Melbourne School of Psychological Sciences - Theses
Permanent URI for this collection
Search Results
1 - 1 of 1
-
ItemMeasuring attention in sport( 1997)Concentration, or attentional functioning, is an integral aspect of successful sporting performance yet there is not commensurate research into attention in sport. To assist in promoting research and to provide an instrument for applied purposes, a self-report questionnaire measuring attentional tendencies in sport was developed. To lay the ground work, a comprehensive review of the Test of Attentional and Interpersonal Style (TAIS - Nideffer, 1976; the major instrument used to assess attentional style in sport) was conducted. The clear conclusion was that the TAIS generally lacks external validity. A study examining the internal validity of the TAIS using two samples of21 0 psychology students in a cross-validation design failed to support the TAIS measurement model. A number of items failed to load satisfactorily (<0.30) in a Confirmatory Factor Analysis (CFA). The most damaging outcome was that 44% of items correlated better with a non-target subscale, indicating poor discriminant validity. Using the same samples, the TAIS was restructured in a cross-validation design producing a three factor model measuring broad attention, narrow attention, and poor focus, which proved to be the 'best' model. The possibility of measuring at least three aspects of attention provided the stimulus for the new test. The new test (Attention and Concentration Tendencies Survey - ACTS) was designed to measure nine dimensions: broad and narrow attention, alertness, flexibility, task-relevant attention, internal and external attention, distractibility, and flow concentration. Two hundred and eighty-six items designed specifically for sport were written to assess these dimensions. The new test went through successive stages of refinement involving content, structural, and external validation. To provide content validation, seven experts in Sport Psychology independently made item-scale ratings that were converted to factor scores indicating each item's relationship with each dimension. These factor scores were then used to develop "core" scales (i.e., the best 15 items on each scale). Empirical validation was then conducted to develop the scales based on these nine' core' scales. Data from an initial sample of university athletes (n= 100) was used to add items to the "core" scales where there was a sufficient relationship with the target 'core' scale, and demonstrated discriminant validity. These provisional scales (1=210) were cross-validated on a second sample of university athletes (n=50), resulting in an insufficient number of items representing the narrow attention and flow concentration scales. The narrow attention scale items were combined with task-relevant attention items and the flow concentration items were re-distributed to other scales. This reassignment was conducted on the initial sample of 100 and cross-validated in the second sample. After this reiteration, 110 items remained representing seven scales. Further refinement, based on the combined samples (n=150) reduced the test to 90 items. A study using two samples of university athletes (n1=72 & 112=74) indicated no problems with social desirability. The final stage of test development cross-validated the measurement model using a large sample of333 athletes of which approximately 50% were 'high' or 'elite' - standard athletes. The test was refined to 73 items representing seven scales (broad attention, narrow attention, flexibility, alertness, internal distraction, external distraction, and distractibility). The CFA overall goodness of fit statistics were only moderate (e.g., CFI=O.66). However, this fit was superior to one and two-factor models. Parameter fit was very good and reliabilities were excellent (0.87-0.94). Despite the high discriminant validity of the parameters, some high latent interscale correlations (0.60-0.72) were evident. It was argued that the test represented a good fit given the size of the model (i=73) and other supporting evidence. Higher-order models were examined and although they were not the 'best' fitting models, they did suggest some dimensions may be higher-order aspects of attentional functioning, namely, alertness and flexible attention. Preliminary external validation of the test with competitive trait anxiety, attentional style, sporting standard, and social desirability revealed some evidence for convergent and discriminant validity.