News and Updates


Intensive Longitudinal Data and Fitness Trackers

November 17, 2016

ild2Nearly all of us carry a powerful computer throughout our waking hours in the form a smart phone. Remarkably, even the most basic iPhone has vastly more computational power than the entire set of computers that guided Apollo 11 to the moon. The ubiquity of smart phones has led to the easy (and potentially too easy) collection of vast amounts of data collected over time. Indeed, one of the most vexing current challenges in longitudinal data analysis is determining how to best fit meaningful statistical models to high-density repeated measures to test specific hypotheses of interest. For example, a recent article in the New York Times summarized two published studies that examined the impact of fitness tracking on health and well-being. In one study, 4000 subjects were followed over a decade and, on average, those who exercised at least 150 minutes-per-week were associated with a one-third decrease in premature death. In a second study, subjects who were paid cash to meet their exercise goals showed slight increases in their activity as recorded by the fitness trackers. However, an array of study limitations exist, not the least of which is validly establishing the proper causal direction of effect. Regardless, smart phones offer a plethora of exciting opportunities for the collection of high-density repeated measures data, yet the subsequent analysis and interpretation remains a challenge.

Read More

Psychometrics in Everyday Life

November 6, 2016

cfaThe term psychometrics is broadly defined as the theory and methods of mental measurement. The historical roots of psychometrics date back more than a century, the birth of which is sometimes marked by Spearman’s seminal 1904 paper titled General Intelligence, Objectively Determined and Measured in which factor analysis was first described as a formal statistical method. Over the following decades, psychometric methods and principles have been applied to virtually every aspect of human emotion, personality, and cognition. In a recent article in The Economist, the use of psychometrics for personality assessment is explored as a new form of data to consider when financial institutions make loans. This information is particularly salient in many areas of the developing world where credit scores are not readily available. The article suggests, for example, that young people who are deemed optimists are a high risk for loan repayment, but older optimists are a safer bet. This interesting application is just one example of how psychometric methods are being used to address real-world problems, and there are a myriad of other applications in hiring, admission to college, setting of insurance rates, and end-of-grade testing. Given the importance of psychometrics in everyday life, it is critically important that trained psychometricians be involved in all aspects of scale development and assessment. However, there is a well-documented shortage of individuals with the necessary level of training to accomplish this challenging task. Being fully informed about the potential strengths and inherent limitations of psychometrics is critical regardless of application or field of study.

Read More

Summer Workshop Schedule

October 18, 2016

networkWe’re pleased to announce our expanded summer workshop schedule for 2017:

  • May 15 — May 19: Structural Equation Modeling
  • May 22 — May 26: Longitudinal Structural Equation Modeling — Now offering Stata demonstrations
  • May 31 — June 2: Network Analysis — New this year
  • June 5 — June 9: Multilevel Modeling
  • June 12 — June 16: Cluster Analysis and Mixture Modeling
  • Register Now

    Read More

    Syntax for Computing Random Effect Estimates in SPSS

    September 27, 2016

    histoMany programs can be used to fit multilevel models. For instance, in our multilevel modeling summer workshop, we demonstrate three programs: SAS, SPSS, and Stata. Unlike SAS, Stata, and many other programs, however, SPSS does not currently offer the option to output estimates of the random effects. Obtaining estimates of the random effects can be useful for a variety of purposes, for instance when producing plots or evaluating model assumptions. To overcome this limitation of SPSS, we have developed documentation and syntax for computing random effects in SPSS based upon output that is easily obtained from the MIXED command.

    Read More

    The Human Desire to Categorize

    July 4, 2016

    traj2Whether colors in the rainbow, notes in a musical scale, types of coffee, or country versus pop singers, there is a natural human desire to categorize objects and experiences. A story that recently appeared in the New York Times by Tom Vanderbilt presents a wonderful exploration of how we all find comfort in defining, seeking out, and confirming categories in what he calls the “psychology of genre”. Within many research applications it is challenging to know when categorizing individuals is appropriate, or how best to discern these categories with the data at hand. Extracting categories when variation is really continuous presents risks, but so too does failing to identify meaningfully distinct subgroups of individuals within the population. Further, many approaches exist for empirically identifying subgroups, including cluster analysis, latent class analysis, and finite mixture modeling. These techniques bring a level of statistical rigor to our natural desire to categorize, but can also be complex to implement in practice. As an accessible initial resource on this topic, we recommend Everitt et al. (2011) Cluster Analysis (5th Edition), published by Wiley.

    Read More