Recently, an article was published in Frontiers in Psychology of deflation-corrected estimators of reliability (Article Deflation-Corrected Estimators of Reliability
). It has been reported that in some cases, when the test items are difficult or easy, reliability may be deflated from 0,86 to 0.26 when using traditional estimators alpha, theta, omega or maximal reliability. In the article, I suggest some correction factors. Would yo come with better ideas?