Tests for consistent measurement of external subjective software quality attributes

摘要: One reason that researchers may wish to demonstrate an external software quality attribute can be measured consistently is so they validate a prediction system for the attribute. However, attempts at validating systems subjective attributes have tended rely on experts indicating values provided by informally agree with experts' intuition about These are undertaken without pre-defined scale which it known consistently. Consequently, valid unbiased estimate of predictive capability cannot given because measurement process not independent system's values. Usually, no justification checking see if measure It seems assumed that: isn't proper or quantified one knows true anyway and estimated. even though classification systems' artefacts' subjective, possible quantify measurements in terms conditional probabilities. then possible, using statistical approach, assess formally whether considered consistent. If consistent, also identify estimates values, system. used In this paper we use Bayesian inference, Markov chain Monte Carlo simulation missing data imputation develop tests consistent ordinal attributes.