University of Twente Student Theses

Login

Fractional Bayes factor to assess the significance of Pearson correlation coefficient

Menke, Janosch (2017) Fractional Bayes factor to assess the significance of Pearson correlation coefficient.

[img]
Preview
PDF
37MB
Abstract:The Pearson’s correlation is on of the most used statistical tests in social sciences. The null hypothesis significance testing (NHST) is the conventional test to establish whether the observed effect, the correlation, is significant. The NHST uses p-values to calculate how likely it is that the observed effect is due to variance in the sample or is actually present in the population. This approach has some disadvantages. It is impossible to quantify evidence in favor of the null hypothesis. Another issue is that increasing the sample size, also increases the significance, thus the sampling procedure can have an impact on the significance. To overcome the issues of the NHST, other approaches were developed using a Bayes factor. The Bayes factor is a ratio of the probability of the null hypothesis being true divided by the probability of the alternative being true. A new approach using an uninformative prior was tested in this simulation study and compared with the NHST and another Bayes factor test. The power of the three different tests was compared with each other. Both Bayes factor approaches appeared to be better suited in judging significance. However, the newly developed approach had some shortcomings in comparison with the preexisting Bayes factor test
Item Type:Essay (Bachelor)
Clients:
Jahr, Deutschland
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:31 mathematics
Programme:Psychology BSc (56604)
Link to this item:http://purl.utwente.nl/essays/72649
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page