Thread by @DejonckheerEgon

1/ THREAD: Why the role of emotion dynamics in psychological well-being may be smaller than previously thought: a summary of our (@merijn_mestdagh and me) BEAT THE MEAN project in 12 tweets.

ED = emotion dynamics, PSY WB = psychological well-being.

2/ Original propositions:

(a) Feelings come and go. Emotions are inherently time-dynamic.

(b) Emotions are important markers for PSY WB.

-> Therefore (a+b): next to how good or bad we feel on average, ED may convey important information too about our PSY WB!

3/ The rising interest in the link between ED and PSY WB lead to a plethora of newly proposed measures, each claiming to evaluate a unique dynamical aspect of a one’s emotional life.

(e.g., emotion variability, instability, inertia, differentiation, density, diversity, etc.)

4/ Although this surge in ED suggests scientific progress, it remains unclear whether they all add to our understanding of well-being and psychopathology.

(a) How do all ED measures empirically interrelate?

(b) What is their incremental value over average PA and NA?

5/ To address both issues we combined affective time series data from 15 different studies (N = 1,777), and evaluated the relation between 16 commonly studied ED measures and 3 prominent markers of PSY WB (life satisfaction, depressive and borderline symptoms).

6/ Regarding (a): all ED are interrelated considerably, suggesting many dynamics merely approach the same data from a different angle. Some measures, however, summarize affective time series relatively independent of other ED measures (e.g., affective bipolarity & inertia).

7/ Regarding (a): interrelatedness between ED is also supported by PCA, adequately summarizing all measures in 6 independent components (R² = 82%). Four of the underlying components directly relate to average levels of, and/or variability in PA and NA.

8/ Regarding (b): meta-analysis shows that all ED individually show a meaningful relation with various indicators of PSY WB (e.g., Houben et al., 2015). Yet, when controlling for average levels of (and variance in) PA an NA, many relations disappear.

ncbi.nlm.nih.gov/pubmed/25822133

9/ Regarding (b): Finally, LASSO analyses indicate that mean levels of PA and NA are indeed the most prominent indicators of PSY WB. Moreover, in many cases, adding other ED will not yield a more accurate prediction of individual differences in PSY WB.

10/ In sum, our findings put the added value of ED measures in perspective. Based on Ockham’s razor, we believe caution is warranted in developing novel and complex measures of ED for the prediction / explanation of well-being and emotion disorder.

11/ Do the current results completely reject the importance of ED in understanding PSY WB? No! We merely established that conventional emotion research is currently unable to demonstrate meaningful and independent relations between ED and PSY WB. Future studies should look into:

12/ Read the full paper in @NatureHumBehav here: nature.com/articles/s4156…

Access code + some of the materials at OSF here: osf.io/zm6uw/.

Or come to our invited talk at the APS @PsychScience conference May 23rd!