2017/05/02 by Cristian Gugiu, P. Cristian Gugiu, Mihaiela Gugiu +1 · 18 citations
Decision Sciences · Engineering · Mathematics · #Computer science #Econometrics #Engineering #Function (biology) #Generalizability theory #Machine learning #Margin (machine learning) #Mathematics #Multi-Criteria Decision Making #Power (physics) #Reliability (semiconductor) #Reliability engineering #Set (abstract data type) #Standard error #Statistics
paper · doi:10.1080/00220973.2017.1315712
published in The Journal of Experimental Education 86(3), 458-472 (Taylor & Francis)
openalex publication_date 2017/05/02 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/25
The .70 and .80 reliability standards, proposed by Jum Nunnally, are widely employed across a spectrum of research domains. Nonetheless, due to their arbitrary nature, both standards fail to satisfy the needs of researchers. This paper presents a set of formulas that can be used to compute minimum reliability standards as a function of a decision criterion (margin of error). An example based on a study of preliteracy is considered, for which a two-facet generalizability theory design was employed to estimate a g-coefficient. The minimum reliability standard was used to determine whether the decision criterion was satisfied by the observed reliability coefficient.