Perceived neighborhood crime and social cohesion trajectories and mental health during COVID-19: Evidence from a 2020-2024 cohort.

We examined longitudinal changes in perceived neighborhood crime and social cohesion from 2020 to 2024 and their associations with depression, anxiety, and loneliness. U.S. adults (N = 895) recruited in spring 2020 completed at least one annual follow-up surveys through 2024. Group-based trajectory modeling identified four perceived-crime trajectories: low (71.7%), increasing (14.9%), decreasing (8.7%), and high (4.8%), and three stable social-cohesion trajectories: high (18.2%), middle (63.9%), and low (17.9%). Multivariable logistic regression showed that, compared with the low-crime trajectory, high and increasing crime were associated with higher odds of depression (OR = 2.61 and 1.82) and anxiety (OR = 3.08 and 1.87), whereas decreasing crime was associated with lower odds of loneliness (OR = 0.55). Compared with high cohesion, low cohesion was associated with higher odds of depression (OR = 6.38), anxiety (OR = 4.81), and loneliness (OR = 4.17). Complementary within-between mixed-effects models showed that within-person increases in perceived crime were associated with higher depression (β = .12), anxiety (β = .15), and loneliness (β = .19), whereas within-person increases in social cohesion were associated with lower symptoms (β = -0.02 to -0.04). Findings indicate consistent longitudinal associations between neighborhood perceptions and mental health across person-centered and variable-centered approaches.