A Measure Of Similarity Between Individuals When Scoring Directions Of Variables Are Arbitrary
- 1 July 1967
- journal article
- Published by Taylor & Francis in Multivariate Behavioral Research
- Vol. 2 (3) , 377-384
- https://doi.org/10.1207/s15327906mbr0203_6
Abstract
Holley and Guilford have proposed an index, G, originally intended for use in Q type factor analysis when scoring directions are arbitrary. G is appropriate for dichotomous data only. In this paper, a general measure is derived which is identical to G in the case of dichotomous data. The measure is approximately equal to the mean inter-person covariance when all permutations of scoring directions are considered. The measure is also closely connected both with component analysis of individual differences and with the condition of proportionality between two sets of observed values.Keywords
This publication has 11 references indexed in Scilit:
- A REPLY TO JASPER HOLLEY'S DEFENSE OF THE G INDEXScandinavian Journal of Psychology, 1967
- COMMENTS ON FACTORING DICHOTOMOUS SCORE MATRICESScandinavian Journal of Psychology, 1966
- A REPLY TO PHILIP LEVY‐IN DEFENSE OF THE G INDEXScandinavian Journal of Psychology, 1966
- PROPERTIES OF THE HOLLEY-GUILFORD ‘G INDEX OF AGREEMENT’ IN R AND Q FACTOR ANALYSISScandinavian Journal of Psychology, 1966
- NOTE ON THE DOUBLE CENTERING OF DICHOTOMIZED MATRICESScandinavian Journal of Psychology, 1966
- ON THE VALIDITY OF THE RORSCHACH TESTScandinavian Journal of Psychology, 1965
- A Note on the G Index of AgreementEducational and Psychological Measurement, 1964
- A Reply to some Comments of Norman CliffEducational and Psychological Measurement, 1964
- A NOTE ON THE RELATIONSHIP BETWEEN Q AND R FACTORSScandinavian Journal of Psychology, 1964
- Mean Performance and the Factor Analysis of Learning DataPsychometrika, 1964