Article title: Vasopressin and parental expressed emotion in the transition to fatherhood
Authors: Anna M. Lotz, Jolien Rijlaarsdam, Jurriaan Witteman, Willemijn Meijer, Kim van Dijk, Marinus H. van Ijzendoorn
Journal: Attachment & Human Development
Bibliometrics: Volume 23, Number 3, pages 257-273
DOI: https://doi.org/10.1080/14616734.2020.1719427
After publication of the manuscript “Vasopressin and Parental Expressed Emotion in the Transition to Fatherhood”, we discovered some inaccuracies in the reported analyses. The corrections do not change the conclusion of the study.
In the original paper, the description of the conditions that were compared in the statistical analyses was not entirely correct. P8: “The effect of AVP administration on emotional prosody was analyzed using multivariate repeated measures ANOVA with AVP as within-subjects factor and the acoustic parameters as dependent variables. In the final step, we explored the effect of the transition to fatherhood on both FMSS-based expressed emotion and emotional content. Similar analyses as described in step three were performed, only for the multivariate repeated measures ANOVA prenatal and postnatal measures were chosen as within-subjects factor.”
The correct description is ‘The effect of AVP administration on emotional prosody was analyzed using multivariate repeated measures ANOVA with condition (PL or AVP) as within-subjects factor and the acoustic parameters as dependent variables. In the final step, we explored the effects of the transition to fatherhood on both FMSS-based expressed emotion and emotional content. Effects on positive comments, critical comments, and emotional content were examined using Wilcoxon matched-pairs tests comparing data from the prenatal PL condition and the postnatal assessment. A multivariate repeated measures ANOVA was performed to test the effect on emotional prosody. To this end, the first prenatal assessment and the postnatal assessment were compared as an effect of time (first vs second assessment), since no effect of AVP administration was found.
In the context of the repeated measures ANOVA testing the effect of AVP administration on emotional prosody, we compared the measurements of prenatal assessment 1 versus prenatal assessment 2, instead of PL vs AVP (P.9, F[6,18] = 0.64, p = .70, ƞ2 = .18), which is slightly different. We now ran the correct analysis to test the effect of AVP administration on emotional prosody. The correct test statistics are: F[6,19] = 0.68, p = .67, ƞ2 = .18
Moreover, one-tailed p-values instead of two-tailed p-values were reported for two Wilcoxon matched-pairs tests (P.9, the effect hormone administration on subjectivity, Z = −0.06, p = .48, d = 0.01; P.9, the influence of birth on polarity (Z = -0.20,p = .43,d = 0.07).
The correct two-tailed p-values are: Subjectivity (Z = -0.06,p = .97,d = 0.01); Polarity (Z = -0.20,p = .86, d = 0.07)
Lastly, we noted some typos in the Results section and corresponding tables.
P9 & Table 4: Influence of birth on subjectivity (Z = -0.04). Correct statistic: Z = -0.40
P9 & Table 4: Influence of birth on emotional prosody (F[6,14] = 4.12). Correct statistic: F[6,14] = 4.62
P9: reported M and SD of variation in intensity prebirth (M = 9.53, SD = 1.36). Correct statistics: M = 9.63, SD = 1.46
P9: reported average and standard deviation for speak rate pre-birth (PL: M = 2.31,SD = 0.66). Correct statistics: pre M = 2.21,SD = 0.67
Table 1, r statistic for spearman correlation Critical comments – Negative affect PL, r = .33. Correct statistic: r = -.33