Study design

Sites

Samples

Assessments

Data treatment

Ethical considerations

Sites

The six projects of the CENOF Research Study will be carried out at four different sites, namely, Zurich/CH (Project I), Fribourg/CH (Project II), Dortmund/D (Project IV) and Vienna/A (Project III, V and VI). Each site will be using the infrastructure and resources at the respective site. This is most important since the planned projects demand intense quantitative and qualitative research strategies which are time-consuming, and call for exploitation of all possible resources including students’ and interns’ assistance, or students who are carrying out research for bachelor and master theses.

Samples

The CENOF Research Study will stand out due to its overall sample size of N=3.700 participants, designating a large-scale study to ensure generalizability and repre­sentativeness of the expected outcomes. Large-scale studies have been very seldom among human science studies in Europe over the last decades. Although the individual projects are concerned with samples of special features, they will all relate to normal middle class samples in order to draw generalized conclusions. Surely, all subsamples will represent equal sizes of girls and boys making it possible to systematically explore gender differences in parental investment throughout the entire study (with Project VI ground-breaking).

Assessments

The CENOF Research Study will excel due to its carefully conceptualized design and overall structure of which cross-linking of all areas of expertise of the CENOF members are involved, and will thus result in synergy effects. Study design and research questions suggest intense exploration with multi-level assessments, for which each family will be visited several times. Assessments range from interviews and observations over videotaping including computer-based observation equipments (such as Interact by MANGOLD) to psychobiological approaches based on participants’ saliva. Selections of the assessments were decided based on the latest methodological inventions in the fields. Most instruments are internationally recognized by their validation and reliability, and the CENOF members’ laboratories keep long-term experiences on the usage of those measures. However, there will a few assessments which need to be invented or adjusted to the special pur­pose of the CENOF Research Study. For reasons of comparability across sites and research questions, however, the CENOF group is going to share the majority of these assessments.

Data treatment

The data management of the CENOF Research Study is centralized at the CENOF headquarter at Vienna University. Association patterns of both (1) within a singular project, as well as (2) between several projects will be analyzed. (1) Within-data analyses will be provided by multivariate analyses, time table evaluations, and hierarchical linear models as well as structural equation modeling in order to cope with some nested substructure of the data (e.g., Bryant & Yarnold, 1995; Bryk & Raudenbush, 1992; Muthén & Muthén, 2010).

However, one of the distinct challenges in data analysis will be the interconnections of the various data sets - hormonal data as well as paternal time investment variables and the quality of fatherhood (captured by father-child relationship quality) – which are available project-wide. Here, however, we need to additionally test for invariance between the different samples (e.g., Burnham & Anderson, 2004; Millsap, 2011; Widaman & Reise, 1997).

In general, we plan to evaluate associations between the hormonal data, relationship codes and time investment scores by path and structural equations models. Apriori hypotheses will be tested by confirmatory comparisons with the control group. Despite complex data handling, exploratory analyses will describe statistical relations, and confounders will be detected and controlled for. Relevant routine demographic data will be used to estimate the possible influence of selection bias. Considering the missing data rates of comparable international studies as well as experiences from own past research, we will partly use planned missing data designs under the stipulation that this strategy would not harm reliability of the data. 

Ethical considerations

Families receive full information on the goals of the study. They will be asked to sign an agreement letter but also informed that participation is voluntary. In addition, the families receive all information that we have collected. Furthermore, no infant/child will be stressed by the study itself, neither by taking saliva, nor by videotaping or the developmental test and play situations which use enjoyable instruments.

Furthermore, we will explain our data safety concept and how our strategy of anonymization works. We usually code all information when we build up the data pool. In order to avoid mismatches of information over the course of the study, however, we will fill full names and addresses in questionnaires and protocols when working with the families. We will archive all original information discreetly.