In the last decade, the human characteristic of dishonesty has resurfaced as a popular topic of scientific studies. Despite moral values and cognitive control being affirmed predictors, their...Show moreIn the last decade, the human characteristic of dishonesty has resurfaced as a popular topic of scientific studies. Despite moral values and cognitive control being affirmed predictors, their interplay remains subject to debate. At the same time, the replication crisis led to strong criticism towards the quality of psychological research. Since then, there have been various attempts to improve robustness of discoveries in psychology. One enhanced method to assess robustness of a study is the construction of a multiverse with all viable data processing choices. Speer et al. (2023) proposed evidence that acute stress moderates the individual moral default by inhibiting self-control. In favor of this hypothesis, the authors discovered a significant interaction effect between stress and moral default by conducting a Bayesian logistic regression analysis (LRA). The aim of this bachelor's thesis was to reproduce Speer and colleagues’ (2023) findings and to construct a multiverse analysis. 64 unique pathways stemming from 5 alternative processing choices were analysed in a frequentist LRA. Results provide mixed findings for the original hypothesis. The interaction effect is only present, if the floor and ceiling effects are considered. Alternative theories for the observed phenomena are discussed.Show less
The discussed research entails the replication of a previous study investigating the possible effect of the interhemispheric deficit theory of dyslexia, through researching the redundant bilateral...Show moreThe discussed research entails the replication of a previous study investigating the possible effect of the interhemispheric deficit theory of dyslexia, through researching the redundant bilateral advantage (RBA) for dyslexic and control participants. A multiverse analysis was performed on existing research, to provide more robust results, and replicate findings. Participants were excluded, assumptions checked, and outliers removed, resulting in twelve different datasets. The performed multiverse analysis looked at the significant presence of a RBA in controls, a significant difference in RBA between dyslexic and control participants, and a significant difference in between right visual field (RVF) and left visual field (LVF) conditions, in a VHF paradigm. The results concur with the study replicated, and all datasets showed the similar levels of significance. There was a significant RBA in the control group, yet no significant difference was found with the dyslexic group. A significant difference in RVF and LVF difference was found between the groups. This was caused by lower LVF accuracy scores in the dyslexic group, suggesting the presence of an LVF cost in dyslexia. Handedness showed to have an influence on the RBA scores in participants. Future suggestions are made in regards to improved research designs and improvements in multiverse analyses for more robust and transparent research.Show less
Kinderen hebben verschillende manieren om nieuwe dingen te leren: ze kunnen iets zelf onderzoeken, maar ze kunnen ook de informatie van iemand of iets krijgen, zoals een ouder, docent of boek. Als...Show moreKinderen hebben verschillende manieren om nieuwe dingen te leren: ze kunnen iets zelf onderzoeken, maar ze kunnen ook de informatie van iemand of iets krijgen, zoals een ouder, docent of boek. Als de informatie komt van een ander of bijvoorbeeld een boek vergt dit wel het vertrouwen dat deze informatie klopt. Wanneer een kind in een boek iets leest over een land ver weg hebben ze geen manier om dit te checken. Ronfard et al. (2021) onderzocht hoe kinderen reageren als ze dit wel kunnen checken. Hierbij vonden ze dat kinderen meer geneigd zijn het feit te onderzoeken als het tegen hun intuïtie ingaat dan wanneer het heel logisch klinkt. Daarnaast hebben ze ook gevonden dat oudere kinderen sneller op onderzoek uitgaan. Omdat dit specifieke idee niet eerder is getest, is het doel van deze multiverse om te onderzoeken of kinderen meer op onderzoek uitgaan als ze iets horen dat tegen hun intuïtie ingaat. In deze multiverse is gevonden dat de bevindingen robuust zijn voor controle op opleidingsniveau van de ouders en voor een alternatieve berekening van deze variabele. De bevindingen waren niet robuust voor het controleren op de autoritaire waarden van de ouders, het gebruiken van leeftijd in plaats van leeftijdsgroep in de analyse, de exclusie van een kleine groep participanten waar het onderzoek niet goed liep en verschillende tests om hetzelfde effect te meten. De resultaten van de multiverse suggereren dat kinderen wel meer geneigd zijn om op onderzoek uit te gaan als ze iets horen dat tegen hun intuïtie ingaat, maar dat er geen sprake is van een interactie-effect met leeftijd hiervan.Show less
Conducting empirical experiments for statistical hypothesis testing can be resource-intensive and can impose pressure on subjects. To address these challenges, employing sequential designs allow...Show moreConducting empirical experiments for statistical hypothesis testing can be resource-intensive and can impose pressure on subjects. To address these challenges, employing sequential designs allow for repeated statistical analyses during data collection, potentially enabling researches to draw conclusions based on accumulated information, reduce costs, and minimize subject exposure. This study aimed to assess the efficiency and reliability of the Group Sequential Design and the Sequential Bayes Factor Design by comparing them with a fixed-sample design. The results demonstrated that both sequential designs outperformed the fixed-sample design in terms of efficiency, with the Sequential Bayes Factor Design exhibiting the highest level of efficiency. However, the Group Sequential Design demonstrated greater concordance with the fixed-sample design in terms of qualitative conclusions and effect size estimates. Disparities between the sequential designs in terms efficiency gains and concordance in conclusions with the fixed-sample design were attributed to variations in the frequency of interim analyses and the stringency of critical boundaries. The Group Sequential Design emerged as the more reliable approach due to its advantages in minimizing conditional bias and controlling error rates. Given its superior reliability and greater availability of clear implementation guidelines, the Group Sequential Design is generally preferred over the Sequential Bayes Factor Design. Nevertheless, the Sequential Bayes Factor Design offers the advantage of flexibility and the interpretation of the Bayes factor for assessing the strength of evidence, even without reaching critical boundaries, which is a distinctive feature that is absent in the Group Sequential Design. Recommendations were provided to address limitations in implementing the Sequential Bayes Factor Design. Therefore, further research should focus on refining the methodological decisions and parameter settings of the Sequential Bayes Factor Design to enhance its reliability and practical applicability.Show less
