Abstract
This study systematically examines the relationship between the two core dimensions of adult attachment — attachment anxiety and attachment avoidance — and various manifestations of anxiety (social anxiety, trait or general anxiety, existential anxiety, generalized anxiety, and post-traumatic stress symptoms), following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (the PRISMA 2020 reporting guideline). Attachment theory holds that the internal working models formed in early caregiving relationships continue to shape emotion regulation and interpersonal functioning across the lifespan, and a substantial body of research links insecure attachment to elevated anxiety. Large-scale meta-analyses addressing this relationship have, however, focused predominantly on social anxiety, whereas comprehensive syntheses that jointly encompass different types of anxiety and that also incorporate primary studies conducted in Türkiye have remained limited.
To help address this gap, studies published between 2015 and 2025 that measured attachment with a valid dimensional scale and reported a direct correlation coefficient together with the sample size were systematically screened, and those with data amenable to quantitative synthesis were included. Data from seven independent studies (combined sample of 3,464 participants) were synthesized using the Fisher z-transformation and a random-effects model. Attachment anxiety was, overall, more strongly associated with anxiety (pooled correlation of .338; 95% confidence interval [.228, .440]) than was attachment avoidance (pooled correlation of .279; 95% confidence interval [.195, .360]); this pattern is consistent with attachment-based models of emotion regulation and with recent meta-analytic syntheses. The heterogeneity across studies was very high (I² = 85% to 92%), indicating that the findings should be read as a conditional estimate that is sensitive to the specific type of anxiety measured rather than as a single universal coefficient. Because the number of studies is small (seven), the results are presented as an expanding pilot synthesis rather than as a definitive meta-analytic conclusion. The study contributes a deliberately transparent, no-imputation synthesis and brings under-represented Turkish primary studies into meta-analytic view.