(PS2-95) Forecasting hopelessness: An investigation of cognitive bias using intensive longitudinal data

Abstract Text:

Background: Affective forecasting is the process by which individuals predict their future affective states. Prior research indicates that people tend to overweight their current affective state when forecasting future affect. While optimistic forecasts may be associated with mental health, pessimistic predictions may worsen negative affect, potentially leading to hopelessness or thoughts of suicide. This study examines the accuracy of affective forecasts by assessing moment-to-moment fluctuations in hopelessness and negative affect (NA) through intensive longitudinal data.

Methods: We collected five ecological momentary assessments (EMA) per day over ten days from 180 participants (N=8,491). Using dynamic structural equation modelling (DSEM), we evaluated how momentary hopelessness and NA were interrelated. NA was modelled as a latent construct indicated by momentary reports of anger, stress, sadness, worry, and anxiety. Autoregressive coefficients (ϕ) indicating inertia–or stability–of a construct over time and cross-lags (ϕ) indicating spillover effects between hopelessness and NA were calculated. We also computed within-person log-variances (Ψ) to assess variability over time.

Results: All NA items loaded onto the latent variable, with the highest loadings for stress (γ*STR,B=0.99 [0.97, 1.0] and worry γ*STR,B=0.94 [0.91, 0.96]. At the within-subject level, although all autoregressive and cross-lagged relationships were significant, NA exhibited the strongest inertia effect (ϕ*NN=0.54 [0.48, 0.60]). By comparison, hopelessness had lower inertia (ϕ*HH=0.17 [0.12, 0.22]) and more modest spillover effects on NA (ϕHN=0.21 [0.16, 0.25]) and vice versa (ϕ*NH=0.09 [0.04, 0.13]). Within-subject log-variances for NA (Ψ*NA=0.53 [0.49, 0.57]) and hopelessness (Ψ*HPL=0.71 [0.67, 0.74]) suggested significant moment-to-moment variability. At the between-subject level, overall NA and hopelessness were positively associated with their respective variabilities suggesting higher average levels correspond with greater fluctuations.

Conclusion: Our findings suggest that while pessimistic affective forecasts (i.e., hopelessness) do predict future NA, momentary NA is a far stronger predictor of subsequent emotional states. These results underscore the importance of cognitive processes in emotional forecasting, which may have critical implications for understanding emotion regulation and developing interventions in clinical settings. Future research could further explore these dynamics and potential moderating factors.