Osman Saka

Akdeniz Üniveristesi Tıp Fakültesi, Biyoistatistik Anabilim Dalı

Abstract

Although every step of the research process is equally important, evaluation is the major issue for most of the researchers. Hypothesis testing is generally considered as the single method for evaluation by most of the researchers. Actually this is not true. Several graphs, tables and measures such as arithmetic mean, variance, standard deviation, frequency, rate, percentage, incidence or prevalence may be adequate for the evaluation of descriptive studies. However, in some researches many variables may affect the studied subject. In this situation, descriptive statistics, univariate hypothesis testing methods and basic regression - correlation analysis will be inadequate for such studies. Use of multivariate statistical analysis methods must be preferred for evaluating these studies. Most important point is to evaluate the collected data accurately and reach reliable, consistent and unbiased conclusions.