A method for aggregating ordinal process assessment measures. Issue 12 (29th September 2014)

• Record Type: Journal Article
• Authors: McBride, Tom; Varkoi, Timo
• Is Part Of: Journal of software. Volume 26:Issue 12(2014:Dec.)
• Page Start: 1267
• Page End: 1279
• Publication Date: 2014-09-29
• Subjects: Software engineering -- Periodicals; Computer software -- Development -- Periodicals; Software maintenance -- Periodicals; 005.1
• DOI: 10.1002/smr.1676
• Languages: English
• ISSNs: 2047-7473

Abstract

<abstract abstract-type="main"> <title>ABSTRACT</title> <p>Process capability and maturity models invariably use an ordinal scale of capability or maturity. Process assessment methods must include ways to aggregate the base assessment measures through to the capability level. As process reference models and assessment models have become more detailed, process assessments have started to assess the processes at lower levels of process attributes. Also, process assessments can assess different divisions of an organization then try to combine the results to form some conclusion about the overall organizational capability. Both circumstances have exposed the problems of aggregating ordinal measures. We propose that these problems can be overcome by converting ordinal base measures to probability distributions. Doing so allows mathematical aggregation to be used at all levels of the measurement model. Although this overcomes a problem, we caution that it also provides an illusion of precision and accuracy where none is justifiable, leading to a tendency to overlook the original purpose of process assessment, which is to identify potential process improvements. Copyright © 2014 John Wiley &amp; Sons, Ltd.</p> </abstract>