Towards the validation of a measure of challenge and threat in sport

To date no reliable and valid measure of Challenge and Threat in sport has been developed. This study reports 4 stages in the development and validation of a Challenge and Threat in Sport (CAT-Sport) measure. In stage 1 item’s indicative of Challenge and Threat were derived from existing measures (Cognitive Appraisal Scale, Skinner and Brewer, (2004), Challenge and Threat Construal, Ptacek et al, (1994), The Primary and Secondary Appraisal Scale, Gaab et al, 2009) and focus groups with current athletes. This yielded a 21 item instrument with each item scored on a scale of 1 (strongly disagree) through to 6 (strongly agree). In stage 2, the questionnaire was distributed to a sample of 127 (n = 80 male, n = 43 female) 21.1 kilometres (km) runners (Mage = 38.87, SD = 9.69) and 70 (n =56 male, n=14 female) 32.2 km runners (Mage = 42.22, SD = 9.48). A principal components factor analysis with Oblimin rotation yielded a two-component solution, one labelled challenge (variance =α=.84) the other labelled threat (variance = α=.71) leaving a 12-item questionnaire. Criteria for components extraction included: (a) eigenvalues greater than 1.0 to indicate that a component explains more variance than any single item; (b) a minimum of around 5% explained variance per component and; (c) components loading of .4 and above (Guadagnoli and Velicer, 1988). Items that cross-loaded were also removed. In stage 3, this 12-item questionnaire was distributed to a second sample of 57(n=32 male, n=25 female) 10km runners (Mage=38.62, SD=12.86) and 90 male (n = 55 male, n=35 female) 16.1km runners (Mage = 40.53, SD = 10.59). The two-component solution was tested using Confirmatory Factor Analysis (EQS version 6.1). A correlated 2-factor model best fit the data (Satorra-Bentler scaled 115.09 (p<0.01); Goodness of Fit Index (GFI) =0.80, Root Mean Square Error of Approximation (RMSEA) =0.13, Robust Comparative Fit Index (RCFI) =0.85) although the goodness of fit indices fell below commonly agreed threshold criteria (Hu and Bentler, 1999, GFI 0.9, RMSEA 0.10, RCFI= 0.95). Stage 4 provided an exploratory examination of the instrument’s criterion validity in each sample of runners (using the final 12 items across both samples). When running times were standardized, a significant positive association was observed between challenge and running speed with all the distances combined, (r =.173, p =