Comparing Assessment Techniques  

In just the last few years, we have begun to see research studies comparing usability testing to other techniques for assessing usability. In general, usability testing has found more of the most serious usability problems and fewer of the least serious problems than other methods. Heuristic evaluation--having people evaluate the interface either from their own expertise or from a set of guidelines--has achieved mixed results. Although usability testing seems expensive compared to other methods, it may be less expensive when considered on the basis of 'cost per problem that needs to be fixed.'

Redish, Janice C. 'Ginny'. STC Proceedings (1993). Articles>Usability>Testing>Assessment

Examining the Relationship Between Quality Writing and Quality Reading    

This article introduces the ISTE approach as a way to measure the relationship between the design of a document and the way that users handle it. ISTE is an acronym for Information SubTypes and Effects. The article describes the classification of information into a general typology as well as into a specific one, to illustrate how the approach can be adapted to specific usability questions. ISTE also requires a typology for categorizing possible user reactions. This article first positions ISTE within the broader context of usability testing and then describes its basic approach and some recent applications. Unlike most formal research methods, ISTE does not require an inordinate amount of time and effort for analyzing high-fidelity data. Electronic versions such as the ERR instrument allow for the creation of a complete and detailed profile of the use of a document 10 minutes after task completion. These characteristics make ISTE a viable instrument for gathering usability metrics in industrial settings.

van der Meij, Hans. Technical Communication Online (2000). Articles>Usability>Testing>Assessment

The Sixty-Minute Guide to Evaluating Comparative Test Data

Mary wants to compare the average registration times between the two versions to see if the second version is faster than the first. The method typically used when comparing averages is called a t-test of independent means.

Hughes, Michael A. STC Proceedings (2008). Articles>Usability>Testing>Assessment