Featured Articles

Constantino LagoaMarch 5, 2014

In control engineering, devices continually monitor the performance of a system and then operate to control that system. Common examples include automobile thermostats and autopilot systems on commercial airliners. These same principles can be used to design behavioral interventions that adapt over time to help patients alter behaviors that affect their health. Potential applications include maintaining an exercise regimen, maintaining a healthy diet, or abstaining from tobacco or illicit drug use.

 

In a forthcoming article in the Journal of Consulting and Clinical Psychology, Methodology Center Investigators Constantino Lagoa, Stephanie Lanza, Susan Murphy, and their colleague Korkut Bekiroglu describe this new approach to building adaptive, intensive interventions. The authors show how data, in this case simulated smoking-cessation data, can be used to inform the design of an adaptive, intensive intervention by applying control-engineering techniques. This intervention, which is designed to provide treatment only when needed, is shown to improve effectiveness while decreasing patient burden. This approach holds great promise for informing clinical decisions and for informing the development of smartphone-based adaptive interventions. 

Linda CollinsDecember 10, 2013

The multiphase optimization strategy (MOST) provides a framework for building effective and efficacious behavioral interventions by following principles from the field of engineering. MOST emphasizes efficiency and resource management to move intervention science forward, and one of the cornerstones of MOST is selecting the most efficient and appropriate design for each experiment. In many instances this means conducting a factorial experiment. (Read more about factorial experimental designs.) In a new article in Translational Behavioral Medicine, Linda Collins and her coauthors explore how to use data from a factorial screening experiment to decide what should be included in the experimental intervention treatment package. This article is aimed at scientists who are considering conducting a factorial screening experiment.

Will quitting smokers lapse?November 15, 2013

In an article in a forthcoming special issue of Nicotine and Tobacco Research, a team of Methodology Center scientists examines the changing relationships between factors that predict relapse for smokers who are quitting. The researchers, led by Research Associate Sara Vasilenko, used time-varying effect models to analyze ecological momentary assessment data and found that predictors of relapse changed over the two weeks post quit. Cigarette cravings significantly predicted relapse throughout the two weeks. However, baseline dependence was significant early in the quit process but was unimportant after one week. Conversely, negative mood became a stronger predictor of relapse as time progressed.

How to Use Propensity Scores for Causal InferenceOctober 21, 2013

Researchers often would like to draw causal inferences from observational data; however, confounders, variables related to both the predictors and the outcome, typically undermine the validity of these inferences. In a recent article in American Journal of Community Psychology, Methodology Center Scientific Director Stephanie Lanza and her collaborators demonstrate propensity scores as a straightforward method for drawing causal inferences from observational data. As an empirical demonstration, the authors estimate the causal effect of Head Start versus parental care on later reading development.

Bethany BrayStephanie LanzaSeptember 24, 2013

Latent class analysis (LCA) has become a popular tool for identifying subgroups within populations. Despite the fact that these are latent subgroups and, therefore, membership in each class is unknown, research questions sometimes make it necessary to assign individuals to classes and then treat class membership as known in later analyses. Current standard practice is to retain from LCA each individual’s posterior probabilities of class membership, and then either assign them to the most likely class or to repeatedly assign them using pseudo-class draws. These practices, however, are known to attenuate estimates in subsequent analyses.

 

In a new article in Structural Equation Modeling, "Eliminating bias in classify-analyze approaches for latent class analysis," by Center Investigators Bethany Bray and Stephanie Lanza and Xianming Tan of McGill University, the authors propose a straightforward solution that includes all variables used in the subsequent analyses as covariates in the LCA. Then, class assignment is carried out using either the most likely class or pseudo-class draws. With adequate measurement quality and sample size, this approach substantially reduces the bias that results from common approaches to class assignment.

Nicole ButeraStephanie LanzaDonna Coffman

August 1, 2013

Donna Coffman and Stephanie Lanza lead the Methodology Center’s research projects on causal inference and latent class analysis (LCA), respectively. Recently, they began integrating modern causal inference techniques with LCA to reveal the impact that nonrandomized risks have on later, complex, behavioral outcomes. In a new article in Prevention Science, “A Framework for Estimating Causal Effects in Latent Class Analysis: Is There a Causal Link Between Early Sex and Subsequent Profiles of Delinquency?” Center researchers Nicole Butera, Stephanie, and Donna examine data on 1,890 adolescents from the National Longitudinal Study of Adolescent Health to determine whether early sexual initiation impacts later profiles of delinquent behavior.

Featured Article: Can We Prevent Children’s Adjustment Problems by Teaching Parenting Skills? March 7, 2013

Emotional problems between ages one and three are a risk factor for developing more serious problems later in life, and researchers sometimes seek to prevent these problems through parenting-skills education. In the article “The Effects of the Family Foundations Prevention Program on Coparenting and Child Adjustment,” to appear in Prevention Science, a team of Penn State researchers including Methodology Center Principal Investigator Donna Coffman examine the pathways through which one such prevention program works.

Teens Becoming Regular Smokers: The Roles of Smoking Quantity and Nicotine Dependence February 4, 2013

Nicotine dependence plays a central role in the development of regular smoking, but the relationship is reciprocal: nicotine dependence causes people to smoke more, and regular smoking leads to nicotine dependence. To understand how people become regular smokers, it is important to understand the roles that nicotine dependence and smoking quantity play in the frequency of smoking, and how the strengths of these relationships change over time. This issue is addressed in the article in Drug and Alcohol Dependence, “Time-varying effects of smoking quantity and nicotine dependence on adolescent smoking regularity,” by Arielle Selya, Lisa Dierker, Jennifer Rose, Donald Hedeker, Xianming Tan, Runze Li, and Robin Mermelstein.

Understanding Childhood Health Disparities with LCA

January 4, 2013

Research indicates that early childhood is a critical period of life that impacts many long-term health outcomes. By understanding disparities in early childhood health among different segments of the U.S. population, researchers can address the greatest needs within our society. In the article “Measuring Early Childhood Health and Health Disparities: A New Approach” which appears in the Maternal and Child Health Journal, Penn State researchers Marianne Hillemeier, Stephanie Lanza, Nancy Landale, and Sal Oropesa provide one of the most comprehensive examinations to date of child health and health disparities in the United States. Traditionally, disparities have been examined in terms of a specific health problem (e.g., differential asthma or obesity rates across ethnicities or socio-economic groups). By applying latent class analysis (LCA)—a tool for identifying hidden subgroups in a population—to a national sample of four-year-old children, the authors were able to simultaneously examine several important components of health: health conditions (e.g., obesity, asthma), functioning (e.g., vision, hearing, overall activity level), fine motor skills, emotional wellness (e.g., empathy, externalizing behavior), social skills, and cognitive achievement, across ethnicities and socio-economic groups.

Featured Article: Does Attending College Lead to Later Drinking Problems?October 26, 2012

College is often perceived as a risky environment for problem drinking, but recent studies indicate that individuals who attend college go on to engage in this behavior in adulthood at equal or lower rates than those who do not attend college; that is, that college may actually protect individuals from substance use behaviors in adulthood. These studies, however, often fail to account for selection bias: the fact that the people who attend college are different in many ways than people who do not attend college. In the article “Causal Inference in Latent Class Analysis,” which will appear in Structural Equation Modeling, Methodology Center researchers Stephanie Lanza and Donna Coffman implement two propensity score techniques for causal inference in latent class analysis (LCA) to determine whether college enrollment is protective or harmful for substance use behavior later in life.

Follow Us