Improving Statistical Practice in HCI

Abstract

In the last chapter we critically examine some influential statistical papers in HCI with respect to some of the common statistical flaws we introduced in the first chapter. We show how misleading use of statistical methods raises some doubt as to the interpretation and validity of previously published findings. The chapter concludes with six recommendations for improving statistical practices which we hope that readers will employ, in conjunction with the specific tech-niques explained in the previous chapters, to enhance their statistical communica-tion.