Assessing Emotional Outcomes in Intervention Study

During weekly lab visits for the pre- and post-intervention measures, we asked participants to take affect-related questionnaires including State Trait Anxiety Inventory (STAI)^{64 (link)}, Center for Epidemiological Studies-Depression (CESD)^{65 (link)}, Profile of Mood State (POMS)⁶⁶ . As POMS includes multiple mood states, we only averaged negative mood items for this analysis. As the three types of questionnaires were measured at Week 1 and 2, we had a total of six pre-intervention emotion measures for each participant. Likewise, we also have a total of six post-intervention emotion measures. To create summary scores for negative affect, we first standardized each measurement separately for younger and older participants. We next ran a principal component analysis on the six normalized pre-intervention measures separately for younger and older adults. This resulted in the pre-intervention summary score as well as one dominant component whose coefficient matrix elements indicate factor loadings for the six emotional measures. Using those matrix elements as weights, we calculated the weighted average of the six corresponding normalized post-intervention emotional measures (Week 6 and 7) to obtain post-intervention summary scores.