Analyzing a single processed dataset, constructed by the researcher without proper justification for the data processing choices, can be misleading. A multiverse analysis (Steegen, Tuerlinckx, Gelman, & Vanpaemel, 2016; OSF page) increases transparency about the influence of a researcher's choices in data construction on the stability of a finding, by showing the results of all data sets that could reasonably be constructed based on the same raw data.

This app uses the raw data of the study of Durante, Rae, and Griskevicius (2013) to visualize the effect of data processing choices. In two studies, Durante and colleagues investigated how fertility interacted with relationship status with respect to six dependent variables, all related to religious or political attitudes. More information about the dependent variables can be found by clicking on the ?.

For each of the six dependent variables Durante et al. considered, the user can make several data processing choices. Information about the data processing choices can be found by clicking on the ?, and in Table 1 in Steegen et al. (2016). The combination of choices gives rise to a unique data set ready for analysis. The resulting single data set is shown in the Single data set tab, and the results of the analysis in the Single data set analysis tab. In the Multiverse tab, the results of all possible single data set analyses are shown, highlighting how the choices in data construction influence the statistical conclusion.


The single dataset constructed based on the processing choices you made.

First column: The chosen dependent variable.

Second column: Your chosen Next Menstrual Onset variable.

Third column: Your chosen Relationship variable.

Fourth column: Your chosen Fertility variable, based on your choice for the NMO and Fertility variables.

Fifth column: The cycle days.



Graphical summary

The barplot represent the dependency between the chosen dependent variable and the relationship status and fertility status (n= )



Effects

p-value


The histogram shows the p-values for the interaction effect between fertility and relationship status for your chosen dependent variable across all processing choices, barring inconsistent ones.

The red line indicates your chosen significance level. The orange dot indicates the p-value from the single data set analysis based on your data processing choices.

The percentage of p-values under the significance level is % ( out of )