Measurement times
Study 1 was part of a broader longitudinal study with different objectives and a variety of measurement times and variables. Data collection took place between 2016 and 2018. For reasons of readability, only the measurement points and variables relevant for this study are presented here. A complete list of all measurement points, variables, the exact order of the questionnaires, and data can be found at OSF. Measures were obtained in the first two trimesters of pregnancy (t1: e.g., relationship questionnaires), one time within the first week after birth (t2: e.g., labor and birth outcomes), and using EMA within the first six weeks after birth (postpartum adjustment). Note, most participants completed questionnaires in the first half of pregnancy. Gestational age at the first measurement time depended on how early in the pregnancy we could engage the women to participate, and varied between week five (0.7%) and week 25 (0.3%) week. The mode was week 15 (11.8%) and seven participants (2%) started between week 21 and week 25, or at the beginning of the second half of pregnancy.
Study 2 used a retrospective study design, surveying participants who gave birth during the first COVID-19 lockdown in Germany between March 15 and April 8, 2020. Data collection took place from June 29 to August 24, 2020.
For both studies, data collection took place online. For Study 1, t1 and t2 questionnaires had to be completed on the computer due to technical peculiarities. The EMA was performed via mobile phone. For Study 2, participants could choose whether they wanted to answer the questions on a computer or on a mobile phone.
Participants
Both studies presented in this paper were conducted in Germany, using a German sample. In the first study partnered heterosexual dyads participated. In the second study, only the childbearing parent took part, and participants with and without partners were welcome to participate.
Study 1. For the longitudinal study, 304 cisgender, heterosexual, partnered dyads were used. Minor fluctuations in the sample size occurred depending on the time of measurement (t1: n = 304; t2 for females: n = 293, t2 for males: n = 279; EMA for females: n = 293, average response rate: 74%, EMA for males: n = 292, average response rate: 68%). At t1 the mean age for the 304 female participants was 30.30 years (SD = 3.99) and for male participants was 32.58 years (SD = 4.51). The majority of couples were married (62.2%) at the beginning of data collection (t1). Only 3.6% indicated a length of relationship of less than a year, and the length of relationship for the remaining sample varied between one year (3.9%) and 20 (1.0%) years with a mode of six years (9.2%; the length of relationship was also assessed at t1). The vast majority of fathers were present at the birth (97.2%). Prior to data collection women completed a screening questionnaire to assess exclusion criteria. Participants being pregnant with more than one child, artificial insemination, and more than one abortion and/or more than one stillbirth in the past could not take part in the study to avoid unnecessary burden for them. Participants also had to have mobile internet access and be older than 18 years and younger than 38 years. The use of psychotropic drugs was also an exclusion criterion. Women were recruited either by the help of midwives and gynecologists or via Facebook groups and Facebook advertisement. Participating women received 100 euros and participating men 80 euros (as they had to complete fewer questionnaires) as monetary compensation. Incentives were paid after the last measurement time point eight weeks after the birth. Participants were also paid if a measurement point was missed. For the additional assessment point six months after birth no incentive was paid.
Study 2. The second study was completed online by 1,160 participants. Participants were recruited mainly through social networks such as Facebook and Instagram. The survey was accessible from June 29 to August 24, 2020. For 180 of the 1,160 participants, preregistered exclusion criteria were met: they did not give birth between March 15 and April 8, 2020 (the period of the official lockdown due to the COVID-19 pandemic in Germany), gave birth in out-of-hospital settings, were men, and/or indicated their data should not be used. Thus, 980 participants (979 female, 1 third gender; Mage = 31.90 years, Sdage = 4.15 years) remained in the sample. Our target sample size of 250 was clearly exceeded but in favor of power maximization we retained the sample that remained after exclusion based on the defined criteria above.
Measures used in study 1
Measurement time 1 (t1). The measures described immediately below were collected at t1, i.e., between week 5 and 25 of pregnancy.
Relationship attachment. We used the partner-specific [
26] German version of the Relationship Questionnaire (RQ; [
27]) to assess relationship attachment. The scale contains one prototypical description for each of the four attachment styles (secure, anxious, preoccupied, dismissive), and participants responded to each description by rating to what degree it describes themselves on a six-point Likert scale ranging from 1 =
strongly disagree to 6 =
strongly agree (the original answer format was changed for the present study). The responses to the insecure attachment descriptions were recoded and aggregated with secure attachment such that a high score of relationship attachment indicates secure attachment. Cronbach’s α was 0.60 for female participants and 0.59 for male participants.
Attitudes toward romantic partner. For measuring (explicit) attitude toward the romantic partner, we used the scale developed by Banse and Kowalick [
6]. Participants were asked to answer 15 items about their partner (e.g.,
I feel good when I am close to my partner) on a six-point Likert scale ranging from 1 =
strongly disagree to 6 =
strongly agree (the original answer format was changed for the present study). Cronbach’s α was 0.83 for both female and male participants.
Relationship satisfaction. The German version [
28] of the Relationship Assessment Scale (RAS; [
29]) was used to assess relationship satisfaction. The scale consists of seven items (e.g.,
In general, how satisfied are you with your relationship?) that participants answered on a six-point Likert scale. The scale’s endpoint labels depended on the particular question. Cronbach’s α was 0.87 for female and 0.82 for male participants.
Dyadic coping. Dyadic coping was measured with the first two subscales of the Dyadic Coping Inventory (DCI; [
30]). The first subscale comprises four items about the desired involvement of the partner when feeling stressed or burdened (e.g.,
I ask my partner to take over tasks and activities if I am overloaded). The second subscale contains 11 items assessing the partner’s reaction to the expressed stress (e.g.,
She/he gives me the feeling that she/he understands me and that she/he is interested in my stress). The subscales were combined to a single score of dyadic coping, as is also suggested in the test manual [
30] and our study did not aim to obtain precise information about certain coping difficulties. All items were answered on a six-point Likert scale (1 =
strongly disagree and 6 =
strongly agree). Cronbach’s α was 0.87 for female and 0.88 for male participants.
Birth-related mindset. Birth-related mindset was assessed using the Mindset and Birth Questionnaire (MBQ; [
31]), which consists of 18 items and four subscales (for the present study an overall score was used). The scale measures trust in midwives versus doctors, birth-related shame and disgust sensitivity, the participant’s view of drug support and vaginal birth. The answer format is a six-point Likert scale ranging from 1 =
strongly disagree to 6 =
strongly agree. Cronbach’s α for this questionnaire in Study 1 was 0.89.
Measurement time 2 (t2). Low-intervention birth and birth experience were assessed shortly (within the first week) after birth at t2.
Low-intervention birth. To summarize the complex process of labor and birth we combined (effect-coded) different birth variables into one variable indicating whether participants had a low-intervention birth (= 1) or not (= -1). The variable low-intervention birth [
32] is adapted from the normal birth index [
33]. If labor and birth was induced (19.9%), or augmented during the process (33.3%), an epidural (24.7%), or episiotomy (13.9%) was performed, and/or the birth was ended by vacuum or forceps (9.4%) or C-section (17.0%), the birth was counted as a high-intervention birth. The numbers in parentheses refer to the frequencies of the specific interventions in the present sample. If none of the mentioned interventions were performed the birth counted as a low-intervention birth, which was the case for 39.7% of the participating women. The C-section rate was lower than would be expected for the years 2016–2018, at 17.0% compared to 29.1–30.5% in the German population, but the number of vaginal assisted births was minimally higher (about 6% on population level compared to 9.4% in the present study; [
34‐
36]; and no population-level prevalence is available for the other interventions. Note, potential problems with using this binary index of low intervention are addressed in the discussion.
Birth experience. To assess the participants’ general satisfaction with the birth experience (e.g.,
I would wish for another birth like this.), we used the Birth experience scale [
31]. The scale consists of 10 items answered on a six-point Likert scale ranging from 1 =
strongly disagree to 6 =
strongly agree. Cronbach’s α in the present study was 0.90 for women and 0.85 for men.
Postpartum adjustment. For assessing postpartum adjustment within the first six weeks after birth we used EMA [
25], which yielded repeated-measures data about current emotional states or behavior in participant’s natural environments [
37]. In the EMA process, participants received a link to an online questionnaire on their mobile phones at a random time of a day (time-based sampling; [
37]). Links were sent daily in weeks one and two, and weekly in weeks three through six, all between 9am and 8pm. The questionnaire contained questions about participant’s emotional and general well-being, and the perceived infant’s well-being (see below). For determining Cronbach’s α, all measurement times were divided into split halves (odd-even).
For measuring emotional well-being, we used 12 items from the Quality of Life Profile for Chronically Ill Patients [
38]; women: α = 0.93, men: α = 0.96), and for general well-being two (men) and three (women) items, respectively, measuring how pain-free (answered only by women), healthy/fit and resilient participants felt (women: α = 0.95, men: α = 0.95). Both scales were answered on a six-point Likert scale ranging from 1 =
strongly disagree to 6 =
strongly agree. The baby’s perceived well-being was assessed with six items using a semantic differential (six points). Items measured the crying and sleeping behavior of the baby, how satisfied, exhausted, and quiet the baby appeared, and how easy the baby could be comforted (women: α = 0.88, men: α = 0.86).
Being primiparous and medical risk. Analogous to previous studies (e.g., [
31,
32]), giving birth for the first time (primiparous: 54.9% in the present study) and having an identified medical risk were treated as control variables. The questions concerning participants’ prenatal risks were based on the German maternity guidelines [
39]. They assessed e.g., previous C-sections, fetal malposition, and health status of the mother. The questions were answered with yes or no according to the presence or absence of the risk factor. If one or more risks were present (true for 45.3% of participants in Study 1), the variable risk was coded.
Measures and procedure used in study 2
As in the first study, the variables relationship attachment (α = 0.62), attitudes toward the romantic partner (α = 0.88), relationship satisfaction (α = 0.88), birth-related mindset (α = 0.83), being primiparous (51.1%), medical risk (55.6%), low-intervention birth (30.9%; induction: 26.3%; augmentation during labor and birth: 25.6%; epidural: 34.2%; episiotomy: 14.5%; assisted vaginal delivery: 8.3%; C-section: 25.7%), and birth experience (α = 0.93) were assessed. For medical risk, in this study we also asked for a SARS-CoV-2 diagnosis at the time of delivery, and four participants had tested positive. Note again the C-section rate in our study was lower than for the German population, which was 29.5% in 2020, and the rate of vaginal deliveries in this study (8.3%) was slightly higher than typical for the German population, which was 6.4%% in 2020 [
40].
Partner’s presence. Participants were asked to indicate who was planned to accompany them during labor and birth (options: male romantic partner, female romantic partner, other), whether the partner was able to be present, and if so, to what extent (the entire time or temporarily, i.e., during active delivery stage). The majority of the sample (95.1%) had planned for their male partner to accompany them, 1.0% for their female partner, and 3.9% for someone else.
1 In 500 cases (51.0%), the companion was continuously present, in 407 (41.5%) temporarily, and in 73 cases (7.4%) not at all. Based on the participants’ answers, we generated the two dummy coded variables,
presence (1 = partner continuously present at birth or partner temporarily present at birth and 0 = partner not present) and
continuous presence (1 = partner continuously present at birth and 0 = partner temporarily present at birth or partner not present) to conduct further analyses. This procedure makes it possible to test the effect of all possibilities (continuously present, temporarily present, not present at all) on low-intervention birth and on birth experience.
Emotional well-being before birth. To retrospectively assess the participants’ emotional well-being in the days before birth, we again used 10 items of the Quality of Life Profile for Chronically Ill Patients [
38], using the following instruction:
We would like you to recollect the last few days before giving birth. How were you feeling when thinking of the imminent birth? The answer format for the items was again a six-point Likert scale ranging from 1 =
strongly disagree to 6 =
strongly agree. Cronbach’s α in the present study was 0.91.
Perceived support. In Germany it is possible to hire a freelancing midwife for labor and birth. The advantage over the standard care by a hospital midwife is continuous one-to-one care. For the accompaniment, however, extra costs must be paid and the number of freelancing midwives providing this type of care is limited. To assess one-on-one support, we asked what kind of midwifery support participants were given during labor and birth with the following options: freelancing midwife, hospital midwife, and no midwife. For the analyses the variable was dummy coded. In the present study 12.7% of the participants had access to one-on-one support from a midwife.
We also assessed the participants’ perception of different support forms originating from different support sources. The different support forms were based on Hodnett [
41] and included emotional support, comfort measures, information, and advocacy. Participants were asked to rate the perceived support on the four items for freelancing midwives (if present), hospital midwives, other hospital staff, and the partner/other accompanying person (if at least temporarily present) on a six-point Likert scale ranging from 1 =
not at all to 6 =
very much. This allowed us to compare the participants’ perception of their partners’ support on different levels to their perception of the medical staff’s support.
Procedure. The second study was conducted online. Participants were first informed that their participation was voluntary, anonymous, and could be ended at any time. Then they were presented with questions concerning their age and gender. To minimize a potential influence of the birth experience on the birth-related mindset, participants first completed the MBQ before continuing with the birth-related questions, including those concerning the different prenatal risks, the planned companion, and their companion’s presence. They then completed the Quality of Life Profile and the Birth experience scale, followed by the questions concerning participants’ perception of the different supporters and the different support forms. The relationship questionnaires were answered last, and only by those who had at least planned for their partner to accompany them. In the end, all participants could indicate whether their data should be used or not, leave a personal remark, and share their e-mail address if they wanted us to send them a summary of the results. In this case, the address was registered separately.