Corrigendum to “Adult attachment and prolonged grief: A systematic review and meta-analysis” [Personality and Individual Differences 214 (2023) 112315] (Personality and Individual Differences (2023) 214, (S0191886923002386), (10.1016/j.paid.2023.112315))

Correction to Eisma et al. (2023) In the article “Adult attachment and prolonged grief: A systematic review and meta-analysis” by Maarten C. Eisma, Kathrin Bernemann, Lena Aehlig, Antje Janshen, and Bettina K. Doering (Personality and Individual Differences, 2023, Vol. 214: 112315. https://doi.org/10.1016/j.paid.2023.112315), an error occurred in the meta-analysis of the concurrent associations between attachment avoidance and prolonged grief symptoms. The error has not affected the general conclusions from the systematic review and meta-analysis. Nevertheless, the authors regret the error has occurred and hereby wish to correct it. The correlation coefficient between attachment avoidance and prolonged grief symptoms for the study by Delespaux (2013) was entered as r = 0.19 in the calculation of the meta-analysis. The correct coefficient is r = −0.21. The correction of this error resulted in changes in Section 3.3.4 of the results section, Figs. 3, 4 and Table 4. Meta-analysis of concurrent associations between attachment avoidance and prolonged grief symptoms The random effects model demonstrated a pooled correlation coefficient of small size between attachment avoidance and prolonged grief symptoms (r = 0.13; k = 15; 95 % CI: 0.01–0.24, Fig. 3). The LFK index was −1.00, with the Doi plot indicating no asymmetry (i.e., publication bias). Heterogeneity among studies was significant (Q = 206.88, p ≤ .001) with a very high I² (93%). The subgroup analysis comparing studies that measured attachment avoidance in general (r = 0.23; k = 10; 95 % CI: 0.19–0.27; I² = 30 %) and studies that measured attachment avoidance in relation to the deceased (r = −0.07; k = 5; 95 % CI: −0.31−0.19; I² = 94 %) yielded a significant difference between the groups, Q(1) = 122.76, p ≤ .001 (Fig. 4). The sensitivity analysis identified three outliers as outside of the 95 % CI of the pooled correlation coefficient (Delespaux et al., 2013; Meert et al., 2010; Smigelsky et al., 2020). Table 4 presents the results of the MetaXL sensitivity analysis by showing the pooled correlation coefficient when each respective study is excluded. Removal of the three outliers increased the pooled correlation coefficient (r = 0.20; k = 12; 95 % CI: 0.15–0.24) and reduced heterogeneity (Q = 16.65, p = .119; I² = 34 %). After removal of the outliers, the difference in the subgroup analysis comparing attachment assessments (attachment avoidance in general: r = 0.22; k = 9; 95 % CI: 0.18–0.25; I² = 0 %; attachment avoidance in relation to the deceased: r = 0.11; k = 3; 95 % CI: −0.07-0.30; I² = 72 %) was no longer significant (Q = 3.37, p = .066).[Figure