Statistics commonly used in research reports: issues and challenges

Traditional descriptive measures like mean and variance together with the ubiquitous p-value in hypothesis testing are still widely analyzed and reported by researchers. Unfortunately literature has shown that researchers fail to exploit the full potential of statistics to unlock more meaningful information from their data. This study is aimed at looking into this issue in a Malaysian context and determining the type of research designs and statistics used in reporting findings based on an analysis of over 220 papers published in a conference proceeding organized by a local university. Frequency count, tabulations and graphical representations were used to analyse the data. Findings indicated that majority of the papers employed quantitative design with descriptive statistics as their major reporting technique. Among the 128 papers using inferential statistics, 35% of them used univariate tests while only 12% resorted to multivariate tests. Two faculties i.e. Social Sciences & Humanities and Science & Technology are major contributors to the high percentage of usage of statistical-related reports. Furthermore, only 7 papers reported their findings using confidence interval (CI) while 23 papers used effect size (ES) and none reported power of tests. This paper concludes with a short discussion on pertinent issues raised and provided some suggestions with regards to additional reporting measures other than p-value