Magnitude-based inference and its application in user research. Issue 88 (April 2016)

• Record Type: Journal Article
• Authors: van Schaik, Paul; Weston, Matthew
• Is Part Of: International journal of human-computer studies. Issue 88(2016)
• Page Start: 38
• Page End: 50
• Publication Date: 2016-04
• Subjects: User research -- Quantification -- Statistics -- Inference -- Usability testing -- User-experience; Human-machine systems -- Periodicals; Systems engineering -- Periodicals; Human engineering -- Periodicals; Human engineering; Human-machine systems; Systems engineering; Periodicals; Electronic journals; 004.019
• DOI: 10.1016/j.ijhcs.2016.01.002
• Languages: English
• ISSNs: 1071-5819

Abstract

Abstract: Magnitude-based inference offers a theoretically justified and practically useful approach in any behavioural research that involves statistical inference. This approach supports two important types of inference: mechanistic inference and practical inference to support real-world decision-making. Therefore, this approach is especially suitable for user research. We present basic elements of magnitude-based inference and examples of its application in user research as well as its merits. Finally, we discuss other approaches to statistical inference and limitations of magnitude-based inference, and give recommendations on how to use this type of inference in user research. Highlights: Magnitude-based inference is a useful alternative for analysing user-research data. Goal-setting in user research is supported by choosing a smallest important effect. The approach uses the smallest important effect in making an inference. As a consequence, a clear effect is never an artefact of sample size. Practical inference is supported by weighing harm and benefit appropriately.