Reproducibility of the Conconi test: Test repeatability and observer variations

The repeatability of the the speed/heart rate (S/HR) relationship obtained with the incremental test developed by Conconi et al. (5) was examined by having 75 subjects perform the same running test twice in the course of a few days. From the data obtained, comparisons were made of four variables: 1) speed of deflection, 2) heart rate of deflection, 3) slope, and 4) intercept on the y-axis of the linear portion of the S/HR relationship. Straight-line equations, correlation coefficients (r), and technical errors of measurement (TEM) were obtained by comparing data from the two successive tests. Data analysis suggests that in two successive tests there are some variations in the linear portion of the S/HR relationship. However, the straight lines obtained from the two tests converge at the same deflection point. The problem of assigning a single value to the point of deflection, when determined by means of visual analysis of the S/HR graph, was also examined. For this part of the study, speed and heart rate data previously obtained for 65 subjects were used. The selected tests had been carried out during the course of various studies. The 65 sets of test data were plotted on graph paper and given to six observers for independent evaluation of speed and heart rate of deflection. These observers had different levels of experience in interpreting Conconi test results. The speed and heart rate of deflection values assigned by each observer were compared to the values obtained through mathematical analysis of the tests by computer. Straight-line equations, correlation coefficients (r), and technical errors of measurement (TEM) obtained by comparing visually determined data to those determined through mathematical analysis were calculated for each observer. Only for observers with little experience were some differences found between the observer-assigned and computer-determined results; these differences occurred for both speed and heart rate of deflection. We conclude that visual analysis provides information that is very similar to that obtained through computer analysis. The accuracy of the visually obtained information varies according to the observer's experience.