(PS13-9) Multilevel Factor Structure of Negative and Positive Affect, Burdensomeness and Belongingness in a Suicidal Sample

Introduction

The two-factor structure of positive affect (PA) and negative affect (NA) has been widely supported in cross-sectional studies among clinical populations. Research has linked diminished PA, elevated NA, perceived burdensomeness and thwarted belongingness with momentary suicidal ideation (SI). Yet, most studies utilized the composite score of PA and NA subscales without evaluating the potential latent factors. Our understanding of the underlying factor structures of these constructs over time, on an individual level, remains limited. To understand the variability in affective experiences among individuals with SI, we employed multilevel exploratory factor analysis to explore the dynamic factor structure of PA, NA, burdensomeness and belongingness. 

Methods

Participants (N = 157, 37.0 ± 11.1 years old , 70% female, 80% White) who reported active SI or suicidal behaviors in the past year completed 28 days of ecological momentary assessments on their smartphone. They completed a range of 3-15 daily surveys that assessed PA and NA via the Positive and Negative Affect Schedule (10 PA and 10 NA items), perceived burdensomeness (useless; like a burden), and thwarted belongingness (lonely; like I do not belong). We conducted multilevel exploratory factor analysis on 24 items to identify potential factors. 

Results

The model with three between-person factors and four within-person factors demonstrated a good level of fit (CFI = .93, RMSEA = .04) and appeared conceptually coherent. Across individuals, NA, PA, and belongingness/burdensomeness emerged as distinct factors. Within individuals, NA and belongingness/burdensomeness remained, whereas 2 factors emerged underlying PA (physical and cognitive PA, moderately correlated r = .53). 

Discussion

The identification of distinct between- and within-person factors, especially the division of PA into two separate dimensions on an individual basis, underscores the importance of a nuanced understanding of both the physical and cognitive facets of PA. Our findings reveal that NA consistently emerges as a singular factor at both between and within-person levels, distinct from perceptions of burdensomeness and belongingness. These findings not only reinforce the two-factor model of affect in a between-person context but also delineate the unique standing of burdensomeness/belongingness separate from PA and NA. Further, the presence of multiple within-person factors for PA has implications for study design when within-person processes are of interest.