Multimodality and Skewness in Emotion Time Series

Haslbeck, Jonas M B; Ryan, Oisín; Dablander, Fabian

doi:10.31234/osf.io/qudr6

Cited by 10 publications

(12 citation statements)

References 43 publications

(107 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To further explore these themes, such analyses can be complemented by exploring within-individual univariate distributional properties (e.g. 55 ).…”

Section: Limitations and Future Directionsmentioning

confidence: 99%

Delineating Discrete Generalizable States from Intraindividual Time Series: Towards a Science of Moments

Lazarus¹,

Song²,

Jeronimus³

et al. 2023

Preprint

View full text Add to dashboard Cite

The present paper pursues archetypal states—compound sets of concurrent, fixed-distance intervals in temporal variables that predict person-level data across persons. Using unsupervised learning, we identify a set of states defined by varying degrees of negatively correlated positive and negative affect. We demonstrate the consistency of these structures across three samples. Sample 1 (N=838) was split into N=500 training series and N=338 hold-out series. Training data were used to distill archetypal compound emotion states, which were validated across the hold-out sample and two external samples—a naturalistic sample of 179 participants and a sample of 45 individuals with depression and anxiety. Predictions of momentary variation in the out-of-sample data accounted for 40% to 50% of the variance in these unseen data. We propose that the current paradigm serves as a proof of concept for a novel and generative science of moments that provides means for transcending the idiographic-nomothetic divide.

show abstract

“…To further explore these themes, such analyses can be complemented by exploring within-individual univariate distributional properties (e.g. 55 ).…”

Section: Limitations and Future Directionsmentioning

confidence: 99%

Delineating Discrete Generalizable States from Intraindividual Time Series: Towards a Science of Moments

Lazarus¹,

Song²,

Jeronimus³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…The multilevel AR(1) model as presented above is based on several assumptions, two of which are of particular interest to us here: (a) the residuals at level 1 are normally distributed, and as a result, the within-person fluctuations of X t are characterized by the normal distribution; and (b) the random effects, including level-2 means, are assumed to come from a multivariate normal distribution. However, these assumptions are not always met in practice (Haslbeck et al, 2022). To illustrate this, we make use of a subset of the intensive longitudinal dataset collected in the COGITO study , in which 204 adults were measured once a day on various affective and cognitive items for up to 109 days.…”

Section: Normality Assumptionsmentioning

confidence: 99%

“…The first DGM is a generalized AR(1) model with χ 2 residuals (Tiku et al, 1999), which is suitable for generating skewed continuous-valued time series. Since the multilevel AR(1) model is often fit to Likert-scale data (Haslbeck et al, 2022), which can be considered ordinal or discrete in nature rather than continuous Weiß, 2018), the second and third DGMs we consider are for discrete-valued time series. The second DGM is the binomial AR(1) model (McKenzie, 1985), that can generate bounded time series of counts with the floor effect, which can be used to simulate data that represent a variable that is measured on, for instance, a 0 to k discrete scale.…”

Section: Alternative Data Generating Modelsmentioning

confidence: 99%

Skewness and staging: Does the floor effect induce bias in multilevel AR(1) models?

Haqiqatkhah¹,

Ryan²,

Hamaker³

2022

Preprint

View full text Add to dashboard Cite

Multilevel autoregressive models are popular choices for the analysis of intensive longitudinal data in psychology. Empirical studies have found a positive correlation between autoregressive parameters of affective time series and the between-person measures of psychopathology, a phenomenon known as the staging effect. However, it has been argued that such findings may represent a statistical artifact: Although common models assume normal error distributions, empirical data (for instance, measurements of negative affect among healthy individuals) often exhibit the floor effect, that is response distributions with high skewness, low mean, and low variability. In this paper, we investigated whether—and to what extent—the floor effect leads to erroneous conclusions by means of a simulation study. We describe three dynamic models which have meaningful substantive interpretations and can produce floor-effect data. We simulate multilevel data from these models, varying skewness but keeping the autoregressive parameter fixed across individuals, while also varying the number of time points and cases. Analyzing these data with the standard multilevel AR(1) model we found that positive bias only occurs when modeling with random residual variance, whereas modeling with fixed residual variance leads to negative bias. We discuss the implications of our study for data collection and modeling choices.

show abstract

“…First, a fairly common statistical challenge is that data, especially negative affect items, are highly skewed at the population level (e.g., Haslbeck et al, 2022). While estimating VAR models on variables that do not fully meet multivariate normality frequently occurs, given the nature of EMA data that is often ordinal, it likely reduces the power to detect small relations in the data.…”

Section: Limitations and Next Stepsmentioning

confidence: 99%

A Network Study of Family Affect Systems in Daily Life

Veenman¹,

Janssen²,

Houtum³

et al. 2022

Preprint

View full text Add to dashboard Cite

Adolescence is a time period characterized by extremes in affect and increasing prevalence of mental health problems. During this time, the family remains a crucial source of both support and stress. Prior studies have illustrated how affect states of adolescents are related to interactions with parents. However, it remains unclear how affect states among family triads, that is adolescents and their parents, are related in daily life. This study investigated affect state dynamics (happy, sad, relaxed, and irritated) of 60 family triads, including 60 adolescents, fathers and mothers. The families participated in the RE-PAIR study, where they reported their affect states in four ecological momentary assessments per day for 14 days. We performed two studies. First, we used multilevel vector-autoregressive network models to estimate affect dynamics across all families, and for each family individually. Resulting models elucidated how family affect states were related at the same moment, and over time. We identified relations from parents to adolescents and vice versa, for both positive and negative affect, such as the reciprocal relation of irritation between adolescents and their parents, while considering family variation in these relations. Second, we evaluated the statistical performance of the network model via a simulation study, varying the percentage missing data, the number of families, and the number of time points. We conclude with substantive and statistical recommendations for future research on family affect dynamics.

show abstract

Multimodality and Skewness in Emotion Time Series

Cited by 10 publications

References 43 publications

Delineating Discrete Generalizable States from Intraindividual Time Series: Towards a Science of Moments

Delineating Discrete Generalizable States from Intraindividual Time Series: Towards a Science of Moments

Skewness and staging: Does the floor effect induce bias in multilevel AR(1) models?

A Network Study of Family Affect Systems in Daily Life

Contact Info

Product

Resources

About