Premium
THE COMPARATIVE EFFECTIVENESS OF SEVERAL SHORT‐CUT ITEM ANALYSIS PROCEDURES
Author(s) -
CONNAUGHTON I. M.,
SKURNIK L. S.
Publication year - 1969
Publication title -
british journal of educational psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.557
H-Index - 95
eISSN - 2044-8279
pISSN - 0007-0998
DOI - 10.1111/j.2044-8279.1969.tb02072.x
Subject(s) - statistics , psychology , table (database) , mathematics , computer program , sample (material) , arithmetic , computer science , data mining , programming language , chemistry , chromatography
S ummary . A comparison was made between the facility and discrimination values produced by a computer item analysis program, and those produced by several short‐cut procedures based on Fan and Flanagan table values, phi and biserial correlations, normalised z and simple 27 per cent and 25 per cent difference quotients. The computer sample was composed of 584 GCE pupils who had taken the CP66 aptitude test. The short‐cut sample consisted of 185 of these pupils. The CP66 item analysis results indicated that Fan and Flanagan tables gave ‘accurate’ results in comparison with those produced by the computer. The very simple D indices, although more ‘severe’ in discrimination standards than the Fan and Flanagan tables were also quite accurate, as were the phi index and the others reviewed. In general it was concluded that where computer facilities were not available for item analysis, short‐cut methods could safely be used with minimum loss of accuracy of information even though some methods rejected more ‘good’ items than other methods.