A psychometric investigation of two measures of type A behaviour in a British sample

Item-total correlations and principal components factor analyses were carried out on the responses of British males to the Jenkins Activity Survey (JAS) and the Framingham Scale (FRAM). The two male subject groups aged between 35 and 65 years, consisting of 150 and 58 men respectively, took part. The results indicated that neither Type A scale was particularly reliable, as measured by Cronbach's alpha coefficient (JAS alpha = 0.54 for the larger subject group and 0.39 for the smaller; FRAM alpha = 0.68 for the larger group and 0.61 for the smaller). The factor analyses indicated that a single factor was being measured by the scales, but that the amount of variance unaccounted for was large (83% for the JAS and 68% for the FRAM using the larger subject group). It is concluded that the JAS and Framingham Type A scales appear unsuitable for assessing Type A behaviour in individuals. The advantages of questionnaire measures of Type A behaviour generally are discussed and possible improvements to the existing scales are suggested.