Br. J. soc. d i n l’sychol. (1977). 16, 97-98

97

Printed in Great Britain

Ray’s balanced Dogmatism scale re-examined M. J. Kirton Two criticisms have been levelled at Rokeach’s (1960) Dogmatism scale. Some critics (e.g. Schulze, 1962) find it too long at 40 items, and others worry that as the items are all oriented in one direction it is liable to response bias. Despite Rokeach’s (1967) spirited defence, Ray (1970) lists himself among the latter worriers, and has published a scale which is not only fractionally shorter (36 items), but is also ‘balanced’. Half the items are positively worded - 14 taken from Rokeach (one is altered), and four from a revision by Anderson & Western (1967) - while the remaining 18 items -devised by Ray -are negative in form. As part of a larger work (Kirton, 1976), it was found possible not only to include the whole of Rokeach’s Dogmatism scale but also Ray’s own suggested items in a battery of tests given to a sample of 355 subjects. These were drawn from the general population in the UK Home Counties, and well balanced with regard to age. sex, educational level and occupational status. Ray’s items were presented with the same scoring system as Rokeach’s items: a constant of four being added prior to analysis (item range: 1-7). As the original design in which the tests were included called, inter alia, for inter-item correlation, items which yielded response distributions as skewed as 80-20 per cent were eliminated as undesirable (e.g. Maxwell, 1971). It was thought that since this degree of skewness was intolerable to the devisers of such scales, a yield of this order would give a clear hint that an item had ‘dated’ or acquired a change of meaning during transoceanic travel and so had become doubly suspect. For similar reasons items failing to correlate significantly with the rest of their parent test were also to be excluded. Together with four other similar tests included in the battery, Rokeach’s scale weathered this analysis with ease: with a theoretical range of 40-280. it yielded an observed range of 69-236, a mean of 152.78 (s.D. 27.50) for the full test. Only three items proved suspect for any reason, their exclusion moving the observed mean nearer to its theoretical position. By contrast Ray’s items were disappointing, with a theoretical range of 18-126 producing an observed range of SO-I 12 and a mean of 92.70 (s.D. 9.46). This misplaced mean was explained when 16 of the items failed to yield responses better distributed than 80-20 per cent. The remaining two items, Ray’s first and sixteenth (nos. 19 and 34 as listed in his table I), were correlated with Rokeach’s full test, yielding low coefficients (-0.134 and -0.087 respectively). Ray’s best results were cbtained with a sample of university students. Within the sample reported here, 55 students were identified but their responses were found to be somewhat more skewed than those obtained from the rest of the sample. For a general, non-student, sample Ray found low correlations, in the order of 0.2, between the negalively and positively worded items: for this UK sample correlations were lower: -0.036 with the 40-item version and -0.07 with Ray’s selected Rokeach items.

Table 1. Intercorrelations between tests Intolerance of ambiguity Intolerance of ambiguity Conservatism Inflexibility Dogmatism Dogmatism

(Budner) (MacDonald) 0.56 (Wilson) 0.36 (Go@) 0.s I (Rokeach) 0.56 (Ray) -0.13

0.59 0.60 0.48 0.03

I

0.59

-

0.44

0.52 0.06

0.19

Notes: As indicated, original scores are reversed so that all relationships are now expected Rased o n n = 286 who all completed the six tests.

-

0.04 to

be positive.

Ray reported a good internal reliability coefficient (0.64) for the 18 negative items only. This study yields a Kuder-Richardson 20 coefficient of 0.70, but this support must be tempered by knowledge of the underlying item response distribution. An indirect assessment of tests is the correlation obtained with other tests and the extent to which such results accord with theory and common sense. Of the tests included in the original work from which these data derive, six ought by any standards to be significantly, even highly, related. Besides Ray’s D (negative) 4

SCP

I6

98

M. J. Kirton

scale and Rokeach’s D scale, there were also the Inflexibility subscale from Gough’s (1956) CPI, Wilson & Patterson’s (1968) Conservatism scale, and two tests of intolerance of ambiguity: that of Budner (1962), and MacDonald’s (1970) version of Rydell & Rosen’s (1966) test. With appropriate test scores reversed so that all coefficients could be expected to be positive, the above four tests correlated 0.52, 0.44, 0.56 and 0.48 respectively with Rokeach’s D scale, whereas the correlations with Ray’s D scale were 0.06, 0.19, -0.13 and 0.03 respectively. All the tests except Ray’s in fact cluster into a well-knit group, emerging in several factor analyses as loading heavily on one factor. These data point to the inescapable conclusion that researchers would be wise to continue to use the unaltered D scale.

Acknowledgement The parent work was supported by an SSRC grant.

References ANDERSON, D. S. & WESTERN, J. S. (1967). An Inventory to Measure Students’ Attitudes. St Lucia, Brisbane: University of Queensland Press. S. (1962). Intolerance of ambiguity as a BUDNER, personality variable. J. Personality 30,29-50. COUGH,H. G. (1956). California Psychological Inventory. Palo Alto: Consulting Psychologists Press Inc. KIRTON,M. J. (1976). Adaptors and innovators: A description and measure. J . appl. Psychol. 61, 622-629. MACDONALD, A. P., Jr. (1970). Revised scale for ambiguity tolerance. Psychol. Rep. 26, 791-798. MAXWELL, A. E. (1971). Multivariate statistical methods and classification problems. Br. J . Psychiat. 119, 121-127. RAY,J. J. (1970). The development and validation of

a balanced dogmatism scale. Aust. J. Psychof. 22 (3), 253-260. ROKEACH, M. (1960). The Open and Closed Mind. New York: Basic Books. M. (1%7). Authoritarianism scales and ROKEACH, response bias: Comment on Peabody’s paper. Psycho/. Bull. 67, 349-355. RYDELL,S.T. & ROSEN,E. (1966). Measurement and some correlates of need cognition. Psychol. Rep. 19, 139-165. R. H. K. (1962). A shortened version of SCHULZE, the Rokeach Dogmatism scale. J. Psychol. Stud. 13 (2). 93-97. J. R. (1968). A new WILSON,G. D. & PATTERSON, measure of conservatism. Br. J. soc. clin. Psychol. 7 (4). 274-279.

Received 24 September 1975; revised version received 7 November 1975 Requests for reprints should be addressed to M. J. Kirton, The Hatfield Polytechnic, Birklands Annexe, London Road, St Albans, Herts.

Ray's balanced dogmatism scale re-examined.

Br. J. soc. d i n l’sychol. (1977). 16, 97-98 97 Printed in Great Britain Ray’s balanced Dogmatism scale re-examined M. J. Kirton Two criticisms ha...
136KB Sizes 0 Downloads 0 Views