Μελέτη του παράδοξου Berkson στη βιοστατιστική
A study of Berskon’s paradox in biostatistics
View/
Keywords
Παράδοξο BerskonAbstract
This dissertation aims to study a phenomenon known in the literature as
Berkson’s paradox or bias and is a type of bias due to sampling errors. It
often occurs in cohort studies, when samples should be taken in health units
and results in the study group from the hospital to show a higher rate of
exposure to risk or disease rate compared to the general population due to
overrepresentation. A literature analysis of Berkson’s paradox will be made,
while special emphasis will be given to various studies that dealt with this
phenomenon. Finally, real data showing Berkson bias will be studied with
the help of weighted distributions, whose theory will be presented in detail.