Fisher-Yates Normalization for Questionnaire Data

• Activity Number: 350
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Birol Emir*+ and Javier Cabrera
• Companies: Pfizer and Rutgers University
• Keywords: Fisher-Yates ; , Quantile Normalization ; Questionnaire

Abstract

Questionnaire data is a research tool that is utilized by researchers in order to obtain information about certain characteristics of individuals. One of the issues with such scales is that they depend on the individual perception of the scores, some subjects will answer with generally high values, some will answer with low or middle, whereas others will answer in different patterns. We notice that (i) the distribution of answers could differ substantially from subject to subject in level, spread and shape, (ii) the boundary threshold effects at the boundary values make the distribution of the scores either left or right skewed. There are statistical analysis challenges in using these questionnaires data when conducting standard analyses such as modeling or clustering and the main issue is to preprocess the subject data rather than the variable with the objective of defining a feature space or to construct a response score. We show that the traditional transformation methods (centering, scaling, z-scores) are inadequate to fulfill this task and we propose an application of FY scoring that we call FY normalization. We show examples of pain data collected from clinical trials