Common Errors in Using Statistical Tools and Data Presentation

Abstract

AbstractErrors in using the statistical tool can occur at multiple stages of the study. It can occur in sample size calculation, defining the parameter, assigning appropriate statistical tests for them, checking all the assumptions of a particular statistical tool, and finally in proper tabulation and graphical representation of data. Manipulation of alpha and beta values to achieve low sample size, extrapolating conclusions for an objective from the sample size estimated for different objectives, and missing drop out cases are common errors in sample size calculation. In data handling, unnecessary conversion of continuous to categorical data, setting improper cutoff value while categorization, and unnecessary log transformation can lead to loss of data validity. Similarly, choosing the wrong statistical test, not checking the ‘normality distribution’ of data before analysis, and not meeting the test assumptions are often encountered while doing statistical analysis. This chapter summarizes common statistical errors encountered in various biomedical researches and their implications and steps to avoid them.KeywordsSample sizeData representationStatistical testStatistical analysisNormality checkingData transformationError