clinical applications, the use of generalization theory was introduced to physical therapy and encouraged by other workers (Gleeson & Mercer, 1996; Roebroeck, Harlaar, Lankhrost, 1993). This theory distinguishes between generalization (G) studies, which refer to the research phase in which a measuring device or procedure is tested, and decision (D) studies, which apply to the subsequent practical use of those measurements. For a particular D study, several components of reliability should be examined in a prior G study, e.g., instrument, intra-rater, inter-rater, and intra-subject reliability (Domholdt, 1993). Primarily, the intrasubject reliability, describing the actual changes in subject performance from time to time, is studied. In this case, it is important to distinguish between relative and absolute reliability. Relative reliability describes the degree of association between repeated measurements, using a correlation coefficient, for example, the Pearson correlation coefficient or the Intra-class correlation coefficient (ICC) (Domholdt, 1993; Shrout & Fleiss, 1979; Stratford, 1989). Absolute reliability describes the variability of the scores from measurement to measurement. The statistics used are, for example, the standard error of the measurement (SEM), the measurement error (ME), the coefficient of variation (CV), and the smallest detectable difference (SDD) (Birmingham, Kramer, Speechely, Chesworth, MacDermid, 1998; Bland & Altman, 1996; Domholdt, 1993; Holmback, Porter, Downham, Lexell, 1999; Roebroeck, Harlaar, Lankhorst, 1993; Stratford, 1989; van Meeteren, Mens, Stam, 1997).
The aim of this study was to assess the absolute and relative reliability of isokinetic supine bench press in healthy women during two isokinetic velocities using the ACES `multifunction exerciser.' The performance parameters included were peak and average force, peak and average power, and total work.
Materials and methods Subjects
Twenty-four healthy women, aged 19-52 year (mean 31.3 year, standard deviation 9.6 year), from the Department of Physical Therapy at the Karolinska Institute Stockholm, Sweden, signed
an agreement to participate in the study. Eight of the subjects were accustomed to strength training and three of them did isotonic bench press exercises regularly, but none of them were used to isokinetic supine bench press exercises. A local ethics
committee agreed with the design of the study.
Testing procedure
The equipment used in this study was the ACES `multifunction exerciser''. The subjects were lying in supine position with bent legs. Hands and elbows were placed in line with proc. xiphoideus,
and feet placed on a chair to stabilize the lower back (fig. 1). The range of motion (ROM) was standardized so that the movement started from zero position in horizontal adduction and finished
'(Ariel Dynamics Inc., Trabuco Canyon, CA).
Fig. 1. Testing position of the subject performing a supine bench press using the Ariel Computerized Exercise System (ACES) `multifunction exerciser'.
with extended elbows and the shoulders in horizontal adduction2. The mean displacement of the resistance arm was 13.3° or about 34.7 cm. For each degree of displacement of the resistance arm, the hands moved 2.6 cm upwards.
The subjects performed three sets of three repetitions of maximal isokinetic supine bench press at two different velocities: 10°/s (slow) and 25°/s (moderate). The order of these two velocities was randomized at the start of the experiment and was kept the same for each individual during the study. A short break was provided between each repetition (5-10 s) and between each set (at least) 1 min to record the data. The total experiment took 20-30 min. The subjects were instructed to push the bar `as fast as possible' during expiration, and they were verbally encouraged to perform at their maximal capacity.
Illustration of one isokinetic bench press exercise
Figure 2 is a typical example of the maximal performance of one of the subjects during a 10°/s isokinetic bench press. The figure shows the maximal force produced during through the ROM in one repetition (upper curve) and the movement of the resistance arm (lower curve). The lower curve makes it possible to study the velocity of the resistance arm and to see the linearity of the movement. An isokinetic movement does not have constant velocity during the whole ROM because of an acceleration at the beginning of the movement and, in most isokinetic equipment (like Kin-Com), a deceleration at the end (Gransberg & Knutson, 1983). With ACES, there is an acceleration at the end of the movement when the stepper motor opens the valve and the resistance disappears (after approximately 2.2 s).
Only `the way up' was analyzed; there was no resistance on `the way down' (250°/s). The ACES software program, which calculates the mean of the three repetitions, calculated the performance parameters peak and average force, peak and average power, and total work.
The experiment was repeated on two different occasions with a mean interval of 4.3 days between each occasion. In this report, the term `day 1' refers to the first occasion and `day 2' to the second occasion. Three days before the experiment, a pre-test identical to the experiment was done for familiarization with the testing procedure and the ACES equipment.
Statistics
To test the presence of systematic errors (type I errors) a two-way ANOVA was used (Domholdt, 1993). The mean of three sets on each day was used to calculate the Interclass correlation
2Horizontal adduction is a combined movement of "pushing the arms forward" after 80° abduction in the shoulder joint.
9
219
