KLM form analyzer: Automated evaluation of web form filling tasks using human performance models

Abstract

Filling forms is a common and frequent task in web interaction. Therefore, designing web forms that enhance users' efficiency is an important task. This paper presents a tool entitled KLM Form Analyzer (KLM-FA) that enables effortless predictions of execution times of web form filling tasks. To this end, the tool employs established models of human performance, namely the Keystroke Level Model and optionally the Fitts' law. KLM-FA can support various evaluation scenarios, both in a formative and summative context, and according to different interaction strategies or modeled users' characteristics. A study investigated the accuracy of KLM-FA predictions by comparing them to participants' execution times for six form filling tasks in popular social networking websites. The tool produced highly accurate predictions (89.1% agreement with user data) in an efficient manner. © 2013 IFIP International Federation for Information Processing.