| dc.description.abstract | In 2005, Ioannidis revitalised a long-lasting debate concerning whether or not null hypothesis significance testing may serve to uphold accessibility related biases, and in effect the file drawer problem as well. Though several authors have argued that significance testing may consequentially impede further progress within the medical and behavioural sciences, it has had little impact on the general consensus, and significance testing is still the most common approach to analyze and interpret data within psychology. In this thesis it is argued that as the philosophical foundation of the argument has never been extensively analysed, reluctance to accept the proposition is understandable, even justified, until sufficient support may decide its validity. In response to these unresolved philosophical and methodological issues, this thesis argues that continuous problem-solving is the driving force of scientific progress, and that problem-solving is contingent upon pluralism. To ensure pluralism, a representative sample of all experimental findings must necessarily be publicly available. However, several well-documented accessibility related biases obstruct theoretical and empirical diversity, and may accordingly inhibit further progress within psychology. In support of Ioannidis’ claim, it is argued that null hypothesis significance testing is the underlying cause of the current accessibility related biases. In addition, this methodological dogma aids several conjunctive fallacies, such as p-value misinterpretations, cultivation of dichotomous thinking, and the problem of practical significance. Taken as a whole, these premises thoroughly emphasize perhaps the most serious challenge towards further progress within psychology. Lastly, some
methodological substitutes are discussed which might liberate the discipline from the current accessibility related biases, encourage further progress, and promote newfound credibility in psychology. | nb_NO |
