Premium
In defence of ‘none of the above’
Author(s) -
GarcíaPŕrezt Miguel A.
Publication year - 1993
Publication title -
british journal of mathematical and statistical psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 3.157
H-Index - 51
eISSN - 2044-8317
pISSN - 0007-1102
DOI - 10.1111/j.2044-8317.1993.tb01013.x
Subject(s) - multinomial distribution , point estimation , point (geometry) , reading (process) , econometrics , confidence interval , estimation , empirical research , statistics , subject (documents) , interval (graph theory) , mathematics , computer science , psychology , economics , linguistics , philosophy , geometry , management , library science , combinatorics
Based on a review of empirical research on the subject, items with the last option reading ‘none of the above’ (NOTA items) have been claimed to be of no advantage over conventional items (Haladyna & Downing, 1989 a, b ). This paper comments briefly on the weakness of the evidence on which this conclusion is based, and points out some facts that might have contaminated prior empirical research on the advisability of NOTA items, thus possibly invalidating the above conclusion. A theoretical study comparing both item types is also presented in this paper. Finite state theory is used to derive multinomial models for either type of item. These models are then used to obtain maximum likelihood point and interval estimates of the examinees' abilities. Analysis of these estimates shows that NOTA items yield narrower confidence intervals than their conventional counterparts. Thus, they should be preferred over conventional items when accuracy in ability estimation is a goal.