Statistical and Methodological Tools
We value acquiring new skills that help us to ask increasingly sophisticated questions! Some background about the methodological sophistication in our lab follows.
Structural equation models (SEM) have been part of our experience from the earliest days of AMoS. Pruchno, Patrick & Burant (1996) and Patrick & Hayden (1999) are among the first caregiving studies to use SEM. Much of our current work builds on these foundational skills. We are learning how to apply Bayesian statistics to examine meaningful null age effects. We have significant expertise in conducting and analyzing short- and long-term experience sampling (ESM) studies and have recently published on how to plan, implement and analyze them (Nehrkorn-Bailey et al., 2018). Using various ESM data sets, we have applied Hierarchical Linear Models (HLM) to examine fluctuations in gratitude and spirituality (Olson et al., 2018) and we have used Generalized Estimating Equations (GEE; Graf, Long & Patrick, 2017) to understand the influence of hassles and uplifts on self-assessed health. We have used Latent Growth Curve Models (LGCM) to examine how changes in the environment relate to changes in well-being over 17 years (Knepple Carney et al., 2017) and to examine stability in well-being in the early days of COVID-19 (Ebert, Bernstein, & Patrick, 2020). We have also explored different patterns of responses using Profile Analyses (Graf et al., 2015; Clark & Patrick (2022). We have also examined age as a moderator of bereavement outcomes (Patrick & Henrie, 2015) and of religious doubt (Patrick, Bernstein & Moore, 2021). Finally, an early paper (Ayotte et al., 2013) used the Actor-Partner Interaction Model (APIM in SEM) to examine dyadic influences on physical activity, as shown in the figure to the right. Thus, members of the Healthy Aging Lab are fearless in acquiring the tools we need to answer important questions.