Issue: December 2012

eNews from The Methodology Center


"Holiday Goodies" for Latent Class Analysis

Stephanie Lanza and Bethany Bray

Podcast: Latent Class Analysis (LCA) Part 1 - Common Questions about LCA

In our latest podcast, Methodology Center scientists Stephanie Lanza and Bethany Bray discuss common, practical issues that arise in latent class analysis (LCA). Issues discussed include selecting indicator variables, selecting a model, determining minimum sample size, finding LCA software, and getting started in LCA. This is the first in a two-part podcast; the next podcast will address some of our recent research on LCA.

Download the podcast.

PROC LCA for SAS: Basic Features

Tutorial Video: Basic Features of PROC LCA

Are you new to PROC LCA? The SAS procedure for latent class analysis (LCA) developed by the Methodology Center is available for free. This video demonstrates the main features of the procedure including running an LCA model, using a grouping variable, specifying measurement invariance, running LCA with covariates, and specifying a reference class. The example in the video is based on Lanza, Collins, Lemmon, and Schafer (2007). Future videos will address other features of PROC LCA. If there are features you would like to see in a video, email with your suggestions.

Watch the video.


Interest Group: Optimizing Behavioral Interventions

Linda Collins

For those at Penn State, Linda Collins is organizing a discussion group on optimizing behavioral interventions that will meet in the spring semester. The purpose of the group is to bring intervention scientists and methodologists together to promote substantive and methodological research on optimizing behavioral interventions using approaches such as the multiphase optimization strategy (MOST), the sequential, multiple assignment, randomized trial (SMART), and related methods. The group will meet on a monthly basis (dates and times TBA).

Interested? Email and ask to be added to the group’s email list.


Featured Articles: Building Interventions That Adapt to Meet Individuals' Needs

Inbal Nahum-Shani, Daniel Almirall, and Susan Murphy

In adaptive interventions, the intervention options (such as the type, dosage, intensity, or content of the intervention) are individualized and adapted over time in response to the ongoing performance and changing needs of the participant. In recent years, investigators have become increasingly interested in obtaining empirical evidence that informs the construction of high-quality adaptive interventions, especially adaptive interventions that optimize long-term outcomes. Methodology Center scientists recently published two related papers in Psychological Methods in which they discuss experimental designs and data analysis methods specifically developed to help investigators construct adaptive interventions based on empirical evidence.

In the first, "Experimental Design and Primary Data Analysis Methods for Comparing Adaptive Interventions," Inbal Nahum-Shani, Daniel Almirall, Susan Murphy, and their colleagues describe the sequential, multiple assignment, randomized trial (SMART), an innovative experimental design that enables investigators to address critical questions concerning the sequencing and adaptation of an intervention over time. In the second article, "Q-Learning: A Data Analysis Method for Constructing Adaptive Interventions," the same authors discuss using Q-learning, a method drawn from computer science, to analyze data from SMART designs and to assess the quality (in terms of optimizing long-term outcome) of adaptive interventions.

Read more.



Journal Features SMART

Journal of Child and Adolescent Psychopharmacology

The most recent issue of Journal of Child and Adolescent Psychopharmacology selected the article "SMARTer Discontinuation Trials for Developing and Adaptive Treatment Strategy," as one of their features. The article, by Methodology Center researchers Daniel Almirall and Susan Murphy and their collaborators, describes using SMART for pharmaceutical discontinuation trials.

Read more on the journal’s website.


Find us on YouTube

Find us on YouTube

In the coming months, we are planning to release a greater number of videos. All of these - along with our existing videos - will be available on YouTube. So, if you are looking for help with specifying measurement invariance in your LCA model at 3 AM, we’ll be wherever the Internet is.

See the Methodology Center YouTube Channel.


Introduction to Time-Varying Effect Models (TVEM)

Our TVEM webpage explains some of the basic concepts underlying time-varying effect models. Questions answered include

  • Why does TVEM matter?
  • What is ILD and how many data points do I need?
  • If I have enough data points, how many subjects do I need?
  • If the macro does not provide a p-value, how do I publish my results

The page also contains an example based on smoking data.

View the webpage.


Follow us

Find us on Facebook View our profile on LinkedIn Follow us on Twitter

Methodology Minutes

The Methodology Center
204 East Calder Way, Suite 400
State College, PA 16801

Phone: 814-865-3253
Fax: 814-863-0000