an issue.<br />

This session is intended as a brief showcase of the possibilities that a<br />

latent variable model framework can offer for research in the social and<br />

behavioral sciences. Each presentation will fill in and illustrate one of the most<br />

familiar model instances in the framework (see Figure).

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!